bump 260 (#9002)

The ReAct reference references to MRKL paper. Corrected so that it
points to the actual ReAct paper #8964.

.github/workflows/langchain_experimental_ci.yml

@@ -1,5 +1,5 @@
 ---
-name: libs/langchain-experimental CI
+name: libs/experimental CI
 
 on:
   push:

.github/workflows/langchain_experimental_release.yml

@@ -1,5 +1,5 @@
 ---
-name: libs/langchain-experimental Release
+name: libs/experimental Release
 
 on:
   pull_request:

.github/workflows/scheduled_test.yml

@@ -0,0 +1,38 @@
+name: Scheduled tests
+
+on:
+  schedule:
+    - cron:  '0 13 * * *'
+
+env:
+  POETRY_VERSION: "1.4.2"
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    environment: Scheduled testing
+    strategy:
+      matrix:
+        python-version:
+          - "3.8"
+          - "3.9"
+          - "3.10"
+          - "3.11"
+    name: Python ${{ matrix.python-version }}
+    steps:
+      - uses: actions/checkout@v3
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: "./.github/actions/poetry_setup"
+        with:
+          python-version: ${{ matrix.python-version }}
+          poetry-version: "1.4.2"
+          install-command: |
+            echo "Running scheduled tests, installing dependencies with poetry..."
+            poetry install -E scheduled_testing
+      - name: Run tests
+        env:
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+        run: |
+          make scheduled_tests
+        shell: bash
+    secrets: inherit

Makefile

@@ -43,7 +43,12 @@ spell_fix:
 
 help:
 	@echo '----'
-	@echo 'coverage                     - run unit tests and generate coverage report'
+	@echo 'clean                        - run docs_clean and api_docs_clean'
 	@echo 'docs_build                   - build the documentation'
 	@echo 'docs_clean                   - clean the documentation build artifacts'
 	@echo 'docs_linkcheck               - run linkchecker on the documentation'
+	@echo 'api_docs_build               - build the API Reference documentation'
+	@echo 'api_docs_clean               - clean the API Reference documentation build artifacts'
+	@echo 'api_docs_linkcheck           - run linkchecker on the API Reference documentation'
+	@echo 'spell_check               	- run codespell on the project'
+	@echo 'spell_fix               		- run codespell on the project and fix the errors'

docs/api_reference/conf.py

@@ -100,6 +100,9 @@ extensions = [
 ]
 source_suffix = [".rst"]
 
+# some autodoc pydantic options are repeated in the actual template.
+# potentially user error, but there may be bugs in the sphinx extension
+# with options not being passed through correctly (from either the location in the code)
 autodoc_pydantic_model_show_json = False
 autodoc_pydantic_field_list_validators = False
 autodoc_pydantic_config_members = False
@@ -112,13 +115,6 @@ autodoc_member_order = "groupwise"
 autoclass_content = "both"
 autodoc_typehints_format = "short"
 
-autodoc_default_options = {
-    "members": True,
-    "show-inheritance": True,
-    "inherited-members": "BaseModel",
-    "undoc-members": True,
-    "special-members": "__call__",
-}
 # autodoc_typehints = "description"
 # Add any paths that contain templates here, relative to this directory.
 templates_path = ["templates"]

docs/api_reference/create_api_rst.py

@@ -1,49 +1,209 @@
-"""Script for auto-generating api_reference.rst"""
-import glob
-import re
+"""Script for auto-generating api_reference.rst."""
+import importlib
+import inspect
+import typing
 from pathlib import Path
+from typing import TypedDict, Sequence, List, Dict, Literal, Union
+from enum import Enum
+
+from pydantic import BaseModel
 
 ROOT_DIR = Path(__file__).parents[2].absolute()
+HERE = Path(__file__).parent
+
 PKG_DIR = ROOT_DIR / "libs" / "langchain" / "langchain"
 EXP_DIR = ROOT_DIR / "libs" / "experimental" / "langchain_experimental"
-WRITE_FILE = Path(__file__).parent / "api_reference.rst"
-EXP_WRITE_FILE = Path(__file__).parent / "experimental_api_reference.rst"
+WRITE_FILE = HERE / "api_reference.rst"
+EXP_WRITE_FILE = HERE / "experimental_api_reference.rst"
 
 
-def load_members(dir: Path) -> dict:
-    members: dict = {}
-    for py in glob.glob(str(dir) + "/**/*.py", recursive=True):
-        module = py[len(str(dir)) + 1 :].replace(".py", "").replace("/", ".")
-        top_level = module.split(".")[0]
-        if top_level not in members:
-            members[top_level] = {"classes": [], "functions": []}
-        with open(py, "r") as f:
-            for line in f.readlines():
-                cls = re.findall(r"^class ([^_].*)\(", line)
-                members[top_level]["classes"].extend([module + "." + c for c in cls])
-                func = re.findall(r"^def ([^_].*)\(", line)
-                afunc = re.findall(r"^async def ([^_].*)\(", line)
-                func_strings = [module + "." + f for f in func + afunc]
-                members[top_level]["functions"].extend(func_strings)
-    return members
+ClassKind = Literal["TypedDict", "Regular", "Pydantic", "enum"]
 
 
-def construct_doc(pkg: str, members: dict) -> str:
+class ClassInfo(TypedDict):
+    """Information about a class."""
+
+    name: str
+    """The name of the class."""
+    qualified_name: str
+    """The fully qualified name of the class."""
+    kind: ClassKind
+    """The kind of the class."""
+    is_public: bool
+    """Whether the class is public or not."""
+
+
+class FunctionInfo(TypedDict):
+    """Information about a function."""
+
+    name: str
+    """The name of the function."""
+    qualified_name: str
+    """The fully qualified name of the function."""
+    is_public: bool
+    """Whether the function is public or not."""
+
+
+class ModuleMembers(TypedDict):
+    """A dictionary of module members."""
+
+    classes_: Sequence[ClassInfo]
+    functions: Sequence[FunctionInfo]
+
+
+def _load_module_members(module_path: str, namespace: str) -> ModuleMembers:
+    """Load all members of a module.
+
+    Args:
+        module_path: Path to the module.
+        namespace: the namespace of the module.
+
+    Returns:
+        list: A list of loaded module objects.
+    """
+    classes_: List[ClassInfo] = []
+    functions: List[FunctionInfo] = []
+    module = importlib.import_module(module_path)
+    for name, type_ in inspect.getmembers(module):
+        if not hasattr(type_, "__module__"):
+            continue
+        if type_.__module__ != module_path:
+            continue
+
+        if inspect.isclass(type_):
+            if type(type_) == typing._TypedDictMeta:  # type: ignore
+                kind: ClassKind = "TypedDict"
+            elif issubclass(type_, Enum):
+                kind = "enum"
+            elif issubclass(type_, BaseModel):
+                kind = "Pydantic"
+            else:
+                kind = "Regular"
+
+            classes_.append(
+                ClassInfo(
+                    name=name,
+                    qualified_name=f"{namespace}.{name}",
+                    kind=kind,
+                    is_public=not name.startswith("_"),
+                )
+            )
+        elif inspect.isfunction(type_):
+            functions.append(
+                FunctionInfo(
+                    name=name,
+                    qualified_name=f"{namespace}.{name}",
+                    is_public=not name.startswith("_"),
+                )
+            )
+        else:
+            continue
+
+    return ModuleMembers(
+        classes_=classes_,
+        functions=functions,
+    )
+
+
+def _merge_module_members(
+    module_members: Sequence[ModuleMembers],
+) -> ModuleMembers:
+    """Merge module members."""
+    classes_: List[ClassInfo] = []
+    functions: List[FunctionInfo] = []
+    for module in module_members:
+        classes_.extend(module["classes_"])
+        functions.extend(module["functions"])
+
+    return ModuleMembers(
+        classes_=classes_,
+        functions=functions,
+    )
+
+
+def _load_package_modules(
+    package_directory: Union[str, Path]
+) -> Dict[str, ModuleMembers]:
+    """Recursively load modules of a package based on the file system.
+
+    Traversal based on the file system makes it easy to determine which
+    of the modules/packages are part of the package vs. 3rd party or built-in.
+
+    Parameters:
+        package_directory: Path to the package directory.
+
+    Returns:
+        list: A list of loaded module objects.
+    """
+    package_path = (
+        Path(package_directory)
+        if isinstance(package_directory, str)
+        else package_directory
+    )
+    modules_by_namespace = {}
+
+    package_name = package_path.name
+
+    for file_path in package_path.rglob("*.py"):
+        if not file_path.name.startswith("__"):
+            relative_module_name = file_path.relative_to(package_path)
+            # Get the full namespace of the module
+            namespace = str(relative_module_name).replace(".py", "").replace("/", ".")
+            # Keep only the top level namespace
+            top_namespace = namespace.split(".")[0]
+
+            try:
+                module_members = _load_module_members(
+                    f"{package_name}.{namespace}", namespace
+                )
+                # Merge module members if the namespace already exists
+                if top_namespace in modules_by_namespace:
+                    existing_module_members = modules_by_namespace[top_namespace]
+                    _module_members = _merge_module_members(
+                        [existing_module_members, module_members]
+                    )
+                else:
+                    _module_members = module_members
+
+                modules_by_namespace[top_namespace] = _module_members
+
+            except ImportError as e:
+                print(f"Error: Unable to import module '{namespace}' with error: {e}")
+
+    return modules_by_namespace
+
+
+def _construct_doc(pkg: str, members_by_namespace: Dict[str, ModuleMembers]) -> str:
+    """Construct the contents of the reference.rst file for the given package.
+
+    Args:
+        pkg: The package name
+        members_by_namespace: The members of the package, dict organized by top level
+                              module contains a list of classes and functions
+                              inside of the top level namespace.
+
+    Returns:
+        The contents of the reference.rst file.
+    """
     full_doc = f"""\
-=============
+=======================
 ``{pkg}`` API Reference
-=============
+=======================
 
 """
-    for module, _members in sorted(members.items(), key=lambda kv: kv[0]):
-        classes = _members["classes"]
+    namespaces = sorted(members_by_namespace)
+
+    for module in namespaces:
+        _members = members_by_namespace[module]
+        classes = _members["classes_"]
         functions = _members["functions"]
         if not (classes or functions):
             continue
         section = f":mod:`{pkg}.{module}`"
+        underline = "=" * (len(section) + 1)
         full_doc += f"""\
 {section}
-{'=' * (len(section) + 1)}
+{underline}
 
 .. automodule:: {pkg}.{module}
     :no-members:
@@ -52,7 +212,6 @@ def construct_doc(pkg: str, members: dict) -> str:
 """
 
         if classes:
-            cstring = "\n    ".join(sorted(classes))
             full_doc += f"""\
 Classes
 --------------
@@ -60,13 +219,31 @@ Classes
 
 .. autosummary::
     :toctree: {module}
-    :template: class.rst
-
-    {cstring}
-
 """
+
+            for class_ in classes:
+                if not class_['is_public']:
+                    continue
+                    
+                if class_["kind"] == "TypedDict":
+                    template = "typeddict.rst"
+                elif class_["kind"] == "enum":
+                    template = "enum.rst"
+                elif class_["kind"] == "Pydantic":
+                    template = "pydantic.rst"
+                else:
+                    template = "class.rst"
+
+                full_doc += f"""\
+    :template: {template}
+    
+    {class_["qualified_name"]}
+    
+"""
+
         if functions:
-            fstring = "\n    ".join(sorted(functions))
+            _functions = [f["qualified_name"] for f in functions if f["is_public"]]
+            fstring = "\n    ".join(sorted(_functions))
             full_doc += f"""\
 Functions
 --------------
@@ -83,12 +260,15 @@ Functions
 
 
 def main() -> None:
-    lc_members = load_members(PKG_DIR)
-    lc_doc = ".. _api_reference:\n\n" + construct_doc("langchain", lc_members)
+    """Generate the reference.rst file for each package."""
+    lc_members = _load_package_modules(PKG_DIR)
+    lc_doc = ".. _api_reference:\n\n" + _construct_doc("langchain", lc_members)
     with open(WRITE_FILE, "w") as f:
         f.write(lc_doc)
-    exp_members = load_members(EXP_DIR)
-    exp_doc = ".. _experimental_api_reference:\n\n" + construct_doc("langchain_experimental", exp_members)
+    exp_members = _load_package_modules(EXP_DIR)
+    exp_doc = ".. _experimental_api_reference:\n\n" + _construct_doc(
+        "langchain_experimental", exp_members
+    )
     with open(EXP_WRITE_FILE, "w") as f:
         f.write(exp_doc)
 

docs/api_reference/requirements.txt

@@ -1,4 +1,5 @@
 -e libs/langchain
+-e libs/experimental
 autodoc_pydantic==1.8.0
 myst_parser
 nbsphinx==0.8.9
@@ -10,4 +11,4 @@ sphinx-panels
 toml
 myst_nb
 sphinx_copybutton
-pydata-sphinx-theme==0.13.1
+pydata-sphinx-theme==0.13.1

docs/api_reference/templates/class.rst

@@ -5,17 +5,6 @@
 
 .. autoclass:: {{ objname }}
 
-   {% block methods %}
-   {% if methods %}
-   .. rubric:: {{ _('Methods') }}
-
-   .. autosummary::
-   {% for item in methods %}
-      ~{{ name }}.{{ item }}
-   {%- endfor %}
-   {% endif %}
-   {% endblock %}
-
    {% block attributes %}
    {% if attributes %}
    .. rubric:: {{ _('Attributes') }}
@@ -27,4 +16,21 @@
    {% endif %}
    {% endblock %}
 
+   {% block methods %}
+   {% if methods %}
+   .. rubric:: {{ _('Methods') }}
+
+   .. autosummary::
+   {% for item in methods %}
+      ~{{ name }}.{{ item }}
+   {%- endfor %}
+
+   {% for item in methods %}
+   .. automethod:: {{ name }}.{{ item }}
+   {%- endfor %}
+
+   {% endif %}
+   {% endblock %}
+
+
 .. example_links:: {{ objname }}

docs/api_reference/templates/enum.rst

@@ -0,0 +1,14 @@
+:mod:`{{module}}`.{{objname}}
+{{ underline }}==============
+
+.. currentmodule:: {{ module }}
+
+.. autoclass:: {{ objname }}
+
+    {% block attributes %}
+    {% for item in attributes %}
+    .. autoattribute:: {{ item }}
+    {% endfor %}
+    {% endblock %}
+
+.. example_links:: {{ objname }}

docs/api_reference/templates/pydantic.rst

@@ -0,0 +1,22 @@
+:mod:`{{module}}`.{{objname}}
+{{ underline }}==============
+
+.. currentmodule:: {{ module }}
+
+.. autopydantic_model:: {{ objname }}
+    :model-show-json: False
+    :model-show-config-summary: False
+    :model-show-validator-members: False
+    :model-show-field-summary: False
+    :field-signature-prefix: param
+    :members:
+    :undoc-members:
+    :inherited-members:
+    :member-order: groupwise
+    :show-inheritance: True
+    :special-members: __call__
+
+    {% block attributes %}
+    {% endblock %}
+
+.. example_links:: {{ objname }}

docs/api_reference/templates/typeddict.rst

@@ -0,0 +1,14 @@
+:mod:`{{module}}`.{{objname}}
+{{ underline }}==============
+
+.. currentmodule:: {{ module }}
+
+.. autoclass:: {{ objname }}
+
+    {% block attributes %}
+   {% for item in attributes %}
+  .. autoattribute:: {{ item }}
+   {% endfor %}
+   {% endblock %}
+
+.. example_links:: {{ objname }}

docs/api_reference/themes/scikit-learn-modern/layout.html

@@ -19,7 +19,7 @@
   {% block htmltitle %}
   <title>{{ title|striptags|e }}{{ titlesuffix }}</title>
   {% endblock %}
-  <link rel="canonical" href="http://scikit-learn.org/stable/{{pagename}}.html" />
+  <link rel="canonical" href="https://api.python.langchain.com/en/latest/{{pagename}}.html" />
 
   {% if favicon_url %}
   <link rel="shortcut icon" href="{{ favicon_url|e }}"/>

docs/api_reference/themes/scikit-learn-modern/nav.html

@@ -6,17 +6,6 @@
   {%- set top_container_cls = "sk-landing-container" %}
 {%- endif %}
 
-{% if theme_link_to_live_contributing_page|tobool %}
-{# Link to development page for live builds #}
-  {%- set development_link = "https://scikit-learn.org/dev/developers/index.html" %}
-{# Open on a new development page in new window/tab for live builds #}
-  {%- set development_attrs = 'target="_blank" rel="noopener noreferrer"' %}
-{%- else %}
-  {%- set development_link = pathto('developers/index') %}
-  {%- set development_attrs = '' %}
-{%- endif %}
-
-
 <nav id="navbar" class="{{ nav_bar_class }} navbar navbar-expand-md navbar-light bg-light py-0">
   <div class="container-fluid {{ top_container_cls }} px-0">
     {%- if logo_url %}

docs/docs_skeleton/docs/modules/agents/agent_types/index.mdx

@@ -12,7 +12,7 @@ Here are the agents available in LangChain.
 
 ### [Zero-shot ReAct](/docs/modules/agents/agent_types/react.html)
 
-This agent uses the [ReAct](https://arxiv.org/pdf/2205.00445.pdf) framework to determine which tool to use
+This agent uses the [ReAct](https://arxiv.org/pdf/2210.03629) framework to determine which tool to use
 based solely on the tool's description. Any number of tools can be provided.
 This agent requires that a description is provided for each tool.
 

docs/docs_skeleton/docs/modules/index.mdx

@@ -18,5 +18,3 @@ Let chains choose which tools to use given high-level directives
 Persist application state between runs of a chain
 #### [Callbacks](/docs/modules/callbacks/)
 Log and stream intermediate steps of any chain
-#### [Evaluation](/docs/modules/evaluation/)
-Evaluate the performance of a chain.

docs/docs_skeleton/docs/modules/model_io/prompts/index.mdx

@@ -3,10 +3,12 @@ sidebar_position: 0
 ---
 # Prompts
 
-The new way of programming models is through prompts.
-A **prompt** refers to the input to the model.
-This input is often constructed from multiple components.
-LangChain provides several classes and functions to make constructing and working with prompts easy.
+A prompt for a language model is a set of instructions or input provided by a user to
+guide the model's response, helping it understand the context and generate relevant
+and coherent language-based output, such as answering questions, completing sentences,
+or engaging in a conversation.
 
-- [Prompt templates](/docs/modules/model_io/prompts/prompt_templates/): Parametrize model inputs
+LangChain provides several classes and functions to help construct and work with prompts.
+
+- [Prompt templates](/docs/modules/model_io/prompts/prompt_templates/): Parametrized model inputs
 - [Example selectors](/docs/modules/model_io/prompts/example_selectors/): Dynamically select examples to include in prompts

docs/docs_skeleton/docs/modules/model_io/prompts/prompt_templates/index.mdx

@@ -4,18 +4,15 @@ sidebar_position: 0
 
 # Prompt templates
 
-Language models take text as input - that text is commonly referred to as a prompt.
-Typically this is not simply a hardcoded string but rather a combination of a template, some examples, and user input.
-LangChain provides several classes and functions to make constructing and working with prompts easy.
+Prompt templates are pre-defined recipes for generating prompts for language models.
 
-## What is a prompt template?
+A template may include instructions, few shot examples, and specific context and
+questions appropriate for a given task.
 
-A prompt template refers to a reproducible way to generate a prompt. It contains a text string ("the template"), that can take in a set of parameters from the end user and generates a prompt.
+LangChain provides tooling to create and work with prompt templates.
 
-A prompt template can contain:
-- instructions to the language model,
-- a set of few shot examples to help the language model generate a better response,
-- a question to the language model.
+LangChain strives to create model agnostic templates to make it easy to reuse
+existing templates across different language models.
 
 import GetStarted from "@snippets/modules/model_io/prompts/prompt_templates/get_started.mdx"
 

docs/docs_skeleton/docusaurus.config.js

@@ -128,6 +128,10 @@ const config = {
           hideable: true,
         },
       },
+      colorMode: {
+        disableSwitch: false,
+        respectPrefersColorScheme: true,
+      },
       prism: {
         theme: {
           ...baseLightCodeBlockTheme,

docs/docs_skeleton/package-lock.json

@@ -12,7 +12,7 @@
         "@docusaurus/preset-classic": "2.4.0",
         "@docusaurus/remark-plugin-npm2yarn": "^2.4.0",
         "@mdx-js/react": "^1.6.22",
-        "@mendable/search": "^0.0.125",
+        "@mendable/search": "^0.0.137",
         "clsx": "^1.2.1",
         "json-loader": "^0.5.7",
         "process": "^0.11.10",
@@ -3212,10 +3212,11 @@
       }
     },
     "node_modules/@mendable/search": {
-      "version": "0.0.125",
-      "resolved": "https://registry.npmjs.org/@mendable/search/-/search-0.0.125.tgz",
-      "integrity": "sha512-Mb1J3zDhOyBZV9cXqJocSOBNYGpe8+LQDqd9n9laPWxosSJcSTUewqtlIbMerrYsScBsxskoSiWgRsc7xF5z0Q==",
+      "version": "0.0.137",
+      "resolved": "https://registry.npmjs.org/@mendable/search/-/search-0.0.137.tgz",
+      "integrity": "sha512-2J2fd5eqToK+mLzrSDA6NAr4F1kfql7QRiHpD7AUJJX0nqpvInhr/mMJKBCUSCv2z76UKCmF5wLuPSw+C90Qdg==",
       "dependencies": {
+        "html-react-parser": "^4.2.0",
         "posthog-js": "^1.45.1"
       },
       "peerDependencies": {
@@ -8332,6 +8333,33 @@
         "safe-buffer": "~5.1.0"
       }
     },
+    "node_modules/html-dom-parser": {
+      "version": "4.0.0",
+      "resolved": "https://registry.npmjs.org/html-dom-parser/-/html-dom-parser-4.0.0.tgz",
+      "integrity": "sha512-TUa3wIwi80f5NF8CVWzkopBVqVAtlawUzJoLwVLHns0XSJGynss4jiY0mTWpiDOsuyw+afP+ujjMgRh9CoZcXw==",
+      "dependencies": {
+        "domhandler": "5.0.3",
+        "htmlparser2": "9.0.0"
+      }
+    },
+    "node_modules/html-dom-parser/node_modules/htmlparser2": {
+      "version": "9.0.0",
+      "resolved": "https://registry.npmjs.org/htmlparser2/-/htmlparser2-9.0.0.tgz",
+      "integrity": "sha512-uxbSI98wmFT/G4P2zXx4OVx04qWUmyFPrD2/CNepa2Zo3GPNaCaaxElDgwUrwYWkK1nr9fft0Ya8dws8coDLLQ==",
+      "funding": [
+        "https://github.com/fb55/htmlparser2?sponsor=1",
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/fb55"
+        }
+      ],
+      "dependencies": {
+        "domelementtype": "^2.3.0",
+        "domhandler": "^5.0.3",
+        "domutils": "^3.1.0",
+        "entities": "^4.5.0"
+      }
+    },
     "node_modules/html-entities": {
       "version": "2.4.0",
       "resolved": "https://registry.npmjs.org/html-entities/-/html-entities-2.4.0.tgz",
@@ -8375,6 +8403,20 @@
         "node": ">= 12"
       }
     },
+    "node_modules/html-react-parser": {
+      "version": "4.2.0",
+      "resolved": "https://registry.npmjs.org/html-react-parser/-/html-react-parser-4.2.0.tgz",
+      "integrity": "sha512-gzU55AS+FI6qD7XaKe5BLuLFM2Xw0/LodfMWZlxV9uOHe7LCD5Lukx/EgYuBI3c0kLu0XlgFXnSzO0qUUn3Vrg==",
+      "dependencies": {
+        "domhandler": "5.0.3",
+        "html-dom-parser": "4.0.0",
+        "react-property": "2.0.0",
+        "style-to-js": "1.1.3"
+      },
+      "peerDependencies": {
+        "react": "0.14 || 15 || 16 || 17 || 18"
+      }
+    },
     "node_modules/html-tags": {
       "version": "3.3.1",
       "resolved": "https://registry.npmjs.org/html-tags/-/html-tags-3.3.1.tgz",
@@ -11762,6 +11804,11 @@
         "webpack": ">=4.41.1 || 5.x"
       }
     },
+    "node_modules/react-property": {
+      "version": "2.0.0",
+      "resolved": "https://registry.npmjs.org/react-property/-/react-property-2.0.0.tgz",
+      "integrity": "sha512-kzmNjIgU32mO4mmH5+iUyrqlpFQhF8K2k7eZ4fdLSOPFrD1XgEuSBv9LDEgxRXTMBqMd8ppT0x6TIzqE5pdGdw=="
+    },
     "node_modules/react-router": {
       "version": "5.3.4",
       "resolved": "https://registry.npmjs.org/react-router/-/react-router-5.3.4.tgz",
@@ -13127,6 +13174,22 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
+    "node_modules/style-to-js": {
+      "version": "1.1.3",
+      "resolved": "https://registry.npmjs.org/style-to-js/-/style-to-js-1.1.3.tgz",
+      "integrity": "sha512-zKI5gN/zb7LS/Vm0eUwjmjrXWw8IMtyA8aPBJZdYiQTXj4+wQ3IucOLIOnF7zCHxvW8UhIGh/uZh/t9zEHXNTQ==",
+      "dependencies": {
+        "style-to-object": "0.4.1"
+      }
+    },
+    "node_modules/style-to-js/node_modules/style-to-object": {
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/style-to-object/-/style-to-object-0.4.1.tgz",
+      "integrity": "sha512-HFpbb5gr2ypci7Qw+IOhnP2zOU7e77b+rzM+wTzXzfi1PrtBCX0E7Pk4wL4iTLnhzZ+JgEGAhX81ebTg/aYjQw==",
+      "dependencies": {
+        "inline-style-parser": "0.1.1"
+      }
+    },
     "node_modules/style-to-object": {
       "version": "0.3.0",
       "resolved": "https://registry.npmjs.org/style-to-object/-/style-to-object-0.3.0.tgz",

docs/docs_skeleton/package.json

@@ -23,7 +23,7 @@
     "@docusaurus/preset-classic": "2.4.0",
     "@docusaurus/remark-plugin-npm2yarn": "^2.4.0",
     "@mdx-js/react": "^1.6.22",
-    "@mendable/search": "^0.0.125",
+    "@mendable/search": "^0.0.137",
     "clsx": "^1.2.1",
     "json-loader": "^0.5.7",
     "process": "^0.11.10",

docs/docs_skeleton/vercel.json

@@ -556,6 +556,14 @@
       "source": "/docs/integrations/llamacpp",
       "destination": "/docs/integrations/providers/llamacpp"
     },
+    {
+      "source": "/en/latest/integrations/log10.html",
+      "destination": "/docs/integrations/providers/log10"
+    },
+    {
+      "source": "/docs/integrations/log10",
+      "destination": "/docs/integrations/providers/log10"
+    },
     {
       "source": "/en/latest/integrations/mediawikidump.html",
       "destination": "/docs/integrations/providers/mediawikidump"

docs/extras/guides/expression_language/cookbook.ipynb

@@ -22,7 +22,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 1,
    "id": "466b65b3",
    "metadata": {},
    "outputs": [],
@@ -171,9 +171,7 @@
    "cell_type": "code",
    "execution_count": 9,
    "id": "decf7710",
-   "metadata": {
-    "scrolled": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "data": {
@@ -202,7 +200,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": 10,
    "id": "f799664d",
    "metadata": {},
    "outputs": [],
@@ -347,7 +345,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 13,
+   "execution_count": 12,
    "id": "5d3d8ffe",
    "metadata": {},
    "outputs": [],
@@ -368,7 +366,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
+   "execution_count": 2,
    "id": "33be32af",
    "metadata": {},
    "outputs": [],
@@ -380,7 +378,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
+   "execution_count": 3,
    "id": "df3f3fa2",
    "metadata": {},
    "outputs": [],
@@ -424,9 +422,7 @@
    "cell_type": "code",
    "execution_count": 18,
    "id": "f3040b0c",
-   "metadata": {
-    "scrolled": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stderr",
@@ -477,9 +473,7 @@
    "cell_type": "code",
    "execution_count": 20,
    "id": "7ee8b2d4",
-   "metadata": {
-    "scrolled": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stderr",
@@ -515,7 +509,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 66,
+   "execution_count": 4,
    "id": "3f30c348",
    "metadata": {},
    "outputs": [],
@@ -526,7 +520,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 21,
+   "execution_count": 5,
    "id": "64ab1dbf",
    "metadata": {},
    "outputs": [],
@@ -544,7 +538,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 22,
+   "execution_count": 6,
    "id": "7d628c97",
    "metadata": {},
    "outputs": [],
@@ -559,7 +553,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 68,
+   "execution_count": 7,
    "id": "f60a5d0f",
    "metadata": {},
    "outputs": [],
@@ -572,7 +566,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 69,
+   "execution_count": 8,
    "id": "7d007db6",
    "metadata": {},
    "outputs": [],
@@ -589,25 +583,29 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 70,
+   "execution_count": 16,
    "id": "5c32cc89",
    "metadata": {},
    "outputs": [],
    "source": [
-    "conversational_qa_chain = RunnableMap({\n",
-    "    \"standalone_question\": {\n",
-    "        \"question\": lambda x: x[\"question\"],\n",
-    "        \"chat_history\": lambda x: _format_chat_history(x['chat_history'])\n",
-    "    } | CONDENSE_QUESTION_PROMPT | ChatOpenAI(temperature=0) | StrOutputParser(),\n",
-    "}) | {\n",
+    "_inputs = RunnableMap(\n",
+    "    {\n",
+    "        \"standalone_question\": {\n",
+    "            \"question\": lambda x: x[\"question\"],\n",
+    "            \"chat_history\": lambda x: _format_chat_history(x['chat_history'])\n",
+    "        } | CONDENSE_QUESTION_PROMPT | ChatOpenAI(temperature=0) | StrOutputParser(),\n",
+    "    }\n",
+    ")\n",
+    "_context = {\n",
     "    \"context\": itemgetter(\"standalone_question\") | retriever | _combine_documents,\n",
     "    \"question\": lambda x: x[\"standalone_question\"]\n",
-    "} | ANSWER_PROMPT | ChatOpenAI()"
+    "}\n",
+    "conversational_qa_chain = _inputs | _context | ANSWER_PROMPT | ChatOpenAI()"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 71,
+   "execution_count": 17,
    "id": "135c8205",
    "metadata": {},
    "outputs": [
@@ -624,7 +622,7 @@
        "AIMessage(content='Harrison was employed at Kensho.', additional_kwargs={}, example=False)"
       ]
      },
-     "execution_count": 71,
+     "execution_count": 17,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -638,7 +636,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 62,
+   "execution_count": 15,
    "id": "424e7e7a",
    "metadata": {},
    "outputs": [
@@ -655,7 +653,7 @@
        "AIMessage(content='Harrison worked at Kensho.', additional_kwargs={}, example=False)"
       ]
      },
-     "execution_count": 62,
+     "execution_count": 15,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -667,6 +665,149 @@
     "})"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "c5543183",
+   "metadata": {},
+   "source": [
+    "### With Memory and returning source documents\n",
+    "\n",
+    "This shows how to use memory with the above. For memory, we need to manage that outside at the memory. For returning the retrieved documents, we just need to pass them through all the way."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "e31dd17c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.memory import ConversationBufferMemory"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 44,
+   "id": "d4bffe94",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "memory = ConversationBufferMemory(return_messages=True, output_key=\"answer\", input_key=\"question\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 45,
+   "id": "733be985",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# First we add a step to load memory\n",
+    "# This needs to be a RunnableMap because its the first input\n",
+    "loaded_memory = RunnableMap(\n",
+    "    {\n",
+    "        \"question\": itemgetter(\"question\"),\n",
+    "        \"memory\": memory.load_memory_variables,\n",
+    "    }\n",
+    ")\n",
+    "# Next we add a step to expand memory into the variables\n",
+    "expanded_memory = {\n",
+    "    \"question\": itemgetter(\"question\"),\n",
+    "    \"chat_history\": lambda x: x[\"memory\"][\"history\"]\n",
+    "}\n",
+    "\n",
+    "# Now we calculate the standalone question\n",
+    "standalone_question = {\n",
+    "    \"standalone_question\": {\n",
+    "        \"question\": lambda x: x[\"question\"],\n",
+    "        \"chat_history\": lambda x: _format_chat_history(x['chat_history'])\n",
+    "    } | CONDENSE_QUESTION_PROMPT | ChatOpenAI(temperature=0) | StrOutputParser(),\n",
+    "}\n",
+    "# Now we retrieve the documents\n",
+    "retrieved_documents = {\n",
+    "    \"docs\": itemgetter(\"standalone_question\") | retriever,\n",
+    "    \"question\": lambda x: x[\"standalone_question\"]\n",
+    "}\n",
+    "# Now we construct the inputs for the final prompt\n",
+    "final_inputs = {\n",
+    "    \"context\": lambda x: _combine_documents(x[\"docs\"]),\n",
+    "    \"question\": itemgetter(\"question\")\n",
+    "}\n",
+    "# And finally, we do the part that returns the answers\n",
+    "answer = {\n",
+    "    \"answer\": final_inputs | ANSWER_PROMPT | ChatOpenAI(),\n",
+    "    \"docs\": itemgetter(\"docs\"),\n",
+    "}\n",
+    "# And now we put it all together!\n",
+    "final_chain = loaded_memory | expanded_memory | standalone_question | retrieved_documents | answer"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 46,
+   "id": "806e390c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Number of requested results 4 is greater than number of elements in index 1, updating n_results = 1\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'answer': AIMessage(content='Harrison was employed at Kensho.', additional_kwargs={}, example=False),\n",
+       " 'docs': [Document(page_content='harrison worked at kensho', metadata={})]}"
+      ]
+     },
+     "execution_count": 46,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "inputs = {\"question\": \"where did harrison work?\"}\n",
+    "result = final_chain.invoke(inputs)\n",
+    "result"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 47,
+   "id": "977399fd",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Note that the memory does not save automatically\n",
+    "# This will be improved in the future\n",
+    "# For now you need to save it yourself\n",
+    "memory.save_context(inputs, {\"answer\": result[\"answer\"].content})"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 48,
+   "id": "f94f7de4",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'history': [HumanMessage(content='where did harrison work?', additional_kwargs={}, example=False),\n",
+       "  AIMessage(content='Harrison was employed at Kensho.', additional_kwargs={}, example=False)]}"
+      ]
+     },
+     "execution_count": 48,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "memory.load_memory_variables({})"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "0f2bf8d3",
@@ -1391,10 +1532,299 @@
     "response"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "4927a727-b4c8-453c-8c83-bd87b4fcac14",
+   "metadata": {},
+   "source": [
+    "## Moderation\n",
+    "\n",
+    "This shows how to add in moderation (or other safeguards) around your LLM application."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "4f5f6449-940a-4f5c-97c0-39b71c3e2a68",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.chains import OpenAIModerationChain\n",
+    "from langchain.llms import OpenAI"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 35,
+   "id": "fcb8312b-7e7a-424f-a3ec-76738c9a9d21",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "moderate = OpenAIModerationChain()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 32,
+   "id": "b24b9148-f6b0-4091-8ea8-d3fb281bd950",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model = OpenAI()\n",
+    "prompt = ChatPromptTemplate.from_messages([\n",
+    "    (\"system\", \"repeat after me: {input}\")\n",
+    "])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 33,
+   "id": "1c8ed87c-9ca6-4559-bf60-d40e94a0af08",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "chain = prompt | model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "id": "5256b9bd-381a-42b0-bfa8-7e6d18f853cb",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'\\n\\nYou are stupid.'"
+      ]
+     },
+     "execution_count": 34,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "chain.invoke({\"input\": \"you are stupid\"})"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 36,
+   "id": "fe6e3b33-dc9a-49d5-b194-ba750c58a628",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "moderated_chain = chain | moderate"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 37,
+   "id": "d8ba0cbd-c739-4d23-be9f-6ae092bd5ffb",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'input': '\\n\\nYou are stupid.',\n",
+       " 'output': \"Text was found that violates OpenAI's content policy.\"}"
+      ]
+     },
+     "execution_count": 37,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "moderated_chain.invoke({\"input\": \"you are stupid\"})"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "179d3c03",
+   "id": "a0a85ba4-f782-47b8-b16f-8b7a61d6dab7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "## Conversational Retrieval With Memory"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "92c87dd8-bb6f-4f32-a30d-8f5459ce6265",
+   "metadata": {},
+   "source": [
+    "## Fallbacks\n",
+    "\n",
+    "With LCEL you can easily introduce fallbacks for any Runnable component, like an LLM."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "1b1cb744-31fc-4261-ab25-65fe1fcad559",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "AIMessage(content='To get to the other side.', additional_kwargs={}, example=False)"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from langchain.chat_models import ChatOpenAI\n",
+    "\n",
+    "bad_llm = ChatOpenAI(model_name=\"gpt-fake\")\n",
+    "good_llm = ChatOpenAI(model_name=\"gpt-3.5-turbo\")\n",
+    "llm = bad_llm.with_fallbacks([good_llm])\n",
+    "\n",
+    "llm.invoke(\"Why did the the chicken cross the road?\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b8cf3982-03f6-49b3-8ff5-7cd12444f19c",
+   "metadata": {},
+   "source": [
+    "Looking at the trace, we can see that the first model failed but the second succeeded, so we still got an output: https://smith.langchain.com/public/dfaf0bf6-d86d-43e9-b084-dd16a56df15c/r\n",
+    "\n",
+    "We can add an arbitrary sequence of fallbacks, which will be executed in order:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "31819be0-7f40-4e67-b5ab-61340027b948",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "AIMessage(content='To get to the other side.', additional_kwargs={}, example=False)"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "llm = bad_llm.with_fallbacks([bad_llm, bad_llm, good_llm])\n",
+    "\n",
+    "llm.invoke(\"Why did the the chicken cross the road?\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "acad6e88-8046-450e-b005-db7e50f33b80",
+   "metadata": {},
+   "source": [
+    "Trace: https://smith.langchain.com/public/c09efd01-3184-4369-a225-c9da8efcaf47/r\n",
+    "\n",
+    "We can continue to use our Runnable with fallbacks the same way we use any Runnable, mean we can include it in sequences:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "bab114a1-bb93-4b7e-a639-e7e00f21aebc",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "AIMessage(content='To show off its incredible jumping skills! Kangaroos are truly amazing creatures.', additional_kwargs={}, example=False)"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from langchain.prompts import ChatPromptTemplate\n",
+    "\n",
+    "prompt = ChatPromptTemplate.from_messages(\n",
+    "    [\n",
+    "        (\"system\", \"You're a nice assistant who always includes a compliment in your response\"),\n",
+    "        (\"human\", \"Why did the {animal} cross the road\"),\n",
+    "    ]\n",
+    ")\n",
+    "chain = prompt | llm\n",
+    "chain.invoke({\"animal\": \"kangaroo\"})"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "58340afa-8187-4ffe-9bd2-7912fb733a15",
+   "metadata": {},
+   "source": [
+    "Trace: https://smith.langchain.com/public/ba03895f-f8bd-4c70-81b7-8b930353eabd/r\n",
+    "\n",
+    "Note, since every sequence of Runnables is itself a Runnable, we can create fallbacks for whole Sequences. We can also continue using the full interface, including asynchronous calls, batched calls, and streams:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "45aa3170-b2e6-430d-887b-bd879048060a",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[\"\\n\\nAnswer: The rabbit crossed the road to get to the other side. That's quite clever of him!\",\n",
+       " '\\n\\nAnswer: The turtle crossed the road to get to the other side. You must be pretty clever to come up with that riddle!']"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from langchain.llms import OpenAI\n",
+    "from langchain.prompts import PromptTemplate\n",
+    "\n",
+    "chat_prompt = ChatPromptTemplate.from_messages(\n",
+    "    [\n",
+    "        (\"system\", \"You're a nice assistant who always includes a compliment in your response\"),\n",
+    "        (\"human\", \"Why did the {animal} cross the road\"),\n",
+    "    ]\n",
+    ")\n",
+    "chat_model = ChatOpenAI(model_name=\"gpt-fake\")\n",
+    "\n",
+    "prompt_template = \"\"\"Instructions: You should always include a compliment in your response.\n",
+    "\n",
+    "Question: Why did the {animal} cross the road?\"\"\"\n",
+    "prompt = PromptTemplate.from_template(prompt_template)\n",
+    "llm = OpenAI()\n",
+    "\n",
+    "bad_chain = chat_prompt | chat_model\n",
+    "good_chain = prompt | llm\n",
+    "chain = bad_chain.with_fallbacks([good_chain])\n",
+    "await chain.abatch([{\"animal\": \"rabbit\"}, {\"animal\": \"turtle\"}])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "af6731c6-0c73-4b1d-a433-6e8f6ecce2bb",
+   "metadata": {},
+   "source": [
+    "Traces: \n",
+    "1. https://smith.langchain.com/public/ccd73236-9ae5-48a6-94b5-41210be18a46/r\n",
+    "2. https://smith.langchain.com/public/f43f608e-075c-45c7-bf73-b64e4d3f3082/r"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3d2fe1fe-506b-4ee5-8056-8b9df801765f",
    "metadata": {},
    "outputs": [],
    "source": []

docs/extras/guides/expression_language/interface.ipynb

@@ -108,7 +108,7 @@
    ],
    "source": [
     "for s in chain.stream({\"topic\": \"bears\"}):\n",
-    "    print(s.content, end=\"\")"
+    "    print(s.content, end=\"\", flush=True)"
    ]
   },
   {
@@ -196,7 +196,7 @@
    ],
    "source": [
     "async for s in chain.astream({\"topic\": \"bears\"}):\n",
-    "    print(s.content, end=\"\")"
+    "    print(s.content, end=\"\", flush=True)"
    ]
   },
   {

docs/extras/integrations/callbacks/streamlit.md

@@ -71,3 +71,6 @@ or any other local ENV management tool.
 
 Currently `StreamlitCallbackHandler` is geared towards use with a LangChain Agent Executor. Support for additional agent types,
 use directly with Chains, etc will be added in the future.
+
+You may also be interested in using
+[StreamlitChatMessageHistory](/docs/integrations/memory/streamlit_chat_message_history) for LangChain.

docs/extras/integrations/chat/anyscale.ipynb

@@ -0,0 +1,225 @@
+{
+ "cells": [
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "id": "642fd21c-600a-47a1-be96-6e1438b421a9",
+   "metadata": {},
+   "source": [
+    "# Anyscale\n",
+    "\n",
+    "This notebook demonstrates the use of `langchain.chat_models.ChatAnyscale` for [Anyscale Endpoints](https://endpoints.anyscale.com/).\n",
+    "\n",
+    "* Set `ANYSCALE_API_KEY` environment variable\n",
+    "* or use the `anyscale_api_key` keyword argument"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "# !pip install openai"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "d00d850917865298"
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "72340871-ae2f-415f-b399-0777d32dc379",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdin",
+     "output_type": "stream",
+     "text": [
+      " ········\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "from getpass import getpass\n",
+    "\n",
+    "os.environ[\"ANYSCALE_API_KEY\"] = getpass()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5d7fc704-3ea0-4c35-96e7-89fcae6c73fa",
+   "metadata": {},
+   "source": [
+    "# Let's try out each model offered on Anyscale Endpoints"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "0dc9428d-4217-47d2-97de-f784b1764186",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "dict_keys(['meta-llama/Llama-2-70b-chat-hf', 'meta-llama/Llama-2-7b-chat-hf', 'meta-llama/Llama-2-13b-chat-hf'])\n"
+     ]
+    }
+   ],
+   "source": [
+    "from langchain.chat_models import ChatAnyscale\n",
+    "\n",
+    "chats = {\n",
+    "    model: ChatAnyscale(model_name=model, temperature=1.0)\n",
+    "    for model in ChatAnyscale.get_available_models()\n",
+    "}\n",
+    "\n",
+    "print(chats.keys())"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7c4f124a-eaf7-4d78-a2c0-b0aa23fb25c4",
+   "metadata": {},
+   "source": [
+    "# We can use async methods and other stuff supported by ChatOpenAI\n",
+    "\n",
+    "This way, the three requests will only take as long as the longest individual request."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "1f94f5d2-569e-4a2c-965e-de53c2845fbb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import asyncio\n",
+    "\n",
+    "from langchain.schema import SystemMessage, HumanMessage\n",
+    "\n",
+    "messages = [\n",
+    "    SystemMessage(\n",
+    "        content=\"You are a helpful AI that shares everything you know.\"\n",
+    "    ),\n",
+    "    HumanMessage(\n",
+    "        content=\"Tell me technical facts about yourself. Are you a transformer model? How many billions of parameters do you have?\"\n",
+    "    ),\n",
+    "]\n",
+    "\n",
+    "async def get_msgs():\n",
+    "    tasks = [\n",
+    "        chat.apredict_messages(messages)\n",
+    "        for chat in chats.values()\n",
+    "    ]\n",
+    "    responses = await asyncio.gather(*tasks)\n",
+    "    return dict(zip(chats.keys(), responses))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "b2ced871-869a-4ca6-a2ec-6bfececdf7da",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import nest_asyncio\n",
+    "\n",
+    "nest_asyncio.apply()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "bc605fa5-9501-470d-a6c9-cd868d2145ef",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\tmeta-llama/Llama-2-70b-chat-hf\n",
+      "\n",
+      "Greetings! I'm just an AI, I don't have a personal identity like humans do, but I'm here to help you with any questions you have.\n",
+      "\n",
+      "I'm a large language model, which means I'm trained on a large corpus of text data to generate language outputs that are coherent and natural-sounding. My architecture is based on a transformer model, which is a type of neural network that's particularly well-suited for natural language processing tasks.\n",
+      "\n",
+      "As for my parameters, I have a few billion parameters, but I don't have access to the exact number as it's not relevant to my functioning. My training data includes a vast amount of text from various sources, including books, articles, and websites, which I use to learn patterns and relationships in language.\n",
+      "\n",
+      "I'm designed to be a helpful tool for a variety of tasks, such as answering questions, providing information, and generating text. I'm constantly learning and improving my abilities through machine learning algorithms and feedback from users like you.\n",
+      "\n",
+      "I hope this helps! Is there anything else you'd like to know about me or my capabilities?\n",
+      "\n",
+      "---\n",
+      "\n",
+      "\tmeta-llama/Llama-2-7b-chat-hf\n",
+      "\n",
+      "Ah, a fellow tech enthusiast! *adjusts glasses* I'm glad to share some technical details about myself. 🤓\n",
+      "Indeed, I'm a transformer model, specifically a BERT-like language model trained on a large corpus of text data. My architecture is based on the transformer framework, which is a type of neural network designed for natural language processing tasks. 🏠\n",
+      "As for the number of parameters, I have approximately 340 million. *winks* That's a pretty hefty number, if I do say so myself! These parameters allow me to learn and represent complex patterns in language, such as syntax, semantics, and more. 🤔\n",
+      "But don't ask me to do math in my head – I'm a language model, not a calculating machine! 😅 My strengths lie in understanding and generating human-like text, so feel free to chat with me anytime you'd like. 💬\n",
+      "Now, do you have any more technical questions for me? Or would you like to engage in a nice chat? 😊\n",
+      "\n",
+      "---\n",
+      "\n",
+      "\tmeta-llama/Llama-2-13b-chat-hf\n",
+      "\n",
+      "Hello! As a friendly and helpful AI, I'd be happy to share some technical facts about myself.\n",
+      "\n",
+      "I am a transformer-based language model, specifically a variant of the BERT (Bidirectional Encoder Representations from Transformers) architecture. BERT was developed by Google in 2018 and has since become one of the most popular and widely-used AI language models.\n",
+      "\n",
+      "Here are some technical details about my capabilities:\n",
+      "\n",
+      "1. Parameters: I have approximately 340 million parameters, which are the numbers that I use to learn and represent language. This is a relatively large number of parameters compared to some other languages models, but it allows me to learn and understand complex language patterns and relationships.\n",
+      "2. Training: I was trained on a large corpus of text data, including books, articles, and other sources of written content. This training allows me to learn about the structure and conventions of language, as well as the relationships between words and phrases.\n",
+      "3. Architectures: My architecture is based on the transformer model, which is a type of neural network that is particularly well-suited for natural language processing tasks. The transformer model uses self-attention mechanisms to allow the model to \"attend\" to different parts of the input text, allowing it to capture long-range dependencies and contextual relationships.\n",
+      "4. Precision: I am capable of generating text with high precision and accuracy, meaning that I can produce text that is close to human-level quality in terms of grammar, syntax, and coherence.\n",
+      "5. Generative capabilities: In addition to being able to generate text based on prompts and questions, I am also capable of generating text based on a given topic or theme. This allows me to create longer, more coherent pieces of text that are organized around a specific idea or concept.\n",
+      "\n",
+      "Overall, I am a powerful and versatile language model that is capable of a wide range of natural language processing tasks. I am constantly learning and improving, and I am here to help answer any questions you may have!\n",
+      "\n",
+      "---\n",
+      "\n",
+      "CPU times: user 371 ms, sys: 15.5 ms, total: 387 ms\n",
+      "Wall time: 12 s\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%time\n",
+    "\n",
+    "response_dict = asyncio.run(get_msgs())\n",
+    "\n",
+    "for model_name, response in response_dict.items():\n",
+    "    print(f'\\t{model_name}')\n",
+    "    print()\n",
+    "    print(response.content)\n",
+    "    print('\\n---\\n')"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.4"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/chat/azure_chat_openai.ipynb

@@ -74,6 +74,124 @@
    "metadata": {},
    "outputs": [],
    "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f27fa24d",
+   "metadata": {},
+   "source": [
+    "## Model Version\n",
+    "Azure OpenAI responses contain `model` property, which is name of the model used to generate the response. However unlike native OpenAI responses, it does not contain the version of the model, which is set on the deplyoment in Azure. This makes it tricky to know which version of the model was used to generate the response, which as result can lead to e.g. wrong total cost calculation with `OpenAICallbackHandler`.\n",
+    "\n",
+    "To solve this problem, you can pass `model_version` parameter to `AzureChatOpenAI` class, which will be added to the model name in the llm output. This way you can easily distinguish between different versions of the model."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0531798a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.callbacks import get_openai_callback"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "3fd97dfc",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "BASE_URL = \"https://{endpoint}.openai.azure.com\"\n",
+    "API_KEY = \"...\"\n",
+    "DEPLOYMENT_NAME = \"gpt-35-turbo\" # in Azure, this deployment has version 0613 - input and output tokens are counted separately"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "aceddb72",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total Cost (USD): $0.000054\n"
+     ]
+    }
+   ],
+   "source": [
+    "model = AzureChatOpenAI(\n",
+    "    openai_api_base=BASE_URL,\n",
+    "    openai_api_version=\"2023-05-15\",\n",
+    "    deployment_name=DEPLOYMENT_NAME,\n",
+    "    openai_api_key=API_KEY,\n",
+    "    openai_api_type=\"azure\",\n",
+    ")\n",
+    "with get_openai_callback() as cb:\n",
+    "    model(\n",
+    "        [\n",
+    "            HumanMessage(\n",
+    "                content=\"Translate this sentence from English to French. I love programming.\"\n",
+    "            )\n",
+    "        ]\n",
+    "    )\n",
+    "    print(f\"Total Cost (USD): ${format(cb.total_cost, '.6f')}\") # without specifying the model version, flat-rate 0.002 USD per 1k input and output tokens is used\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2e61eefd",
+   "metadata": {},
+   "source": [
+    "We can provide the model version to `AzureChatOpenAI` constructor. It will get appended to the model name returned by Azure OpenAI and cost will be counted correctly."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "8d5e54e9",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total Cost (USD): $0.000044\n"
+     ]
+    }
+   ],
+   "source": [
+    "model0613 = AzureChatOpenAI(\n",
+    "    openai_api_base=BASE_URL,\n",
+    "    openai_api_version=\"2023-05-15\",\n",
+    "    deployment_name=DEPLOYMENT_NAME,\n",
+    "    openai_api_key=API_KEY,\n",
+    "    openai_api_type=\"azure\",\n",
+    "    model_version=\"0613\"\n",
+    ")\n",
+    "with get_openai_callback() as cb:\n",
+    "    model0613(\n",
+    "        [\n",
+    "            HumanMessage(\n",
+    "                content=\"Translate this sentence from English to French. I love programming.\"\n",
+    "            )\n",
+    "        ]\n",
+    "    )\n",
+    "    print(f\"Total Cost (USD): ${format(cb.total_cost, '.6f')}\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "99682534",
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {
@@ -92,7 +210,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.9.1"
+   "version": "3.8.10"
   }
  },
  "nbformat": 4,

docs/extras/integrations/document_loaders/airbyte_cdk.ipynb

@@ -0,0 +1,226 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte CDK"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "A lot of source connectors are implemented using the [Airbyte CDK](https://docs.airbyte.com/connector-development/cdk-python/). This loader allows to run any of these connectors and return the data as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-cdk` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-cdk"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "085aa658",
+   "metadata": {},
+   "source": [
+    "Then, either install an existing connector from the [Airbyte Github repository](https://github.com/airbytehq/airbyte/tree/master/airbyte-integrations/connectors) or create your own connector using the [Airbyte CDK](https://docs.airbyte.io/connector-development/connector-development).\n",
+    "\n",
+    "For example, to install the Github connector, run"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f6d04ef4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install \"source_github@git+https://github.com/airbytehq/airbyte.git@master#subdirectory=airbyte-integrations/connectors/source-github\""
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "36069b74",
+   "metadata": {},
+   "source": [
+    "Some sources are also published as regular packages on PyPI"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Now you can create an `AirbyteCDKLoader` based on the imported source. It takes a `config` object that's passed to the connector. You also have to pick the stream you want to retrieve records from by name (`stream_name`). Check the connectors documentation page and spec definition for more information on the config object and available streams. For the Github connectors these are:\n",
+    "* [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-github/source_github/spec.json](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-github/source_github/spec.json).\n",
+    "* [https://docs.airbyte.com/integrations/sources/github/](https://docs.airbyte.com/integrations/sources/github/)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteCDKLoader\n",
+    "from source_github.source import SourceGithub # plug in your own source here\n",
+    "\n",
+    "config = {\n",
+    "    # your github configuration\n",
+    "    \"credentials\": {\n",
+    "        \"api_url\": \"api.github.com\",\n",
+    "        \"personal_access_token\": \"<token>\"\n",
+    "    },\n",
+    "    \"repository\": \"<repo>\",\n",
+    "    \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\"\n",
+    "}\n",
+    "\n",
+    "issues_loader = AirbyteCDKLoader(source_class=SourceGithub, config=config, stream_name=\"issues\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = issues_loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = issues_loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To create documents in a different, pass in a record_handler function when creating the loader:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"] + \"\\n\" + (record.data[\"body\"] or \"\"), metadata=record.data)\n",
+    "\n",
+    "issues_loader = AirbyteCDKLoader(source_class=SourceGithub, config=config, stream_name=\"issues\", record_handler=handle_record)\n",
+    "\n",
+    "docs = issues_loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = issues_loader.last_state # store safely\n",
+    "\n",
+    "incremental_issue_loader = AirbyteCDKLoader(source_class=SourceGithub, config=config, stream_name=\"issues\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_issue_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/airbyte_gong.ipynb

@@ -0,0 +1,206 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte Gong"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "This loader exposes the Gong connector as a document loader, allowing you to load various Gong objects as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6847a40c",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-source-gong` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-source-gong"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Check out the [Airbyte documentation page](https://docs.airbyte.com/integrations/sources/gong/) for details about how to configure the reader.\n",
+    "The JSON schema the config object should adhere to can be found on Github: [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-gong/source_gong/spec.yaml](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-gong/source_gong/spec.yaml).\n",
+    "\n",
+    "The general shape looks like this:\n",
+    "```python\n",
+    "{\n",
+    "  \"access_key\": \"<access key name>\",\n",
+    "  \"access_key_secret\": \"<access key secret>\",\n",
+    "  \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\",\n",
+    "}\n",
+    "```\n",
+    "\n",
+    "By default all fields are stored as metadata in the documents and the text is set to an empty string. Construct the text of the document by transforming the documents returned by the reader."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteGongLoader\n",
+    "\n",
+    "config = {\n",
+    "    # your gong configuration\n",
+    "}\n",
+    "\n",
+    "loader = AirbyteGongLoader(config=config, stream_name=\"calls\") # check the documentation linked above for a list of all streams"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To process documents, create a class inheriting from the base loader and implement the `_handle_records` method yourself:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"], metadata=record.data)\n",
+    "\n",
+    "loader = AirbyteGongLoader(config=config, record_handler=handle_record, stream_name=\"calls\")\n",
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = loader.last_state # store safely\n",
+    "\n",
+    "incremental_loader = AirbyteGongLoader(config=config, stream_name=\"calls\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/airbyte_hubspot.ipynb

@@ -0,0 +1,208 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte Hubspot"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "This loader exposes the Hubspot connector as a document loader, allowing you to load various Hubspot objects as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6847a40c",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-source-hubspot` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-source-hubspot"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Check out the [Airbyte documentation page](https://docs.airbyte.com/integrations/sources/hubspot/) for details about how to configure the reader.\n",
+    "The JSON schema the config object should adhere to can be found on Github: [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-hubspot/source_hubspot/spec.yaml](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-hubspot/source_hubspot/spec.yaml).\n",
+    "\n",
+    "The general shape looks like this:\n",
+    "```python\n",
+    "{\n",
+    "  \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\",\n",
+    "  \"credentials\": {\n",
+    "    \"credentials_title\": \"Private App Credentials\",\n",
+    "    \"access_token\": \"<access token of your private app>\"\n",
+    "  }\n",
+    "}\n",
+    "```\n",
+    "\n",
+    "By default all fields are stored as metadata in the documents and the text is set to an empty string. Construct the text of the document by transforming the documents returned by the reader."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteHubspotLoader\n",
+    "\n",
+    "config = {\n",
+    "    # your hubspot configuration\n",
+    "}\n",
+    "\n",
+    "loader = AirbyteHubspotLoader(config=config, stream_name=\"products\") # check the documentation linked above for a list of all streams"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To process documents, create a class inheriting from the base loader and implement the `_handle_records` method yourself:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"], metadata=record.data)\n",
+    "\n",
+    "loader = AirbyteHubspotLoader(config=config, record_handler=handle_record, stream_name=\"products\")\n",
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = loader.last_state # store safely\n",
+    "\n",
+    "incremental_loader = AirbyteHubspotLoader(config=config, stream_name=\"products\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/airbyte_salesforce.ipynb

@@ -0,0 +1,213 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte Salesforce"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "This loader exposes the Salesforce connector as a document loader, allowing you to load various Salesforce objects as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6847a40c",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-source-salesforce` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-source-salesforce"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Check out the [Airbyte documentation page](https://docs.airbyte.com/integrations/sources/salesforce/) for details about how to configure the reader.\n",
+    "The JSON schema the config object should adhere to can be found on Github: [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-salesforce/source_salesforce/spec.yaml](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-salesforce/source_salesforce/spec.yaml).\n",
+    "\n",
+    "The general shape looks like this:\n",
+    "```python\n",
+    "{\n",
+    "  \"client_id\": \"<oauth client id>\",\n",
+    "  \"client_secret\": \"<oauth client secret>\",\n",
+    "  \"refresh_token\": \"<oauth refresh token>\",\n",
+    "  \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\",\n",
+    "  \"is_sandbox\": False, # set to True if you're using a sandbox environment\n",
+    "  \"streams_criteria\": [ # Array of filters for salesforce objects that should be loadable\n",
+    "    {\"criteria\": \"exacts\", \"value\": \"Account\"}, # Exact name of salesforce object\n",
+    "    {\"criteria\": \"starts with\", \"value\": \"Asset\"}, # Prefix of the name\n",
+    "    # Other allowed criteria: ends with, contains, starts not with, ends not with, not contains, not exacts\n",
+    "  ],\n",
+    "}\n",
+    "```\n",
+    "\n",
+    "By default all fields are stored as metadata in the documents and the text is set to an empty string. Construct the text of the document by transforming the documents returned by the reader."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteSalesforceLoader\n",
+    "\n",
+    "config = {\n",
+    "    # your salesforce configuration\n",
+    "}\n",
+    "\n",
+    "loader = AirbyteSalesforceLoader(config=config, stream_name=\"asset\") # check the documentation linked above for a list of all streams"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To create documents in a different, pass in a record_handler function when creating the loader:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"], metadata=record.data)\n",
+    "\n",
+    "loader = AirbyteSalesforceLoader(config=config, record_handler=handle_record, stream_name=\"asset\")\n",
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = loader.last_state # store safely\n",
+    "\n",
+    "incremental_loader = AirbyteSalesforceLoader(config=config, stream_name=\"asset\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/airbyte_shopify.ipynb

@@ -0,0 +1,209 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte Shopify"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "This loader exposes the Shopify connector as a document loader, allowing you to load various Shopify objects as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6847a40c",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-source-shopify` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-source-shopify"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Check out the [Airbyte documentation page](https://docs.airbyte.com/integrations/sources/shopify/) for details about how to configure the reader.\n",
+    "The JSON schema the config object should adhere to can be found on Github: [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-shopify/source_shopify/spec.json](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-shopify/source_shopify/spec.json).\n",
+    "\n",
+    "The general shape looks like this:\n",
+    "```python\n",
+    "{\n",
+    "    \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\",\n",
+    "    \"shop\": \"<name of the shop you want to retrieve documents from>\",\n",
+    "    \"credentials\": {\n",
+    "        \"auth_method\": \"api_password\",\n",
+    "        \"api_password\": \"<your api password>\"\n",
+    "    }\n",
+    "}\n",
+    "```\n",
+    "\n",
+    "By default all fields are stored as metadata in the documents and the text is set to an empty string. Construct the text of the document by transforming the documents returned by the reader."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteShopifyLoader\n",
+    "\n",
+    "config = {\n",
+    "    # your shopify configuration\n",
+    "}\n",
+    "\n",
+    "loader = AirbyteShopifyLoader(config=config, stream_name=\"orders\") # check the documentation linked above for a list of all streams"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To create documents in a different, pass in a record_handler function when creating the loader:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"], metadata=record.data)\n",
+    "\n",
+    "loader = AirbyteShopifyLoader(config=config, record_handler=handle_record, stream_name=\"orders\")\n",
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = loader.last_state # store safely\n",
+    "\n",
+    "incremental_loader = AirbyteShopifyLoader(config=config, stream_name=\"orders\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/airbyte_stripe.ipynb

@@ -0,0 +1,206 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte Stripe"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "This loader exposes the Stripe connector as a document loader, allowing you to load various Stripe objects as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6847a40c",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-source-stripe` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-source-stripe"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Check out the [Airbyte documentation page](https://docs.airbyte.com/integrations/sources/stripe/) for details about how to configure the reader.\n",
+    "The JSON schema the config object should adhere to can be found on Github: [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-stripe/source_stripe/spec.yaml](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-stripe/source_stripe/spec.yaml).\n",
+    "\n",
+    "The general shape looks like this:\n",
+    "```python\n",
+    "{\n",
+    "  \"client_secret\": \"<secret key>\",\n",
+    "  \"account_id\": \"<account id>\",\n",
+    "  \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\",\n",
+    "}\n",
+    "```\n",
+    "\n",
+    "By default all fields are stored as metadata in the documents and the text is set to an empty string. Construct the text of the document by transforming the documents returned by the reader."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteStripeLoader\n",
+    "\n",
+    "config = {\n",
+    "    # your stripe configuration\n",
+    "}\n",
+    "\n",
+    "loader = AirbyteStripeLoader(config=config, stream_name=\"invoices\") # check the documentation linked above for a list of all streams"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To create documents in a different, pass in a record_handler function when creating the loader:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"], metadata=record.data)\n",
+    "\n",
+    "loader = AirbyteStripeLoader(config=config, record_handler=handle_record, stream_name=\"invoices\")\n",
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = loader.last_state # store safely\n",
+    "\n",
+    "incremental_loader = AirbyteStripeLoader(config=config, record_handler=handle_record, stream_name=\"invoices\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/airbyte_typeform.ipynb

@@ -0,0 +1,209 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte Typeform"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "This loader exposes the Typeform connector as a document loader, allowing you to load various Typeform objects as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6847a40c",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-source-typeform` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-source-typeform"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Check out the [Airbyte documentation page](https://docs.airbyte.com/integrations/sources/typeform/) for details about how to configure the reader.\n",
+    "The JSON schema the config object should adhere to can be found on Github: [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-typeform/source_typeform/spec.json](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-typeform/source_typeform/spec.json).\n",
+    "\n",
+    "The general shape looks like this:\n",
+    "```python\n",
+    "{\n",
+    "  \"credentials\": {\n",
+    "    \"auth_type\": \"Private Token\",\n",
+    "    \"access_token\": \"<your auth token>\"\n",
+    "  },\n",
+    "  \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\",\n",
+    "  \"form_ids\": [\"<id of form to load records for>\"] # if omitted, records from all forms will be loaded\n",
+    "}\n",
+    "```\n",
+    "\n",
+    "By default all fields are stored as metadata in the documents and the text is set to an empty string. Construct the text of the document by transforming the documents returned by the reader."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteTypeformLoader\n",
+    "\n",
+    "config = {\n",
+    "    # your typeform configuration\n",
+    "}\n",
+    "\n",
+    "loader = AirbyteTypeformLoader(config=config, stream_name=\"forms\") # check the documentation linked above for a list of all streams"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To create documents in a different, pass in a record_handler function when creating the loader:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"], metadata=record.data)\n",
+    "\n",
+    "loader = AirbyteTypeformLoader(config=config, record_handler=handle_record, stream_name=\"forms\")\n",
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = loader.last_state # store safely\n",
+    "\n",
+    "incremental_loader = AirbyteTypeformLoader(config=config, record_handler=handle_record, stream_name=\"forms\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/airbyte_zendesk_support.ipynb

@@ -0,0 +1,210 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "1f3a5ebf",
+   "metadata": {},
+   "source": [
+    "# Airbyte Zendesk Support"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "35ac77b1-449b-44f7-b8f3-3494d55c286e",
+   "metadata": {},
+   "source": [
+    ">[Airbyte](https://github.com/airbytehq/airbyte) is a data integration platform for ELT pipelines from APIs, databases & files to warehouses & lakes. It has the largest catalog of ELT connectors to data warehouses and databases.\n",
+    "\n",
+    "This loader exposes the Zendesk Support connector as a document loader, allowing you to load various objects as documents."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6847a40c",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3b06fbde",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3e9dc79",
+   "metadata": {},
+   "source": [
+    "First, you need to install the `airbyte-source-zendesk-support` python package."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4d35e4e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install airbyte-source-zendesk-support"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae855210",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02208f52",
+   "metadata": {},
+   "source": [
+    "Check out the [Airbyte documentation page](https://docs.airbyte.com/integrations/sources/zendesk-support/) for details about how to configure the reader.\n",
+    "The JSON schema the config object should adhere to can be found on Github: [https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-zendesk-support/source_zendesk_support/spec.json](https://github.com/airbytehq/airbyte/blob/master/airbyte-integrations/connectors/source-zendesk-support/source_zendesk_support/spec.json).\n",
+    "\n",
+    "The general shape looks like this:\n",
+    "```python\n",
+    "{\n",
+    "  \"subdomain\": \"<your zendesk subdomain>\",\n",
+    "  \"start_date\": \"<date from which to start retrieving records from in ISO format, e.g. 2020-10-20T00:00:00Z>\",\n",
+    "  \"credentials\": {\n",
+    "    \"credentials\": \"api_token\",\n",
+    "    \"email\": \"<your email>\",\n",
+    "    \"api_token\": \"<your api token>\"\n",
+    "  }\n",
+    "}\n",
+    "```\n",
+    "\n",
+    "By default all fields are stored as metadata in the documents and the text is set to an empty string. Construct the text of the document by transforming the documents returned by the reader."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89a99e58",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.document_loaders.airbyte import AirbyteZendeskSupportLoader\n",
+    "\n",
+    "config = {\n",
+    "    # your zendesk-support configuration\n",
+    "}\n",
+    "\n",
+    "loader = AirbyteZendeskSupportLoader(config=config, stream_name=\"tickets\") # check the documentation linked above for a list of all streams"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2cea23fc",
+   "metadata": {},
+   "source": [
+    "Now you can load documents the usual way"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dae75cdb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a93dc2a",
+   "metadata": {},
+   "source": [
+    "As `load` returns a list, it will block until all documents are loaded. To have better control over this process, you can also you the `lazy_load` method which returns an iterator instead:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1782db09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_iterator = loader.lazy_load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a124086",
+   "metadata": {},
+   "source": [
+    "Keep in mind that by default the page content is empty and the metadata object contains all the information from the record. To create documents in a different, pass in a record_handler function when creating the loader:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5671395d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.docstore.document import Document\n",
+    "\n",
+    "def handle_record(record, id):\n",
+    "    return Document(page_content=record.data[\"title\"], metadata=record.data)\n",
+    "\n",
+    "loader = AirbyteZendeskSupportLoader(config=config, record_handler=handle_record, stream_name=\"tickets\")\n",
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "223eb8bc",
+   "metadata": {},
+   "source": [
+    "## Incremental loads\n",
+    "\n",
+    "Some streams allow incremental loading, this means the source keeps track of synced records and won't load them again. This is useful for sources that have a high volume of data and are updated frequently.\n",
+    "\n",
+    "To take advantage of this, store the `last_state` property of the loader and pass it in when creating the loader again. This will ensure that only new records are loaded."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7061e735",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "last_state = loader.last_state # store safely\n",
+    "\n",
+    "incremental_loader = AirbyteZendeskSupportLoader(config=config, stream_name=\"tickets\", state=last_state)\n",
+    "\n",
+    "new_docs = incremental_loader.load()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/google_cloud_storage_file.ipynb

@@ -73,13 +73,27 @@
     "loader.load()"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "41c8a46f",
+   "metadata": {},
+   "source": [
+    "If you want to use an alternative loader, you can provide a custom function, for example:"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
    "id": "eba3002d",
    "metadata": {},
    "outputs": [],
-   "source": []
+   "source": [
+    "from langchain.document_loaders import PyPDFLoader\n",
+    "def load_pdf(file_path):\n",
+    "    return PyPDFLoader(file_path)\n",
+    "\n",
+    "loader = GCSFileLoader(project_name=\"aist\", bucket=\"testing-hwc\", blob=\"fake.pdf\", loader_func=load_pdf)"
+   ]
   }
  ],
  "metadata": {

docs/extras/integrations/document_loaders/nuclia.ipynb

@@ -0,0 +1,144 @@
+{
+ "cells": [
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Nuclia Understanding API document loader\n",
+    "\n",
+    "[Nuclia](https://nuclia.com) automatically indexes your unstructured data from any internal and external source, providing optimized search results and generative answers. It can handle video and audio transcription, image content extraction, and document parsing.\n",
+    "\n",
+    "The Nuclia Understanding API supports the processing of unstructured data, including text, web pages, documents, and audio/video contents. It extracts all texts wherever they are (using speech-to-text or OCR when needed), it also extracts metadata, embedded files (like images in a PDF), and web links. If machine learning is enabled, it identifies entities, provides a summary of the content and generates embeddings for all the sentences.\n",
+    "\n",
+    "To use the Nuclia Understanding API, you need to have a Nuclia account. You can create one for free at [https://nuclia.cloud](https://nuclia.cloud), and then [create a NUA key](https://docs.nuclia.dev/docs/docs/using/understanding/intro)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install --upgrade protobuf\n",
+    "#!pip install nucliadb-protos"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "\n",
+    "os.environ[\"NUCLIA_ZONE\"] = \"<YOUR_ZONE>\"  # e.g. europe-1\n",
+    "os.environ[\"NUCLIA_NUA_KEY\"] = \"<YOUR_API_KEY>\""
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "To use the Nuclia document loader, you need to instantiate a `NucliaUnderstandingAPI` tool:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.tools.nuclia import NucliaUnderstandingAPI\n",
+    "\n",
+    "nua = NucliaUnderstandingAPI(enable_ml=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.document_loaders.nuclia import NucliaLoader\n",
+    "\n",
+    "loader = NucliaLoader(\"./interview.mp4\", nua)"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You can now call the `load` the document in a loop until you get the document."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import time\n",
+    "\n",
+    "pending = True\n",
+    "while pending:\n",
+    "    time.sleep(15)\n",
+    "    docs = loader.load()\n",
+    "    if len(docs) > 0:\n",
+    "        print(docs[0].page_content)\n",
+    "        print(docs[0].metadata)\n",
+    "        pending = False\n",
+    "    else:\n",
+    "        print(\"waiting...\")"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Retrieved information\n",
+    "\n",
+    "Nuclia returns the following information:\n",
+    "\n",
+    "- file metadata\n",
+    "- extracted text\n",
+    "- nested text (like text in an embedded image)\n",
+    "- paragraphs and sentences splitting (defined by the position of their first and last characters, plus start time and end time for a video or audio file)\n",
+    "- links\n",
+    "- a thumbnail\n",
+    "- embedded files\n",
+    "\n",
+    "Note:\n",
+    "\n",
+    "  Generated files (thumbnail, extracted embedded files, etc.) are provided as a token. You can download them with the [`/processing/download` endpoint](https://docs.nuclia.dev/docs/api#operation/Download_binary_file_processing_download_get).\n",
+    "\n",
+    "  Also at any level, if an attribute exceeds a certain size, it will be put in a downloadable file and will be replaced in the document by a file pointer. This will consist of `{\"file\": {\"uri\": \"JWT_TOKEN\"}}`. The rule is that if the size of the message is greater than 1000000 characters, the biggest parts will be moved to downloadable files. First, the compression process will target vectors. If that is not enough, it will target large field metadata, and finally it will target extracted text.\n"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "langchain",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.5"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

docs/extras/integrations/document_loaders/pubmed.ipynb

@@ -0,0 +1,139 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "3df0dcf8",
+   "metadata": {},
+   "source": [
+    "# PubMed\n",
+    "\n",
+    ">[PubMed®](https://pubmed.ncbi.nlm.nih.gov/) by `The National Center for Biotechnology Information, National Library of Medicine` comprises more than 35 million citations for biomedical literature from `MEDLINE`, life science journals, and online books. Citations may include links to full text content from `PubMed Central` and publisher web sites."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "aecaff63",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.document_loaders import PubMedLoader"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "f2f7e8d3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "loader = PubMedLoader(\"chatgpt\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "ed115aa1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs = loader.load()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "b68d3264-b893-45e4-8ab0-077b25a586dc",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "3"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(docs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "9f4626d2-068d-4aed-9ffe-ad754ad4b4cd",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'uid': '37548997',\n",
+       " 'Title': 'Performance of ChatGPT on the Situational Judgement Test-A Professional Dilemmas-Based Examination for Doctors in the United Kingdom.',\n",
+       " 'Published': '2023-08-07',\n",
+       " 'Copyright Information': '©Robin J Borchert, Charlotte R Hickman, Jack Pepys, Timothy J Sadler. Originally published in JMIR Medical Education (https://mededu.jmir.org), 07.08.2023.'}"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "docs[1].metadata"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "8000f687-b500-4cce-841b-70d6151304da",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "\"BACKGROUND: ChatGPT is a large language model that has performed well on professional examinations in the fields of medicine, law, and business. However, it is unclear how ChatGPT would perform on an examination assessing professionalism and situational judgement for doctors.\\nOBJECTIVE: We evaluated the performance of ChatGPT on the Situational Judgement Test (SJT): a national examination taken by all final-year medical students in the United Kingdom. This examination is designed to assess attributes such as communication, teamwork, patient safety, prioritization skills, professionalism, and ethics.\\nMETHODS: All questions from the UK Foundation Programme Office's (UKFPO's) 2023 SJT practice examination were inputted into ChatGPT. For each question, ChatGPT's answers and rationales were recorded and assessed on the basis of the official UK Foundation Programme Office scoring template. Questions were categorized into domains of Good Medical Practice on the basis of the domains referenced in the rationales provided in the scoring sheet. Questions without clear domain links were screened by reviewers and assigned one or multiple domains. ChatGPT's overall performance, as well as its performance across the domains of Good Medical Practice, was evaluated.\\nRESULTS: Overall, ChatGPT performed well, scoring 76% on the SJT but scoring full marks on only a few questions (9%), which may reflect possible flaws in ChatGPT's situational judgement or inconsistencies in the reasoning across questions (or both) in the examination itself. ChatGPT demonstrated consistent performance across the 4 outlined domains in Good Medical Practice for doctors.\\nCONCLUSIONS: Further research is needed to understand the potential applications of large language models, such as ChatGPT, in medical education for standardizing questions and providing consistent rationales for examinations assessing professionalism and ethics.\""
+      ]
+     },
+     "execution_count": 9,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "docs[1].page_content"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1070e571-697d-4c33-9a4f-0b2dd6909629",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.12"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_loaders/recursive_url_loader.ipynb

@@ -9,7 +9,7 @@
     "\n",
     "We may want to process load all URLs under a root directory.\n",
     "\n",
-    "For example, let's look at the [LangChain JS documentation](https://js.langchain.com/docs/).\n",
+    "For example, let's look at the [Python 3.9 Document](https://docs.python.org/3.9/).\n",
     "\n",
     "This has many interesting child pages that we may want to read in bulk.\n",
     "\n",
@@ -19,13 +19,28 @@
     " \n",
     "We do this using the `RecursiveUrlLoader`.\n",
     "\n",
-    "This also gives us the flexibility to exclude some children (e.g., the `api` directory with > 800 child pages)."
+    "This also gives us the flexibility to exclude some children, customize the extractor, and more."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1be8094f",
+   "metadata": {},
+   "source": [
+    "# Parameters\n",
+    "- url: str, the target url to crawl.\n",
+    "- exclude_dirs: Optional[str], webpage directories to exclude.\n",
+    "- use_async: Optional[bool], wether to use async requests, using async requests is usually faster in large tasks. However, async will disable the lazy loading feature(the function still works, but it is not lazy). By default, it is set to False.\n",
+    "- extractor: Optional[Callable[[str], str]], a function to extract the text of the document from the webpage, by default it returns the page as it is. It is recommended to use tools like goose3 and beautifulsoup to extract the text. By default, it just returns the page as it is.\n",
+    "- max_depth: Optional[int] = None, the maximum depth to crawl. By default, it is set to 2. If you need to crawl the whole website, set it to a number that is large enough would simply do the job.\n",
+    "- timeout: Optional[int] = None, the timeout for each request, in the unit of seconds. By default, it is set to 10.\n",
+    "- prevent_outside: Optional[bool] = None, whether to prevent crawling outside the root url. By default, it is set to True."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
-   "id": "2e3532b2",
+   "execution_count": null,
+   "id": "23c18539",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -42,13 +57,15 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
-   "id": "d69e5620",
+   "execution_count": null,
+   "id": "55394afe",
    "metadata": {},
    "outputs": [],
    "source": [
-    "url = \"https://js.langchain.com/docs/modules/memory/examples/\"\n",
-    "loader = RecursiveUrlLoader(url=url)\n",
+    "from bs4 import BeautifulSoup as Soup\n",
+    "\n",
+    "url = \"https://docs.python.org/3.9/\"\n",
+    "loader = RecursiveUrlLoader(url=url, max_depth=2, extractor=lambda x: Soup(x, \"html.parser\").text)\n",
     "docs = loader.load()"
    ]
   },
@@ -61,7 +78,7 @@
     {
      "data": {
       "text/plain": [
-       "12"
+       "'\\n\\n\\n\\n\\nPython Frequently Asked Questions — Python 3.'"
       ]
      },
      "execution_count": 3,
@@ -70,19 +87,21 @@
     }
    ],
    "source": [
-    "len(docs)"
+    "docs[0].page_content[:50]"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 4,
-   "id": "89355b7c",
+   "id": "13bd7e16",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "'\\n\\n\\n\\n\\nBuffer Window Memory | 🦜️🔗 Langchain\\n\\n\\n\\n\\n\\nSki'"
+       "{'source': 'https://docs.python.org/3.9/library/index.html',\n",
+       " 'title': 'The Python Standard Library — Python 3.9.17 documentation',\n",
+       " 'language': None}"
       ]
      },
      "execution_count": 4,
@@ -91,137 +110,48 @@
     }
    ],
    "source": [
-    "docs[0].page_content[:50]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "id": "13bd7e16",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "{'source': 'https://js.langchain.com/docs/modules/memory/examples/buffer_window_memory',\n",
-       " 'title': 'Buffer Window Memory | 🦜️🔗 Langchain',\n",
-       " 'description': 'BufferWindowMemory keeps track of the back-and-forths in conversation, and then uses a window of size k to surface the last k back-and-forths to use as memory.',\n",
-       " 'language': 'en'}"
-      ]
-     },
-     "execution_count": 5,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "docs[0].metadata"
+    "docs[-1].metadata"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "40fc13ef",
+   "id": "5866e5a6",
    "metadata": {},
    "source": [
-    "Now, let's try a more extensive example, the `docs` root dir.\n",
-    "\n",
-    "We will skip everything under `api`.\n",
-    "\n",
-    "For this, we can `lazy_load`  each page as we crawl the tree, using `WebBaseLoader` to load each as we go."
+    "However, since it's hard to perform a perfect filter, you may still see some irrelevant results in the results. You can perform a filter on the returned documents by yourself, if it's needed. Most of the time, the returned results are good enough."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4ec8ecef",
+   "metadata": {},
+   "source": [
+    "Testing on LangChain docs."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
-   "id": "5c938b9f",
+   "execution_count": 2,
+   "id": "349b5598",
    "metadata": {},
-   "outputs": [],
-   "source": [
-    "url = \"https://js.langchain.com/docs/\"\n",
-    "exclude_dirs = [\"https://js.langchain.com/docs/api/\"]\n",
-    "loader = RecursiveUrlLoader(url=url, exclude_dirs=exclude_dirs)\n",
-    "# Lazy load each\n",
-    "docs = [print(doc) or doc for doc in loader.lazy_load()]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "id": "30ff61d3",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Load all pages\n",
-    "docs = loader.load()"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "id": "457e30f3",
-   "metadata": {
-    "scrolled": true
-   },
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "188"
+       "8"
       ]
      },
-     "execution_count": 8,
+     "execution_count": 2,
      "metadata": {},
      "output_type": "execute_result"
     }
    ],
    "source": [
+    "url = \"https://js.langchain.com/docs/modules/memory/integrations/\"\n",
+    "loader = RecursiveUrlLoader(url=url)\n",
+    "docs = loader.load()\n",
     "len(docs)"
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "id": "bca80b4a",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "'\\n\\n\\n\\n\\nAgent Simulations | 🦜️🔗 Langchain\\n\\n\\n\\n\\n\\nSkip t'"
-      ]
-     },
-     "execution_count": 9,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "docs[0].page_content[:50]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 10,
-   "id": "df97cf22",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "{'source': 'https://js.langchain.com/docs/use_cases/agent_simulations/',\n",
-       " 'title': 'Agent Simulations | 🦜️🔗 Langchain',\n",
-       " 'description': 'Agent simulations involve taking multiple agents and having them interact with each other.',\n",
-       " 'language': 'en'}"
-      ]
-     },
-     "execution_count": 10,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "docs[0].metadata"
-   ]
   }
  ],
  "metadata": {

docs/extras/integrations/document_loaders/tensorflow_datasets.ipynb

@@ -0,0 +1,320 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "bda1f3f5",
+   "metadata": {},
+   "source": [
+    "# TensorFlow Datasets\n",
+    "\n",
+    ">[TensorFlow Datasets](https://www.tensorflow.org/datasets) is a collection of datasets ready to use, with TensorFlow or other Python ML frameworks, such as Jax. All datasets are exposed as [tf.data.Datasets](https://www.tensorflow.org/api_docs/python/tf/data/Dataset), enabling easy-to-use and high-performance input pipelines. To get started see the [guide](https://www.tensorflow.org/datasets/overview) and the [list of datasets](https://www.tensorflow.org/datasets/catalog/overview#all_datasets).\n",
+    "\n",
+    "This notebook shows how to load `TensorFlow Datasets` into a Document format that we can use downstream."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1b7a1eef-7bf7-4e7d-8bfc-c4e27c9488cb",
+   "metadata": {},
+   "source": [
+    "## Installation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2abd5578-aa3d-46b9-99af-8b262f0b3df8",
+   "metadata": {},
+   "source": [
+    "You need to install `tensorflow` and `tensorflow-datasets` python packages."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2e589036-351e-4c63-b734-c9a05fadb880",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!pip install tensorflow"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b674aaea-ed3a-4541-8414-260a8f67f623",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "!pip install tensorflow-datasets"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "95f05e1c-195e-4e2b-ae8e-8d6637f15be6",
+   "metadata": {},
+   "source": [
+    "## Example"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e66e211e-9419-4dbb-b3cd-afc3cf984305",
+   "metadata": {},
+   "source": [
+    "As an example, we use the [`mlqa/en` dataset](https://www.tensorflow.org/datasets/catalog/mlqa#mlqaen).\n",
+    "\n",
+    ">`MLQA` (`Multilingual Question Answering Dataset`) is a benchmark dataset for evaluating multilingual question answering performance. The dataset consists of 7 languages: Arabic, German, Spanish, English, Hindi, Vietnamese, Chinese.\n",
+    ">\n",
+    ">- Homepage: https://github.com/facebookresearch/MLQA\n",
+    ">- Source code: `tfds.datasets.mlqa.Builder`\n",
+    ">- Download size: 72.21 MiB\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8968d645-c81c-4e3b-82bc-a3cbb5ddd93a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Feature structure of `mlqa/en` dataset:\n",
+    "\n",
+    "FeaturesDict({\n",
+    "    'answers': Sequence({\n",
+    "        'answer_start': int32,\n",
+    "        'text': Text(shape=(), dtype=string),\n",
+    "    }),\n",
+    "    'context': Text(shape=(), dtype=string),\n",
+    "    'id': string,\n",
+    "    'question': Text(shape=(), dtype=string),\n",
+    "    'title': Text(shape=(), dtype=string),\n",
+    "})"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "30fcaba5-cc9b-4a0e-a8f4-c047018451c2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import tensorflow as tf\n",
+    "import tensorflow_datasets as tfds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 78,
+   "id": "e307dd67-029e-4ee3-a65f-e085c09b0b8b",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<_TakeDataset element_spec={'answers': {'answer_start': TensorSpec(shape=(None,), dtype=tf.int32, name=None), 'text': TensorSpec(shape=(None,), dtype=tf.string, name=None)}, 'context': TensorSpec(shape=(), dtype=tf.string, name=None), 'id': TensorSpec(shape=(), dtype=tf.string, name=None), 'question': TensorSpec(shape=(), dtype=tf.string, name=None), 'title': TensorSpec(shape=(), dtype=tf.string, name=None)}>"
+      ]
+     },
+     "execution_count": 78,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# try directly access this dataset:\n",
+    "ds = tfds.load('mlqa/en', split='test')\n",
+    "ds = ds.take(1)  # Only take a single example\n",
+    "ds"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5c9c4b08-d94f-4b53-add0-93769811644e",
+   "metadata": {},
+   "source": [
+    "Now we have to create a custom function to convert dataset sample into a Document.\n",
+    "\n",
+    "This is a requirement. There is no standard format for the TF datasets that's why we need to make a custom transformation function.\n",
+    "\n",
+    "Let's use `context` field as the `Document.page_content` and place other fields in the `Document.metadata`.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 72,
+   "id": "78844113-f8d8-48a8-8105-685280b6cfa5",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "page_content='After completing the journey around South America, on 23 February 2006, Queen Mary 2 met her namesake, the original RMS Queen Mary, which is permanently docked at Long Beach, California. Escorted by a flotilla of smaller ships, the two Queens exchanged a \"whistle salute\" which was heard throughout the city of Long Beach. Queen Mary 2 met the other serving Cunard liners Queen Victoria and Queen Elizabeth 2 on 13 January 2008 near the Statue of Liberty in New York City harbour, with a celebratory fireworks display; Queen Elizabeth 2 and Queen Victoria made a tandem crossing of the Atlantic for the meeting. This marked the first time three Cunard Queens have been present in the same location. Cunard stated this would be the last time these three ships would ever meet, due to Queen Elizabeth 2\\'s impending retirement from service in late 2008. However this would prove not to be the case, as the three Queens met in Southampton on 22 April 2008. Queen Mary 2 rendezvoused with Queen Elizabeth 2  in Dubai on Saturday 21 March 2009, after the latter ship\\'s retirement, while both ships were berthed at Port Rashid. With the withdrawal of Queen Elizabeth 2 from Cunard\\'s fleet and its docking in Dubai, Queen Mary 2 became the only ocean liner left in active passenger service.' metadata={'id': '5116f7cccdbf614d60bcd23498274ffd7b1e4ec7', 'title': 'RMS Queen Mary 2', 'question': 'What year did Queen Mary 2 complete her journey around South America?', 'answer': '2006'}\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2023-08-03 14:27:08.482983: W tensorflow/core/kernels/data/cache_dataset_ops.cc:854] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.\n"
+     ]
+    }
+   ],
+   "source": [
+    "def decode_to_str(item: tf.Tensor) -> str:\n",
+    "    return item.numpy().decode('utf-8')\n",
+    "\n",
+    "def mlqaen_example_to_document(example: dict) -> Document:\n",
+    "    return Document(\n",
+    "        page_content=decode_to_str(example[\"context\"]),\n",
+    "        metadata={\n",
+    "            \"id\": decode_to_str(example[\"id\"]),\n",
+    "            \"title\": decode_to_str(example[\"title\"]),\n",
+    "            \"question\": decode_to_str(example[\"question\"]),\n",
+    "            \"answer\": decode_to_str(example[\"answers\"][\"text\"][0]),\n",
+    "        },\n",
+    "    )\n",
+    "    \n",
+    "    \n",
+    "for example in ds:  \n",
+    "    doc = mlqaen_example_to_document(example)\n",
+    "    print(doc)\n",
+    "    break"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 73,
+   "id": "2d43c834-5145-4793-9558-8e301ccaf3b4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.schema import Document\n",
+    "from langchain.document_loaders import TensorflowDatasetLoader\n",
+    "\n",
+    "loader = TensorflowDatasetLoader(\n",
+    "        dataset_name=\"mlqa/en\",\n",
+    "        split_name=\"test\",\n",
+    "        load_max_docs=3,\n",
+    "        sample_to_document_function=mlqaen_example_to_document,\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e29b954c-1407-4797-ae21-6ba8937156be",
+   "metadata": {},
+   "source": [
+    "`TensorflowDatasetLoader` has these parameters:\n",
+    "- `dataset_name`: the name of the dataset to load\n",
+    "- `split_name`: the name of the split to load. Defaults to \"train\".\n",
+    "- `load_max_docs`: a limit to the number of loaded documents. Defaults to 100.\n",
+    "- `sample_to_document_function`: a function that converts a dataset sample to a Document\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 74,
+   "id": "700e4ef2",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2023-08-03 14:27:22.998964: W tensorflow/core/kernels/data/cache_dataset_ops.cc:854] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "3"
+      ]
+     },
+     "execution_count": 74,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "docs = loader.load()\n",
+    "len(docs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 76,
+   "id": "9138940a-e9fe-4145-83e8-77589b5272c9",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'After completing the journey around South America, on 23 February 2006, Queen Mary 2 met her namesake, the original RMS Queen Mary, which is permanently docked at Long Beach, California. Escorted by a flotilla of smaller ships, the two Queens exchanged a \"whistle salute\" which was heard throughout the city of Long Beach. Queen Mary 2 met the other serving Cunard liners Queen Victoria and Queen Elizabeth 2 on 13 January 2008 near the Statue of Liberty in New York City harbour, with a celebratory fireworks display; Queen Elizabeth 2 and Queen Victoria made a tandem crossing of the Atlantic for the meeting. This marked the first time three Cunard Queens have been present in the same location. Cunard stated this would be the last time these three ships would ever meet, due to Queen Elizabeth 2\\'s impending retirement from service in late 2008. However this would prove not to be the case, as the three Queens met in Southampton on 22 April 2008. Queen Mary 2 rendezvoused with Queen Elizabeth 2  in Dubai on Saturday 21 March 2009, after the latter ship\\'s retirement, while both ships were berthed at Port Rashid. With the withdrawal of Queen Elizabeth 2 from Cunard\\'s fleet and its docking in Dubai, Queen Mary 2 became the only ocean liner left in active passenger service.'"
+      ]
+     },
+     "execution_count": 76,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "docs[0].page_content"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 77,
+   "id": "2f7f7832-fe4d-4a58-892d-bb987cdbed0b",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'id': '5116f7cccdbf614d60bcd23498274ffd7b1e4ec7',\n",
+       " 'title': 'RMS Queen Mary 2',\n",
+       " 'question': 'What year did Queen Mary 2 complete her journey around South America?',\n",
+       " 'answer': '2006'}"
+      ]
+     },
+     "execution_count": 77,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "docs[0].metadata"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "125d073c-4f4f-4ae6-a0c7-9e9db3cc8d69",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.12"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/document_transformers/nuclia_transformer.ipynb

@@ -0,0 +1,103 @@
+{
+ "cells": [
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Nuclia Understanding API document transformer\n",
+    "\n",
+    "[Nuclia](https://nuclia.com) automatically indexes your unstructured data from any internal and external source, providing optimized search results and generative answers. It can handle video and audio transcription, image content extraction, and document parsing.\n",
+    "\n",
+    "The Nuclia Understanding API document transformer splits text into paragraphs and sentences, identifies entities, provides a summary of the text and generates embeddings for all the sentences.\n",
+    "\n",
+    "To use the Nuclia Understanding API, you need to have a Nuclia account. You can create one for free at [https://nuclia.cloud](https://nuclia.cloud), and then [create a NUA key](https://docs.nuclia.dev/docs/docs/using/understanding/intro).\n",
+    "\n",
+    "from langchain.document_transformers.nuclia_text_transform import NucliaTextTransformer"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install --upgrade protobuf\n",
+    "#!pip install nucliadb-protos"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "\n",
+    "os.environ[\"NUCLIA_ZONE\"] = \"<YOUR_ZONE>\"  # e.g. europe-1\n",
+    "os.environ[\"NUCLIA_NUA_KEY\"] = \"<YOUR_API_KEY>\""
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "To use the Nuclia document transformer, you need to instantiate a `NucliaUnderstandingAPI` tool with `enable_ml` set to `True`:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.tools.nuclia import NucliaUnderstandingAPI\n",
+    "\n",
+    "nua = NucliaUnderstandingAPI(enable_ml=True)"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The Nuclia document transformer must be called in async mode, so you need to use the `atransform_documents` method:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import asyncio\n",
+    "\n",
+    "from langchain.document_transformers.nuclia_text_transform import NucliaTextTransformer\n",
+    "from langchain.schema.document import Document\n",
+    "\n",
+    "\n",
+    "async def process():\n",
+    "    documents = [\n",
+    "        Document(page_content=\"<TEXT 1>\", metadata={}),\n",
+    "        Document(page_content=\"<TEXT 2>\", metadata={}),\n",
+    "        Document(page_content=\"<TEXT 3>\", metadata={}),\n",
+    "    ]\n",
+    "    nuclia_transformer = NucliaTextTransformer(nua)\n",
+    "    transformed_documents = await nuclia_transformer.atransform_documents(documents)\n",
+    "    print(transformed_documents)\n",
+    "\n",
+    "\n",
+    "asyncio.run(process())"
+   ]
+  }
+ ],
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

docs/extras/integrations/llms/Fireworks.ipynb

@@ -7,7 +7,7 @@
    "source": [
     "# Fireworks\n",
     "\n",
-    ">[Fireworks](https://www.fireworks.ai/) is an AI startup focused on accelerating product development on generative AI by creating an innovative AI experiment and production platform. \n",
+    ">[Fireworks](https://app.fireworks.ai/) accelerates product development on generative AI by creating an innovative AI experiment and production platform. \n",
     "\n",
     "This example goes over how to use LangChain to interact with `Fireworks` models."
    ]
@@ -37,7 +37,7 @@
     "\n",
     "Contact Fireworks AI for the an API Key to access our models\n",
     "\n",
-    "Set up your model using a model id. If the model is not set, the default model is fireworks-llama-v2-13b-chat."
+    "Set up your model using a model id. If the model is not set, the default model is fireworks-llama-v2-7b-chat."
    ]
   },
   {
@@ -49,7 +49,7 @@
    "source": [
     "# Initialize a Fireworks LLM\n",
     "os.environ['FIREWORKS_API_KEY'] = \"\" #change this to your own API KEY\n",
-    "llm = Fireworks(model_id=\"fireworks-llama-v2-13b-chat\")"
+    "llm = Fireworks(model_id=\"accounts/fireworks/models/fireworks-llama-v2-13b-chat\")"
    ]
   },
   {
@@ -73,11 +73,10 @@
     "You can use the LLMs to call the model for specified prompt(s). \n",
     "\n",
     "Current Specified Models: \n",
-    "* fireworks-falcon-7b, fireworks-falcon-40b-w8a16\n",
-    "* fireworks-guanaco-30b, fireworks-guanaco-33b-w8a16\n",
-    "* fireworks-llama-7b, fireworks-llama-13b, fireworks-llama-30b-w8a16\n",
-    "* fireworks-llama-v2-13b, fireworks-llama-v2-13b-chat, fireworks-llama-v2-13b-w8a16, fireworks-llama-v2-13b-chat-w8a16\n",
-    "* fireworks-llama-v2-7b, fireworks-llama-v2-7b-chat, fireworks-llama-v2-7b-w8a16, fireworks-llama-v2-7b-chat-w8a16"
+    "* accounts/fireworks/models/fireworks-falcon-7b, accounts/fireworks/models/fireworks-falcon-40b-w8a16\n",
+    "* accounts/fireworks/models/fireworks-starcoder-1b-w8a16-1gpu, accounts/fireworks/models/fireworks-starcoder-3b-w8a16-1gpu, accounts/fireworks/models/fireworks-starcoder-7b-w8a16-1gpu, accounts/fireworks/models/fireworks-starcoder-16b-w8a16 \n",
+    "* accounts/fireworks/models/fireworks-llama-v2-13b, accounts/fireworks/models/fireworks-llama-v2-13b-chat, accounts/fireworks/models/fireworks-llama-v2-13b-w8a16, accounts/fireworks/models/fireworks-llama-v2-13b-chat-w8a16\n",
+    "* accounts/fireworks/models/fireworks-llama-v2-7b, accounts/fireworks/models/fireworks-llama-v2-7b-chat, accounts/fireworks/models/fireworks-llama-v2-7b-w8a16, accounts/fireworks/models/fireworks-llama-v2-7b-chat-w8a16, accounts/fireworks/models/fireworks-llama-v2-70b-chat-4gpu"
    ]
   },
   {
@@ -147,7 +146,7 @@
    ],
    "source": [
     "#setting parameters: model_id, temperature, max_tokens, top_p\n",
-    "llm = Fireworks(model_id=\"fireworks-llama-v2-13b-chat\", temperature=0.7, max_tokens=15, top_p=1.0)\n",
+    "llm = Fireworks(model_id=\"accounts/fireworks/models/fireworks-llama-v2-13b-chat\", temperature=0.7, max_tokens=15, top_p=1.0)\n",
     "print(llm(\"What's the weather like in Kansas City in December?\"))"
    ]
   },

docs/extras/integrations/llms/llamacpp.ipynb

@@ -4,12 +4,12 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Llama-cpp\n",
+    "# Llama.cpp\n",
     "\n",
-    "[llama-cpp](https://github.com/abetlen/llama-cpp-python) is a Python binding for [llama.cpp](https://github.com/ggerganov/llama.cpp). \n",
+    "[llama-cpp-python](https://github.com/abetlen/llama-cpp-python) is a Python binding for [llama.cpp](https://github.com/ggerganov/llama.cpp). \n",
     "It supports [several LLMs](https://github.com/ggerganov/llama.cpp).\n",
     "\n",
-    "This notebook goes over how to run `llama-cpp` within LangChain."
+    "This notebook goes over how to run `llama-cpp-python` within LangChain."
    ]
   },
   {
@@ -18,7 +18,7 @@
    "source": [
     "## Installation\n",
     "\n",
-    "There is a bunch of options how to install the llama-cpp package: \n",
+    "There are different options on how to install the llama-cpp package: \n",
     "- only CPU usage\n",
     "- CPU + GPU (using one of many BLAS backends)\n",
     "- Metal GPU (MacOS with Apple Silicon Chip) \n",
@@ -61,7 +61,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "**IMPORTANT**: If you have already installed a cpu only version of the package, you need to reinstall it from scratch: consider the following command: "
+    "**IMPORTANT**: If you have already installed the CPU only version of the package, you need to reinstall it from scratch. Consider the following command: "
    ]
   },
   {
@@ -79,7 +79,7 @@
    "source": [
     "### Installation with Metal\n",
     "\n",
-    "`lama.cpp` supports Apple silicon first-class citizen - optimized via ARM NEON, Accelerate and Metal frameworks. Use the `FORCE_CMAKE=1` environment variable to force the use of cmake and install the pip package for the Metal support ([source](https://github.com/abetlen/llama-cpp-python/blob/main/docs/install/macos.md)).\n",
+    "`llama.cpp` supports Apple silicon first-class citizen - optimized via ARM NEON, Accelerate and Metal frameworks. Use the `FORCE_CMAKE=1` environment variable to force the use of cmake and install the pip package for the Metal support ([source](https://github.com/abetlen/llama-cpp-python/blob/main/docs/install/macos.md)).\n",
     "\n",
     "Example installation with Metal Support:"
    ]
@@ -143,7 +143,7 @@
     "\n",
     "#### Compiling and installing\n",
     "\n",
-    "In the same command prompt (anaconda prompt) you set the variables, you can cd into `llama-cpp-python` directory and run the following commands.\n",
+    "In the same command prompt (anaconda prompt) you set the variables, you can `cd` into `llama-cpp-python` directory and run the following commands.\n",
     "\n",
     "```\n",
     "python setup.py clean\n",
@@ -164,7 +164,9 @@
    "source": [
     "Make sure you are following all instructions to [install all necessary model files](https://github.com/ggerganov/llama.cpp).\n",
     "\n",
-    "You don't need an `API_TOKEN`!"
+    "You don't need an `API_TOKEN` as you will run the LLM locally.\n",
+    "\n",
+    "It is worth understanding which models are suitable to be used on the desired machine."
    ]
   },
   {
@@ -227,7 +229,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "`Llama-v2`"
+    "Example using a LLaMA 2 7B model"
    ]
   },
   {
@@ -304,7 +306,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "`Llama-v1`"
+    "Example using a LLaMA v1 model"
    ]
   },
   {
@@ -381,7 +383,7 @@
    "source": [
     "### GPU\n",
     "\n",
-    "If the installation with BLAS backend was correct, you will see an `BLAS = 1` indicator in model properties.\n",
+    "If the installation with BLAS backend was correct, you will see a `BLAS = 1` indicator in model properties.\n",
     "\n",
     "Two of the most important parameters for use with GPU are:\n",
     "\n",
@@ -473,22 +475,15 @@
     "llm_chain.run(question)"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
     "### Metal\n",
     "\n",
-    "If the installation with Metal was correct, you will see an `NEON = 1` indicator in model properties.\n",
+    "If the installation with Metal was correct, you will see a `NEON = 1` indicator in model properties.\n",
     "\n",
-    "Two of the most important parameters for use with GPU are:\n",
+    "Two of the most important GPU parameters are:\n",
     "\n",
     "- `n_gpu_layers` - determines how many layers of the model are offloaded to your Metal GPU, in the most case, set it to `1` is enough for Metal\n",
     "- `n_batch` - how many tokens are processed in parallel, default is 8, set to bigger number.\n",
@@ -522,7 +517,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "The rest are almost same as GPU, the console log will show the following log to indicate the Metal was enable properly.\n",
+    "The console log will show the following log to indicate Metal was enable properly.\n",
     "\n",
     "```\n",
     "ggml_metal_init: allocating\n",
@@ -530,7 +525,9 @@
     "...\n",
     "```\n",
     "\n",
-    "You also could check the `Activity Monitor` by watching the % GPU of the process, the % CPU will drop dramatically after turn on `n_gpu_layers=1`. Also for the first time call LLM, the performance might be slow due to the model compilation in Metal GPU."
+    "You also could check `Activity Monitor` by watching the GPU usage of the process, the CPU usage will drop dramatically after turn on `n_gpu_layers=1`. \n",
+    "\n",
+    "For the first call to the LLM, the performance may be slow due to the model compilation in Metal GPU."
    ]
   }
  ],

docs/extras/integrations/llms/ollama.ipynb

@@ -0,0 +1,340 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Ollama\n",
+    "\n",
+    "[Ollama](https://ollama.ai/) allows you to run open-source large language models, such as Llama 2, locally.\n",
+    "\n",
+    "Ollama bundles model weights, configuration, and data into a single package, defined by a Modelfile. \n",
+    "\n",
+    "It optimizes setup and configuration details, including GPU usage.\n",
+    "\n",
+    "For a complete list of supported models and model variants, see the [Ollama model library](https://github.com/jmorganca/ollama#model-library).\n",
+    "\n",
+    "## Setup\n",
+    "\n",
+    "First, follow [these instructions](https://github.com/jmorganca/ollama) to set up and run a local Ollama instance:\n",
+    "\n",
+    "* [Download](https://ollama.ai/download)\n",
+    "* Fetch a model, e.g., `Llama-7b`: `ollama pull llama2`\n",
+    "* Run `ollama run llama2`\n",
+    "\n",
+    "\n",
+    "## Usage\n",
+    "\n",
+    "You can see a full list of supported parameters on the [API reference page](https://api.python.langchain.com/en/latest/llms/langchain.llms.ollama.Ollama.html)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 38,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.llms import Ollama\n",
+    "from langchain.callbacks.manager import CallbackManager\n",
+    "from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler                                  \n",
+    "llm = Ollama(base_url=\"http://localhost:11434\", \n",
+    "             model=\"llama2\", \n",
+    "             callback_manager = CallbackManager([StreamingStdOutCallbackHandler()]))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "With `StreamingStdOutCallbackHandler`, you will see tokens streamed."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 40,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Great! The history of Artificial Intelligence (AI) is a fascinating and complex topic that spans several decades. Here's a brief overview:\n",
+      "\n",
+      "1. Early Years (1950s-1960s): The term \"Artificial Intelligence\" was coined in 1956 by computer scientist John McCarthy. However, the concept of AI dates back to ancient Greece, where mythical creatures like Talos and Hephaestus were created to perform tasks without any human intervention. In the 1950s and 1960s, researchers began exploring ways to replicate human intelligence using computers, leading to the development of simple AI programs like ELIZA (1966) and PARRY (1972).\n",
+      "2. Rule-Based Systems (1970s-1980s): As computing power increased, researchers developed rule-based systems, such as Mycin (1976), which could diagnose medical conditions based on a set of rules. This period also saw the rise of expert systems, like EDICT (1985), which mimicked human experts in specific domains.\n",
+      "3. Machine Learning (1990s-2000s): With the advent of big data and machine learning algorithms, AI evolved to include neural networks, decision trees, and other techniques for training models on large datasets. This led to the development of applications like speech recognition (e.g., Siri, Alexa), image recognition (e.g., Google Image Search), and natural language processing (e.g., chatbots).\n",
+      "4. Deep Learning (2010s-present): The rise of deep learning techniques, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), has enabled AI to perform complex tasks like image and speech recognition, natural language processing, and even autonomous driving. Companies like Google, Facebook, and Baidu have invested heavily in deep learning research, leading to breakthroughs in areas like facial recognition, object detection, and machine translation.\n",
+      "5. Current Trends (present-future): AI is currently being applied to various industries, including healthcare, finance, education, and entertainment. With the growth of cloud computing, edge AI, and autonomous systems, we can expect to see more sophisticated AI applications in the near future. However, there are also concerns about the ethical implications of AI, such as data privacy, algorithmic bias, and job displacement.\n",
+      "\n",
+      "Remember, AI has a long history, and its development is an ongoing process. As technology advances, we can expect to see even more innovative applications of AI in various fields."
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "'\\nGreat! The history of Artificial Intelligence (AI) is a fascinating and complex topic that spans several decades. Here\\'s a brief overview:\\n\\n1. Early Years (1950s-1960s): The term \"Artificial Intelligence\" was coined in 1956 by computer scientist John McCarthy. However, the concept of AI dates back to ancient Greece, where mythical creatures like Talos and Hephaestus were created to perform tasks without any human intervention. In the 1950s and 1960s, researchers began exploring ways to replicate human intelligence using computers, leading to the development of simple AI programs like ELIZA (1966) and PARRY (1972).\\n2. Rule-Based Systems (1970s-1980s): As computing power increased, researchers developed rule-based systems, such as Mycin (1976), which could diagnose medical conditions based on a set of rules. This period also saw the rise of expert systems, like EDICT (1985), which mimicked human experts in specific domains.\\n3. Machine Learning (1990s-2000s): With the advent of big data and machine learning algorithms, AI evolved to include neural networks, decision trees, and other techniques for training models on large datasets. This led to the development of applications like speech recognition (e.g., Siri, Alexa), image recognition (e.g., Google Image Search), and natural language processing (e.g., chatbots).\\n4. Deep Learning (2010s-present): The rise of deep learning techniques, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), has enabled AI to perform complex tasks like image and speech recognition, natural language processing, and even autonomous driving. Companies like Google, Facebook, and Baidu have invested heavily in deep learning research, leading to breakthroughs in areas like facial recognition, object detection, and machine translation.\\n5. Current Trends (present-future): AI is currently being applied to various industries, including healthcare, finance, education, and entertainment. With the growth of cloud computing, edge AI, and autonomous systems, we can expect to see more sophisticated AI applications in the near future. However, there are also concerns about the ethical implications of AI, such as data privacy, algorithmic bias, and job displacement.\\n\\nRemember, AI has a long history, and its development is an ongoing process. As technology advances, we can expect to see even more innovative applications of AI in various fields.'"
+      ]
+     },
+     "execution_count": 40,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "llm(\"Tell me about the history of AI\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## RAG\n",
+    "\n",
+    "We can use Olama with RAG, [just as shown here](https://python.langchain.com/docs/use_cases/question_answering/how_to/local_retrieval_qa).\n",
+    "\n",
+    "Let's use the 13b model:\n",
+    "\n",
+    "```\n",
+    "ollama pull llama2:13b\n",
+    "ollama run llama2:13b \n",
+    "```\n",
+    "\n",
+    "Let's also use local embeddings from `GPT4AllEmbeddings` and `Chroma`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "! pip install gpt4all chromadb"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 60,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.document_loaders import WebBaseLoader\n",
+    "loader = WebBaseLoader(\"https://lilianweng.github.io/posts/2023-06-23-agent/\")\n",
+    "data = loader.load()\n",
+    "\n",
+    "from langchain.text_splitter import RecursiveCharacterTextSplitter\n",
+    "text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)\n",
+    "all_splits = text_splitter.split_documents(data)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 61,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found model file at  /Users/rlm/.cache/gpt4all/ggml-all-MiniLM-L6-v2-f16.bin\n"
+     ]
+    }
+   ],
+   "source": [
+    "from langchain.vectorstores import Chroma\n",
+    "from langchain.embeddings import GPT4AllEmbeddings\n",
+    "\n",
+    "vectorstore = Chroma.from_documents(documents=all_splits, embedding=GPT4AllEmbeddings())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 62,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "4"
+      ]
+     },
+     "execution_count": 62,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "question = \"What are the approaches to Task Decomposition?\"\n",
+    "docs = vectorstore.similarity_search(question)\n",
+    "len(docs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain import PromptTemplate\n",
+    "\n",
+    "# Prompt\n",
+    "template = \"\"\"Use the following pieces of context to answer the question at the end. \n",
+    "If you don't know the answer, just say that you don't know, don't try to make up an answer. \n",
+    "Use three sentences maximum and keep the answer as concise as possible. \n",
+    "{context}\n",
+    "Question: {question}\n",
+    "Helpful Answer:\"\"\"\n",
+    "QA_CHAIN_PROMPT = PromptTemplate(\n",
+    "    input_variables=[\"context\", \"question\"],\n",
+    "    template=template,\n",
+    ")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 69,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# LLM\n",
+    "from langchain.llms import Ollama\n",
+    "from langchain.callbacks.manager import CallbackManager\n",
+    "from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler\n",
+    "llm = Ollama(base_url=\"http://localhost:11434\",\n",
+    "             model=\"llama2\",\n",
+    "             verbose=True,\n",
+    "             callback_manager=CallbackManager([StreamingStdOutCallbackHandler()]))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 66,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# QA chain\n",
+    "from langchain.chains import RetrievalQA\n",
+    "qa_chain = RetrievalQA.from_chain_type(\n",
+    "    llm,\n",
+    "    retriever=vectorstore.as_retriever(),\n",
+    "    chain_type_kwargs={\"prompt\": QA_CHAIN_PROMPT},\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 70,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Task decomposition can be approached in different ways for AI agents, including:\n",
+      "\n",
+      "1. Using simple prompts like \"Steps for XYZ.\" or \"What are the subgoals for achieving XYZ?\" to guide the LLM.\n",
+      "2. Providing task-specific instructions, such as \"Write a story outline\" for writing a novel.\n",
+      "3. Utilizing human inputs to help the AI agent understand the task and break it down into smaller steps."
+     ]
+    }
+   ],
+   "source": [
+    "question = \"What are the various approaches to Task Decomposition for AI Agents?\"\n",
+    "result = qa_chain({\"query\": question})"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You can also get logging for tokens."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 56,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Task decomposition can be approached in three ways: (1) using simple prompting like \"Steps for XYZ.\\n1.\", \"What are the subgoals for achieving XYZ?\", (2) by using task-specific instructions, or (3) with human inputs.{'model': 'llama2', 'created_at': '2023-08-08T04:01:09.005367Z', 'done': True, 'context': [1, 29871, 1, 13, 9314, 14816, 29903, 6778, 13, 13, 3492, 526, 263, 8444, 29892, 3390, 1319, 322, 15993, 20255, 29889, 29849, 1234, 408, 1371, 3730, 408, 1950, 29892, 1550, 1641, 9109, 29889, 3575, 6089, 881, 451, 3160, 738, 10311, 1319, 29892, 443, 621, 936, 29892, 11021, 391, 29892, 7916, 391, 29892, 304, 27375, 29892, 18215, 29892, 470, 27302, 2793, 29889, 3529, 9801, 393, 596, 20890, 526, 5374, 635, 443, 5365, 1463, 322, 6374, 297, 5469, 29889, 13, 13, 3644, 263, 1139, 947, 451, 1207, 738, 4060, 29892, 470, 338, 451, 2114, 1474, 16165, 261, 296, 29892, 5649, 2020, 2012, 310, 22862, 1554, 451, 1959, 29889, 960, 366, 1016, 29915, 29873, 1073, 278, 1234, 304, 263, 1139, 29892, 3113, 1016, 29915, 29873, 6232, 2089, 2472, 29889, 13, 13, 29966, 829, 14816, 29903, 6778, 13, 13, 29961, 25580, 29962, 4803, 278, 1494, 12785, 310, 3030, 304, 1234, 278, 1139, 472, 278, 1095, 29889, 29871, 13, 3644, 366, 1016, 29915, 29873, 1073, 278, 1234, 29892, 925, 1827, 393, 366, 1016, 29915, 29873, 1073, 29892, 1016, 29915, 29873, 1018, 304, 1207, 701, 385, 1234, 29889, 29871, 13, 11403, 2211, 25260, 7472, 322, 3013, 278, 1234, 408, 3022, 895, 408, 1950, 29889, 29871, 13, 5398, 26227, 508, 367, 2309, 313, 29896, 29897, 491, 365, 26369, 411, 2560, 9508, 292, 763, 376, 7789, 567, 363, 1060, 29979, 29999, 7790, 29876, 29896, 19602, 376, 5618, 526, 278, 1014, 1484, 1338, 363, 3657, 15387, 1060, 29979, 29999, 29973, 613, 313, 29906, 29897, 491, 773, 3414, 29899, 14940, 11994, 29936, 321, 29889, 29887, 29889, 376, 6113, 263, 5828, 27887, 1213, 363, 5007, 263, 9554, 29892, 470, 313, 29941, 29897, 411, 5199, 10970, 29889, 13, 13, 5398, 26227, 508, 367, 2309, 313, 29896, 29897, 491, 365, 26369, 411, 2560, 9508, 292, 763, 376, 7789, 567, 363, 1060, 29979, 29999, 7790, 29876, 29896, 19602, 376, 5618, 526, 278, 1014, 1484, 1338, 363, 3657, 15387, 1060, 29979, 29999, 29973, 613, 313, 29906, 29897, 491, 773, 3414, 29899, 14940, 11994, 29936, 321, 29889, 29887, 29889, 376, 6113, 263, 5828, 27887, 1213, 363, 5007, 263, 9554, 29892, 470, 313, 29941, 29897, 411, 5199, 10970, 29889, 13, 13, 5398, 26227, 508, 367, 2309, 313, 29896, 29897, 491, 365, 26369, 411, 2560, 9508, 292, 763, 376, 7789, 567, 363, 1060, 29979, 29999, 7790, 29876, 29896, 19602, 376, 5618, 526, 278, 1014, 1484, 1338, 363, 3657, 15387, 1060, 29979, 29999, 29973, 613, 313, 29906, 29897, 491, 773, 3414, 29899, 14940, 11994, 29936, 321, 29889, 29887, 29889, 376, 6113, 263, 5828, 27887, 1213, 363, 5007, 263, 9554, 29892, 470, 313, 29941, 29897, 411, 5199, 10970, 29889, 13, 13, 1451, 16047, 267, 297, 1472, 29899, 8489, 18987, 322, 3414, 26227, 29901, 1858, 9450, 975, 263, 3309, 29891, 4955, 322, 17583, 3902, 8253, 278, 1650, 2913, 3933, 18066, 292, 29889, 365, 26369, 29879, 21117, 304, 10365, 13900, 746, 20050, 411, 15668, 4436, 29892, 3907, 963, 3109, 16424, 9401, 304, 25618, 1058, 5110, 515, 14260, 322, 1059, 29889, 13, 16492, 29901, 1724, 526, 278, 13501, 304, 9330, 897, 510, 3283, 29973, 13, 29648, 1319, 673, 29901, 518, 29914, 25580, 29962, 13, 5398, 26227, 508, 367, 26733, 297, 2211, 5837, 29901, 313, 29896, 29897, 773, 2560, 9508, 292, 763, 376, 7789, 567, 363, 1060, 29979, 29999, 7790, 29876, 29896, 19602, 376, 5618, 526, 278, 1014, 1484, 1338, 363, 3657, 15387, 1060, 29979, 29999, 29973, 613, 313, 29906, 29897, 491, 773, 3414, 29899, 14940, 11994, 29892, 470, 313, 29941, 29897, 411, 5199, 10970, 29889, 2], 'total_duration': 1364428708, 'load_duration': 1246375, 'sample_count': 62, 'sample_duration': 44859000, 'prompt_eval_count': 1, 'eval_count': 62, 'eval_duration': 1313002000}\n"
+     ]
+    }
+   ],
+   "source": [
+    "from langchain.schema import LLMResult\n",
+    "from langchain.callbacks.base import BaseCallbackHandler\n",
+    "\n",
+    "class GenerationStatisticsCallback(BaseCallbackHandler):\n",
+    "    def on_llm_end(self, response: LLMResult, **kwargs) -> None:\n",
+    "        print(response.generations[0][0].generation_info)\n",
+    "        \n",
+    "callback_manager = CallbackManager([StreamingStdOutCallbackHandler(), GenerationStatisticsCallback()])\n",
+    "\n",
+    "llm = Ollama(base_url=\"http://localhost:11434\",\n",
+    "             model=\"llama2\",\n",
+    "             verbose=True,\n",
+    "             callback_manager=callback_manager)\n",
+    "\n",
+    "qa_chain = RetrievalQA.from_chain_type(\n",
+    "    llm,\n",
+    "    retriever=vectorstore.as_retriever(),\n",
+    "    chain_type_kwargs={\"prompt\": QA_CHAIN_PROMPT},\n",
+    ")\n",
+    "\n",
+    "question = \"What are the approaches to Task Decomposition?\"\n",
+    "result = qa_chain({\"query\": question})"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "`eval_count` / (`eval_duration`/10e9)  gets `tok / s`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 57,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "47.22003469910937"
+      ]
+     },
+     "execution_count": 57,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "62 / (1313002000/1000/1000/1000)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

docs/extras/integrations/llms/symblai_nebula.ipynb

@@ -0,0 +1,106 @@
+{
+ "cells": [
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "id": "9597802c",
+   "metadata": {},
+   "source": [
+    "# Nebula\n",
+    "\n",
+    "[Nebula](https://symbl.ai/nebula/) is a fully-managed Conversation platform, on which you can build, deploy, and manage scalable AI applications.\n",
+    "\n",
+    "This example goes over how to use LangChain to interact with the [Nebula platform](https://docs.symbl.ai/docs/nebula-llm-overview). \n",
+    "\n",
+    "It will send the requests to Nebula Service endpoint, which concatenates `SYMBLAI_NEBULA_SERVICE_URL` and `SYMBLAI_NEBULA_SERVICE_PATH`, with a token defined in `SYMBLAI_NEBULA_SERVICE_TOKEN`"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f15ebe0d",
+   "metadata": {},
+   "source": [
+    "### Integrate with a LLMChain"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5472a7cd-af26-48ca-ae9b-5f6ae73c74d2",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "\n",
+    "os.environ[\"NEBULA_SERVICE_URL\"] = NEBULA_SERVICE_URL\n",
+    "os.environ[\"NEBULA_SERVICE_PATH\"] = NEBULA_SERVICE_PATH\n",
+    "os.environ[\"NEBULA_SERVICE_API_KEY\"] = NEBULA_SERVICE_API_KEY"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6fb585dd",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from langchain.llms import OpenLLM\n",
+    "\n",
+    "llm = OpenLLM(\n",
+    "    conversation=\"<Drop your text conversation that you want to ask Nebula to analyze here>\",\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "035dea0f",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from langchain import PromptTemplate, LLMChain\n",
+    "\n",
+    "template = \"Identify the {count} main objectives or goals mentioned in this context concisely in less points. Emphasize on key intents.\"\n",
+    "\n",
+    "prompt = PromptTemplate(template=template, input_variables=[\"count\"])\n",
+    "\n",
+    "llm_chain = LLMChain(prompt=prompt, llm=llm)\n",
+    "\n",
+    "generated = llm_chain.run(count=\"five\")\n",
+    "print(generated)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.8"
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "a0a0263b650d907a3bfe41c0f8d6a63a071b884df3cfdc1579f00cdc1aed6b03"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/llms/vllm.ipynb

@@ -0,0 +1,196 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "499c3142-2033-437d-a60a-731988ac6074",
+   "metadata": {},
+   "source": [
+    "# vLLM\n",
+    "\n",
+    "[vLLM](https://vllm.readthedocs.io/en/latest/index.html) is a fast and easy-to-use library for LLM inference and serving, offering:\n",
+    "* State-of-the-art serving throughput \n",
+    "* Efficient management of attention key and value memory with PagedAttention\n",
+    "* Continuous batching of incoming requests\n",
+    "* Optimized CUDA kernels\n",
+    "\n",
+    "This notebooks goes over how to use a LLM with langchain and vLLM.\n",
+    "\n",
+    "To use, you should have the `vllm` python package installed."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "8a3f2666-5c75-4797-967a-7915a247bf33",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "#!pip install vllm -q"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "84e350f7-21f6-455b-b1f0-8b0116a2fd49",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "INFO 08-06 11:37:33 llm_engine.py:70] Initializing an LLM engine with config: model='mosaicml/mpt-7b', tokenizer='mosaicml/mpt-7b', tokenizer_mode=auto, trust_remote_code=True, dtype=torch.bfloat16, use_dummy_weights=False, download_dir=None, use_np_weights=False, tensor_parallel_size=1, seed=0)\n",
+      "INFO 08-06 11:37:41 llm_engine.py:196] # GPU blocks: 861, # CPU blocks: 512\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processed prompts: 100%|██████████| 1/1 [00:00<00:00,  2.00it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "What is the capital of France ? The capital of France is Paris.\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "from langchain.llms import VLLM\n",
+    "\n",
+    "llm = VLLM(model=\"mosaicml/mpt-7b\",\n",
+    "           trust_remote_code=True,  # mandatory for hf models\n",
+    "           max_new_tokens=128,\n",
+    "           top_k=10,\n",
+    "           top_p=0.95,\n",
+    "           temperature=0.8,\n",
+    ")\n",
+    "\n",
+    "print(llm(\"What is the capital of France ?\"))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "94a3b41d-8329-4f8f-94f9-453d7f132214",
+   "metadata": {},
+   "source": [
+    "## Integrate the model in an LLMChain"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "5605b7a1-fa63-49c1-934d-8b4ef8d71dd5",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processed prompts: 100%|██████████| 1/1 [00:01<00:00,  1.34s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "1. The first Pokemon game was released in 1996.\n",
+      "2. The president was Bill Clinton.\n",
+      "3. Clinton was president from 1993 to 2001.\n",
+      "4. The answer is Clinton.\n",
+      "\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "from langchain import PromptTemplate, LLMChain\n",
+    "\n",
+    "template = \"\"\"Question: {question}\n",
+    "\n",
+    "Answer: Let's think step by step.\"\"\"\n",
+    "prompt = PromptTemplate(template=template, input_variables=[\"question\"])\n",
+    "\n",
+    "llm_chain = LLMChain(prompt=prompt, llm=llm)\n",
+    "\n",
+    "question = \"Who was the US president in the year the first Pokemon game was released?\"\n",
+    "\n",
+    "print(llm_chain.run(question))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "56826aba-d08b-4838-8bfa-ca96e463b25d",
+   "metadata": {},
+   "source": [
+    "## Distributed Inference\n",
+    "\n",
+    "vLLM supports distributed tensor-parallel inference and serving. \n",
+    "\n",
+    "To run multi-GPU inference with the LLM class, set the `tensor_parallel_size` argument to the number of GPUs you want to use. For example, to run inference on 4 GPUs"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f8c25c35-47b5-459d-9985-3cf546e9ac16",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.llms import VLLM\n",
+    "\n",
+    "llm = VLLM(model=\"mosaicml/mpt-30b\",\n",
+    "           tensor_parallel_size=4,\n",
+    "           trust_remote_code=True,  # mandatory for hf models\n",
+    ")\n",
+    "\n",
+    "llm(\"What is the future of AI?\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "conda_pytorch_p310",
+   "language": "python",
+   "name": "conda_pytorch_p310"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/memory/rockset_chat_message_history.ipynb

@@ -0,0 +1,67 @@
+{
+ "cells": [
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Rockset Chat Message History\n",
+    "\n",
+    "This notebook goes over how to use [Rockset](https://rockset.com/docs) to store chat message history. \n",
+    "\n",
+    "To begin, with get your API key from the [Rockset console](https://console.rockset.com/apikeys). Find your API region for the Rockset [API reference](https://rockset.com/docs/rest-api#introduction)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "vscode": {
+     "languageId": "plaintext"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "from langchain.memory.chat_message_histories import RocksetChatMessageHistory\n",
+    "from rockset import RocksetClient, Regions\n",
+    "\n",
+    "history = RocksetChatMessageHistory(\n",
+    "    session_id=\"MySession\",\n",
+    "    client=RocksetClient(\n",
+    "        api_key=\"YOUR API KEY\", \n",
+    "        host=Regions.usw2a1 # us-west-2 Oregon\n",
+    "    ),\n",
+    "    collection=\"langchain_demo\",\n",
+    "    sync=True\n",
+    ")\n",
+    "history.add_user_message(\"hi!\")\n",
+    "history.add_ai_message(\"whats up?\")\n",
+    "print(history.messages)"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The output should be something like:\n",
+    "\n",
+    "```python\n",
+    "[\n",
+    "    HumanMessage(content='hi!', additional_kwargs={'id': '2e62f1c2-e9f7-465e-b551-49bae07fe9f0'}, example=False), \n",
+    "    AIMessage(content='whats up?', additional_kwargs={'id': 'b9be8eda-4c18-4cf8-81c3-e91e876927d0'}, example=False)\n",
+    "]\n",
+    "\n",
+    "```"
+   ]
+  }
+ ],
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

docs/extras/integrations/memory/streamlit_chat_message_history.ipynb

@@ -7,8 +7,17 @@
    "source": [
     "# Streamlit Chat Message History\n",
     "\n",
-    "This notebook goes over how to use Streamlit to store chat message history. Note, StreamlitChatMessageHistory only works when run in a Streamlit app. For more on Streamlit check out their\n",
-    "[getting started documentation](https://docs.streamlit.io/library/get-started)."
+    "This notebook goes over how to store and use chat message history in a Streamlit app. StreamlitChatMessageHistory will store messages in\n",
+    "[Streamlit session state](https://docs.streamlit.io/library/api-reference/session-state)\n",
+    "at the specified `key=`. The default key is `\"langchain_messages\"`.\n",
+    "\n",
+    "- Note, StreamlitChatMessageHistory only works when run in a Streamlit app.\n",
+    "- You may also be interested in [StreamlitCallbackHandler](/docs/integrations/callbacks/streamlit) for LangChain.\n",
+    "- For more on Streamlit check out their\n",
+    "[getting started documentation](https://docs.streamlit.io/library/get-started).\n",
+    "\n",
+    "You can see the [full app example running here](https://langchain-st-memory.streamlit.app/), and more examples in\n",
+    "[github.com/langchain-ai/streamlit-agent](https://github.com/langchain-ai/streamlit-agent)."
    ]
   },
   {
@@ -20,7 +29,7 @@
    "source": [
     "from langchain.memory import StreamlitChatMessageHistory\n",
     "\n",
-    "history = StreamlitChatMessageHistory(\"foo\")\n",
+    "history = StreamlitChatMessageHistory(key=\"chat_messages\")\n",
     "\n",
     "history.add_user_message(\"hi!\")\n",
     "history.add_ai_message(\"whats up?\")"
@@ -35,6 +44,90 @@
    "source": [
     "history.messages"
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b60dc735",
+   "metadata": {},
+   "source": [
+    "You can integrate StreamlitChatMessageHistory into ConversationBufferMemory and chains or agents as usual. The history will be persisted across re-runs of the Streamlit app within a given user session. A given StreamlitChatMessageHistory will NOT be persisted or shared across user sessions."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "42ab5bf3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from langchain.memory import ConversationBufferMemory\n",
+    "from langchain.memory.chat_message_histories import StreamlitChatMessageHistory\n",
+    "\n",
+    "# Optionally, specify your own session_state key for storing messages\n",
+    "msgs = StreamlitChatMessageHistory(key=\"special_app_key\")\n",
+    "\n",
+    "memory = ConversationBufferMemory(memory_key=\"history\", chat_memory=msgs)\n",
+    "if len(msgs.messages) == 0:\n",
+    "    msgs.add_ai_message(\"How can I help you?\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a29252de",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.chains import LLMChain\n",
+    "from langchain.llms import OpenAI\n",
+    "from langchain.prompts import PromptTemplate\n",
+    "template = \"\"\"You are an AI chatbot having a conversation with a human.\n",
+    "\n",
+    "{history}\n",
+    "Human: {human_input}\n",
+    "AI: \"\"\"\n",
+    "prompt = PromptTemplate(input_variables=[\"history\", \"human_input\"], template=template)\n",
+    "\n",
+    "# Add the memory to an LLMChain as usual\n",
+    "llm_chain = LLMChain(llm=OpenAI(), prompt=prompt, memory=memory)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7cd99b4b",
+   "metadata": {},
+   "source": [
+    "Conversational Streamlit apps will often re-draw each previous chat message on every re-run. This is easy to do by iterating through `StreamlitChatMessageHistory.messages`:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3bdb637b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import streamlit as st\n",
+    "\n",
+    "for msg in msgs.messages:\n",
+    "    st.chat_message(msg.type).write(msg.content)\n",
+    "\n",
+    "if prompt := st.chat_input():\n",
+    "    st.chat_message(\"human\").write(prompt)\n",
+    "\n",
+    "    # As usual, new messages are added to StreamlitChatMessageHistory when the Chain is called.\n",
+    "    response = llm_chain.run(prompt)\n",
+    "    st.chat_message(\"ai\").write(response)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7adaf3d6",
+   "metadata": {},
+   "source": [
+    "**[View the final app](https://langchain-st-memory.streamlit.app/).**"
+   ]
   }
  ],
  "metadata": {

docs/extras/integrations/providers/bedrock.mdx

@@ -13,7 +13,7 @@ pip install boto3
 See a [usage example](/docs/integrations/llms/bedrock).
 
 ```python
-from langchain import Bedrock
+from langchain.llms.bedrock import Bedrock
 ```
 
 ## Text Embedding Models

docs/extras/integrations/providers/fireworks.md

@@ -19,4 +19,4 @@ llm = Fireworks(model="fireworks-llama-v2-13b-chat", max_tokens=256, temperature
 llm("Name 3 sports.")
 ```
 
-For a more detailed walkthrough, see [here](/docs/extras/modules/model_io/models/llms/integrations/Fireworks.ipynb).
+For a more detailed walkthrough, see [here](/docs/integrations/llms/Fireworks).

docs/extras/integrations/providers/log10.mdx

@@ -0,0 +1,104 @@
+# Log10
+
+This page covers how to use the [Log10](https://log10.io) within LangChain.
+
+## What is Log10?
+
+Log10 is an [open source](https://github.com/log10-io/log10) proxiless LLM data management and application development platform that lets you log, debug and tag your Langchain calls.
+
+## Quick start
+
+1. Create your free account at [log10.io](https://log10.io)
+2. Add your `LOG10_TOKEN` and `LOG10_ORG_ID` from the Settings and Organization tabs respectively as environment variables.
+3. Also add `LOG10_URL=https://log10.io` and your usual LLM API key: for e.g. `OPENAI_API_KEY` or `ANTHROPIC_API_KEY` to your environment
+
+## How to enable Log10 data management for Langchain
+
+Integration with log10 is a simple one-line `log10_callback` integration as shown below:
+
+```python
+from langchain.chat_models import ChatOpenAI
+from langchain.schema import HumanMessage
+
+from log10.langchain import Log10Callback
+from log10.llm import Log10Config
+
+log10_callback = Log10Callback(log10_config=Log10Config())
+
+messages = [
+    HumanMessage(content="You are a ping pong machine"),
+    HumanMessage(content="Ping?"),
+]
+
+llm = ChatOpenAI(model_name="gpt-3.5-turbo", callbacks=[log10_callback])
+```
+
+[Log10 + Langchain + Logs docs](https://github.com/log10-io/log10/blob/main/logging.md#langchain-logger)
+
+[More details + screenshots](https://log10.io/docs/logs) including instructions for self-hosting logs
+
+## How to use tags with Log10
+
+```python
+from langchain import OpenAI
+from langchain.chat_models import ChatAnthropic
+from langchain.chat_models import ChatOpenAI
+from langchain.schema import HumanMessage
+
+from log10.langchain import Log10Callback
+from log10.llm import Log10Config
+
+log10_callback = Log10Callback(log10_config=Log10Config())
+
+messages = [
+    HumanMessage(content="You are a ping pong machine"),
+    HumanMessage(content="Ping?"),
+]
+
+llm = ChatOpenAI(model_name="gpt-3.5-turbo", callbacks=[log10_callback], temperature=0.5, tags=["test"])
+completion = llm.predict_messages(messages, tags=["foobar"])
+print(completion)
+
+llm = ChatAnthropic(model="claude-2", callbacks=[log10_callback], temperature=0.7, tags=["baz"])
+llm.predict_messages(messages)
+print(completion)
+
+llm = OpenAI(model_name="text-davinci-003", callbacks=[log10_callback], temperature=0.5)
+completion = llm.predict("You are a ping pong machine.\nPing?\n")
+print(completion)
+```
+
+You can also intermix direct OpenAI calls and Langchain LLM calls:
+
+```python
+import os
+from log10.load import log10, log10_session
+import openai
+from langchain import OpenAI
+
+log10(openai)
+
+with log10_session(tags=["foo", "bar"]):
+    # Log a direct OpenAI call
+    response = openai.Completion.create(
+        model="text-ada-001",
+        prompt="Where is the Eiffel Tower?",
+        temperature=0,
+        max_tokens=1024,
+        top_p=1,
+        frequency_penalty=0,
+        presence_penalty=0,
+    )
+    print(response)
+
+    # Log a call via Langchain
+    llm = OpenAI(model_name="text-ada-001", temperature=0.5)
+    response = llm.predict("You are a ping pong machine.\nPing?\n")
+    print(response)
+```
+
+## How to debug Langchain calls
+
+[Example of debugging](https://log10.io/docs/prompt_chain_debugging)
+
+[More Langchain examples](https://github.com/log10-io/log10/tree/main/examples#langchain)

docs/extras/integrations/providers/pubmed.md

@@ -0,0 +1,30 @@
+# PubMed
+
+# PubMed
+
+>[PubMed®](https://pubmed.ncbi.nlm.nih.gov/) by `The National Center for Biotechnology Information, National Library of Medicine` 
+> comprises more than 35 million citations for biomedical literature from `MEDLINE`, life science journals, and online books. 
+> Citations may include links to full text content from `PubMed Central` and publisher web sites.
+
+## Setup
+You need to install a python package.
+
+```bash
+pip install xmltodict
+```
+
+### Retriever
+
+See a [usage example](/docs/integrations/retrievers/pubmed).
+
+```python
+from langchain.retrievers import PubMedRetriever
+```
+
+### Document Loader
+
+See a [usage example](/docs/integrations/document_loaders/pubmed).
+
+```python
+from langchain.document_loaders import PubMedLoader
+```

docs/extras/integrations/providers/rockset.mdx

@@ -23,4 +23,11 @@ from langchain.vectorstores import Rockset
 See a [usage example](/docs/integrations/document_loaders/rockset).
 ```python
 from langchain.document_loaders import RocksetLoader
+```
+
+## Chat Message History
+
+See a [usage example](/docs/integrations/memory/rockset_chat_message_history).
+```python
+from langchain.memory.chat_message_histories import RocksetChatMessageHistory
 ```

docs/extras/integrations/providers/symblai_nebula.mdx

@@ -0,0 +1,20 @@
+# Nebula
+
+This page covers how to use [Nebula](https://symbl.ai/nebula), [Symbl.ai](https://symbl.ai/)'s LLM, ecosystem within LangChain.
+It is broken into two parts: installation and setup, and then references to specific Nebula wrappers.
+
+## Installation and Setup
+
+- Get an Nebula API Key and set as environment variables (`SYMBLAI_NEBULA_SERVICE_URL`, `SYMBLAI_NEBULA_SERVICE_PATH`, `SYMBLAI_NEBULA_SERVICE_TOKEN`)
+  - Sign up for a FREE Symbl.ai/Nebula Account: [https://nebula.symbl.ai/playground/](https://nebula.symbl.ai/playground/)
+- Please see the [Nebula documentation](https://docs.symbl.ai/docs/nebula-llm-overview) for more details.
+  - No time? Visit the [Nebula Quickstart Guide](https://docs.symbl.ai/docs/nebula-quickstart).
+
+## Wrappers
+
+### LLM
+
+There exists an Nebula LLM wrapper, which you can access with 
+```python
+from langchain.llms import Nebula
+```

docs/extras/integrations/providers/tensorflow_datasets.mdx

@@ -0,0 +1,31 @@
+# TensorFlow Datasets
+
+>[TensorFlow Datasets](https://www.tensorflow.org/datasets) is a collection of datasets ready to use, 
+> with TensorFlow or other Python ML frameworks, such as Jax. All datasets are exposed 
+> as [tf.data.Datasets](https://www.tensorflow.org/api_docs/python/tf/data/Dataset), 
+> enabling easy-to-use and high-performance input pipelines. To get started see 
+> the [guide](https://www.tensorflow.org/datasets/overview) and 
+> the [list of datasets](https://www.tensorflow.org/datasets/catalog/overview#all_datasets).
+
+
+
+## Installation and Setup
+
+You need to install `tensorflow` and `tensorflow-datasets` python packages.
+
+```bash
+pip install tensorflow
+```
+
+```bash
+pip install tensorflow-dataset
+```
+
+
+## Document Loader
+
+See a [usage example](/docs/integrations/document_loaders/tensorflow_datasets).
+
+```python
+from langchain.document_loaders import TensorflowDatasetLoader
+```

docs/extras/integrations/retrievers/pubmed.ipynb

@@ -7,14 +7,15 @@
    "source": [
     "# PubMed\n",
     "\n",
-    "This notebook goes over how to use `PubMed` as a retriever\n",
     "\n",
-    "`PubMed®` comprises more than 35 million citations for biomedical literature from `MEDLINE`, life science journals, and online books. Citations may include links to full text content from `PubMed Central` and publisher web sites."
+    ">[PubMed®](https://pubmed.ncbi.nlm.nih.gov/) by `The National Center for Biotechnology Information, National Library of Medicine` comprises more than 35 million citations for biomedical literature from `MEDLINE`, life science journals, and online books. Citations may include links to full text content from `PubMed Central` and publisher web sites.\n",
+    "\n",
+    "This notebook goes over how to use `PubMed` as a retriever"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 12,
    "id": "aecaff63",
    "metadata": {},
    "outputs": [],
@@ -24,7 +25,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 34,
    "id": "f2f7e8d3",
    "metadata": {},
    "outputs": [],
@@ -34,19 +35,19 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 35,
    "id": "ed115aa1",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "[Document(page_content='', metadata={'uid': '37268021', 'title': 'Dermatology in the wake of an AI revolution: who gets a say?', 'pub_date': '<Year>2023</Year><Month>May</Month><Day>31</Day>'}),\n",
-       " Document(page_content='', metadata={'uid': '37267643', 'title': 'What is ChatGPT and what do we do with it? Implications of the age of AI for nursing and midwifery practice and education: An editorial.', 'pub_date': '<Year>2023</Year><Month>May</Month><Day>30</Day>'}),\n",
-       " Document(page_content='The nursing field has undergone notable changes over time and is projected to undergo further modifications in the future, owing to the advent of sophisticated technologies and growing healthcare needs. The advent of ChatGPT, an AI-powered language model, is expected to exert a significant influence on the nursing profession, specifically in the domains of patient care and instruction. The present article delves into the ramifications of ChatGPT within the nursing domain and accentuates its capacity and constraints to transform the discipline.', metadata={'uid': '37266721', 'title': 'The Impact of ChatGPT on the Nursing Profession: Revolutionizing Patient Care and Education.', 'pub_date': '<Year>2023</Year><Month>Jun</Month><Day>02</Day>'})]"
+       "[Document(page_content='', metadata={'uid': '37549050', 'Title': 'ChatGPT: \"To Be or Not to Be\" in Bikini Bottom.', 'Published': '--', 'Copyright Information': ''}),\n",
+       " Document(page_content=\"BACKGROUND: ChatGPT is a large language model that has performed well on professional examinations in the fields of medicine, law, and business. However, it is unclear how ChatGPT would perform on an examination assessing professionalism and situational judgement for doctors.\\nOBJECTIVE: We evaluated the performance of ChatGPT on the Situational Judgement Test (SJT): a national examination taken by all final-year medical students in the United Kingdom. This examination is designed to assess attributes such as communication, teamwork, patient safety, prioritization skills, professionalism, and ethics.\\nMETHODS: All questions from the UK Foundation Programme Office's (UKFPO's) 2023 SJT practice examination were inputted into ChatGPT. For each question, ChatGPT's answers and rationales were recorded and assessed on the basis of the official UK Foundation Programme Office scoring template. Questions were categorized into domains of Good Medical Practice on the basis of the domains referenced in the rationales provided in the scoring sheet. Questions without clear domain links were screened by reviewers and assigned one or multiple domains. ChatGPT's overall performance, as well as its performance across the domains of Good Medical Practice, was evaluated.\\nRESULTS: Overall, ChatGPT performed well, scoring 76% on the SJT but scoring full marks on only a few questions (9%), which may reflect possible flaws in ChatGPT's situational judgement or inconsistencies in the reasoning across questions (or both) in the examination itself. ChatGPT demonstrated consistent performance across the 4 outlined domains in Good Medical Practice for doctors.\\nCONCLUSIONS: Further research is needed to understand the potential applications of large language models, such as ChatGPT, in medical education for standardizing questions and providing consistent rationales for examinations assessing professionalism and ethics.\", metadata={'uid': '37548997', 'Title': 'Performance of ChatGPT on the Situational Judgement Test-A Professional Dilemmas-Based Examination for Doctors in the United Kingdom.', 'Published': '2023-08-07', 'Copyright Information': '©Robin J Borchert, Charlotte R Hickman, Jack Pepys, Timothy J Sadler. Originally published in JMIR Medical Education (https://mededu.jmir.org), 07.08.2023.'}),\n",
+       " Document(page_content='', metadata={'uid': '37548971', 'Title': \"Large Language Models Answer Medical Questions Accurately, but Can't Match Clinicians' Knowledge.\", 'Published': '2023-08-07', 'Copyright Information': ''})]"
       ]
      },
-     "execution_count": 9,
+     "execution_count": 35,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -54,6 +55,14 @@
    "source": [
     "retriever.get_relevant_documents(\"chatgpt\")"
    ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a9ff7a25-bb4b-4cd5-896d-72f70f4af49b",
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {
@@ -72,7 +81,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.6"
+   "version": "3.10.12"
   }
  },
  "nbformat": 4,

docs/extras/integrations/text_embedding/bge_huggingface.ipynb

@@ -0,0 +1,84 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "719619d3",
+   "metadata": {},
+   "source": [
+    "# BGE Hugging Face Embeddings\n",
+    "\n",
+    "This notebook shows how to use BGE Embeddings through Hugging Face"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "f7a54279",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [],
+   "source": [
+    "# !pip install sentence_transformers"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "9e1d5b6b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.embeddings import HuggingFaceBgeEmbeddings\n",
+    "\n",
+    "model_name = \"BAAI/bge-small-en\"\n",
+    "model_kwargs = {'device': 'cpu'}\n",
+    "encode_kwargs = {'normalize_embeddings': False}\n",
+    "hf = HuggingFaceBgeEmbeddings(\n",
+    "    model_name=model_name,\n",
+    "    model_kwargs=model_kwargs,\n",
+    "    encode_kwargs=encode_kwargs\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "e59d1a89",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "embedding = hf.embed_query(\"hi this is harrison\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e596315f",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/toolkits/multion.ipynb

@@ -5,7 +5,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Multion Toolkit\n",
+    "# MultiOn Toolkit\n",
     "\n",
     "This notebook walks you through connecting LangChain to the MultiOn Client in your browser\n",
     "\n",
@@ -18,7 +18,32 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "!pip install --upgrade multion > /dev/null"
+    "!pip install --upgrade multion langchain -q"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.agents.agent_toolkits import MultionToolkit\n",
+    "import os\n",
+    "\n",
+    "\n",
+    "toolkit = MultionToolkit()\n",
+    "\n",
+    "toolkit"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "tools = toolkit.get_tools()\n",
+    "tools"
    ]
   },
   {
@@ -38,8 +63,9 @@
    "outputs": [],
    "source": [
     "# Authorize connection to your Browser extention\n",
-    "import multion \n",
-    "multion.login()\n"
+    "import multion\n",
+    "multion.login()\n",
+    "\n"
    ]
   },
   {
@@ -57,38 +83,18 @@
    },
    "outputs": [],
    "source": [
-    "from langchain.agents.agent_toolkits import create_multion_agent\n",
-    "from langchain.tools.multion.tool import MultionClientTool\n",
-    "from langchain.agents.agent_types import AgentType\n",
-    "from langchain.chat_models import ChatOpenAI"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "tags": []
-   },
-   "outputs": [],
-   "source": [
-    "\n",
-    "agent_executor = create_multion_agent(\n",
-    "    llm=ChatOpenAI(temperature=0),\n",
-    "    tool=MultionClientTool(),\n",
-    "    agent_type=AgentType.OPENAI_FUNCTIONS,\n",
-    "    verbose=True\n",
-    ")\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "tags": []
-   },
-   "outputs": [],
-   "source": [
-    "agent.run(\"show me the weather today\")"
+    "from langchain import OpenAI\n",
+    "from langchain.agents import initialize_agent, AgentType\n",
+    "llm = OpenAI(temperature=0)\n",
+    "from langchain.agents.agent_toolkits import MultionToolkit\n",
+    "toolkit = MultionToolkit()\n",
+    "tools=toolkit.get_tools()\n",
+    "agent = initialize_agent(\n",
+    "    tools=toolkit.get_tools(),\n",
+    "    llm=llm,\n",
+    "    agent=AgentType.STRUCTURED_CHAT_ZERO_SHOT_REACT_DESCRIPTION,\n",
+    "    verbose = True\n",
+    ")"
    ]
   },
   {
@@ -100,7 +106,7 @@
    "outputs": [],
    "source": [
     "agent.run(\n",
-    "    \"Tweet about Elon Musk\"\n",
+    "    \"Tweet 'Hi from MultiOn'\"\n",
     ")"
    ]
   }

docs/extras/integrations/tools/awslambda.ipynb

@@ -66,7 +66,7 @@
    "outputs": [],
    "source": [
     "from langchain import OpenAI\n",
-    "from langchain.agents import load_tools, AgentType\n",
+    "from langchain.agents import load_tools, initialize_agent, AgentType\n",
     "\n",
     "llm = OpenAI(temperature=0)\n",
     "\n",

docs/extras/integrations/tools/dalle_image_generator.ipynb

@@ -0,0 +1,181 @@
+{
+  "cells": [
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "# Dall-E Image Generator\n",
+        "\n",
+        "This notebook shows how you can generate images from a prompt synthesized using an OpenAI LLM. The images are generated using Dall-E, which uses the same OpenAI API key as the LLM."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# Needed if you would like to display images in the notebook\n",
+        "!pip install opencv-python scikit-image"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 4,
+      "metadata": {
+        "id": "q-k8wmp0zquh"
+      },
+      "outputs": [],
+      "source": [
+        "from langchain.llms import OpenAI\n",
+        "import os\n",
+        "os.environ[\"OPENAI_API_KEY\"] = \"<your-key-here>\""
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "## Run as a chain"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 5,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "from langchain.utilities.dalle_image_generator import DallEAPIWrapper\n",
+        "from langchain.prompts import PromptTemplate\n",
+        "from langchain.chains import LLMChain\n",
+        "\n",
+        "llm = OpenAI(temperature=0.9)\n",
+        "prompt = PromptTemplate(\n",
+        "    input_variables=[\"image_desc\"],\n",
+        "    template=\"Generate a detailed prompt to generate an image based on the following description: {image_desc}\",\n",
+        ")\n",
+        "chain = LLMChain(llm=llm, prompt=prompt)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 7,
+      "metadata": {},
+      "outputs": [
+        {
+          "name": "stdout",
+          "output_type": "stream",
+          "text": [
+            "https://oaidalleapiprodscus.blob.core.windows.net/private/org-rocrupyvzgcl4yf25rqq6d1v/user-WsxrbKyP2c8rfhCKWDyMfe8N/img-mg1OWiziXxQN1aR2XRsLNndg.png?st=2023-01-31T07%3A34%3A15Z&se=2023-01-31T09%3A34%3A15Z&sp=r&sv=2021-08-06&sr=b&rscd=inline&rsct=image/png&skoid=6aaadede-4fb3-4698-a8f6-684d7786b067&sktid=a48cca56-e6da-484e-a814-9c849652bcb3&skt=2023-01-30T22%3A19%3A44Z&ske=2023-01-31T22%3A19%3A44Z&sks=b&skv=2021-08-06&sig=XDPee5aEng%2BcbXq2mqhh39uHGZTBmJgGAerSd0g%2BMEs%3D\n"
+          ]
+        }
+      ],
+      "source": [
+        "image_url = DallEAPIWrapper().run(chain.run(\"halloween night at a haunted museum\"))"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# You can click on the link above to display the image for\n",
+        "# Or you can try the options below to display the image inline in this notebook\n",
+        "\n",
+        "try:\n",
+        "    import google.colab\n",
+        "    IN_COLAB = True\n",
+        "except:\n",
+        "    IN_COLAB = False\n",
+        "\n",
+        "if IN_COLAB:\n",
+        "    from google.colab.patches import cv2_imshow # for image display\n",
+        "    from skimage import io\n",
+        "\n",
+        "    image = io.imread(image_url) \n",
+        "    cv2_imshow(image)\n",
+        "else:\n",
+        "    import cv2\n",
+        "    from skimage import io\n",
+        "\n",
+        "    image = io.imread(image_url) \n",
+        "    cv2.imshow('image', image)\n",
+        "    cv2.waitKey(0)   #wait for a keyboard input\n",
+        "    cv2.destroyAllWindows()\n"
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "## Run as a tool with an agent"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 9,
+      "metadata": {},
+      "outputs": [
+        {
+          "name": "stdout",
+          "output_type": "stream",
+          "text": [
+            "\n",
+            "\n",
+            "\u001b[1m> Entering new AgentExecutor chain...\u001b[0m\n",
+            "\u001b[32;1m\u001b[1;3m What is the best way to turn this description into an image?\n",
+            "Action: Dall-E Image Generator\n",
+            "Action Input: A spooky Halloween night at a haunted museum\u001b[0mhttps://oaidalleapiprodscus.blob.core.windows.net/private/org-rocrupyvzgcl4yf25rqq6d1v/user-WsxrbKyP2c8rfhCKWDyMfe8N/img-ogKfqxxOS5KWVSj4gYySR6FY.png?st=2023-01-31T07%3A38%3A25Z&se=2023-01-31T09%3A38%3A25Z&sp=r&sv=2021-08-06&sr=b&rscd=inline&rsct=image/png&skoid=6aaadede-4fb3-4698-a8f6-684d7786b067&sktid=a48cca56-e6da-484e-a814-9c849652bcb3&skt=2023-01-30T22%3A19%3A36Z&ske=2023-01-31T22%3A19%3A36Z&sks=b&skv=2021-08-06&sig=XsomxxBfu2CP78SzR9lrWUlbask4wBNnaMsHamy4VvU%3D\n",
+            "\n",
+            "Observation: \u001b[36;1m\u001b[1;3mhttps://oaidalleapiprodscus.blob.core.windows.net/private/org-rocrupyvzgcl4yf25rqq6d1v/user-WsxrbKyP2c8rfhCKWDyMfe8N/img-ogKfqxxOS5KWVSj4gYySR6FY.png?st=2023-01-31T07%3A38%3A25Z&se=2023-01-31T09%3A38%3A25Z&sp=r&sv=2021-08-06&sr=b&rscd=inline&rsct=image/png&skoid=6aaadede-4fb3-4698-a8f6-684d7786b067&sktid=a48cca56-e6da-484e-a814-9c849652bcb3&skt=2023-01-30T22%3A19%3A36Z&ske=2023-01-31T22%3A19%3A36Z&sks=b&skv=2021-08-06&sig=XsomxxBfu2CP78SzR9lrWUlbask4wBNnaMsHamy4VvU%3D\u001b[0m\n",
+            "Thought:\u001b[32;1m\u001b[1;3m With the image generated, I can now make my final answer.\n",
+            "Final Answer: An image of a Halloween night at a haunted museum can be seen here: https://oaidalleapiprodscus.blob.core.windows.net/private/org-rocrupyvzgcl4yf25rqq6d1v/user-WsxrbKyP2c8rfhCKWDyMfe8N/img-ogKfqxxOS5KWVSj4gYySR6FY.png?st=2023-01-31T07%3A38%3A25Z&se=2023-01-31T09%3A38%3A25Z&sp=r&sv=2021-08-06&sr=b&rscd=inline&rsct=image/png&skoid=6aaadede-4fb3-4698-a8f6-684d7786b067&sktid=a48cca56-e6da-484e-a814-9c849652bcb3&skt=2023-01-30T22\u001b[0m\n",
+            "\n",
+            "\u001b[1m> Finished chain.\u001b[0m\n"
+          ]
+        }
+      ],
+      "source": [
+        "from langchain.agents import load_tools\n",
+        "from langchain.agents import initialize_agent\n",
+        "\n",
+        "tools = load_tools(['dalle-image-generator'])\n",
+        "agent = initialize_agent(tools, llm, agent=\"zero-shot-react-description\", verbose=True)\n",
+        "output = agent.run(\"Create an image of a halloween night at a haunted museum\")"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "provenance": []
+    },
+    "kernelspec": {
+      "display_name": "langchain",
+      "language": "python",
+      "name": "python3"
+    },
+    "language_info": {
+      "codemirror_mode": {
+        "name": "ipython",
+        "version": 3
+      },
+      "file_extension": ".py",
+      "mimetype": "text/x-python",
+      "name": "python",
+      "nbconvert_exporter": "python",
+      "pygments_lexer": "ipython3",
+      "version": "3.9.16"
+    },
+    "vscode": {
+      "interpreter": {
+        "hash": "3570c8892273ffbeee7ead61dc7c022b73551d9f55fb2584ac0e8e8920b18a89"
+      }
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

docs/extras/integrations/tools/google_search.ipynb

@@ -9,7 +9,7 @@
     "\n",
     "This notebook goes over how to use the google search component.\n",
     "\n",
-    "First, you need to set up the proper API keys and environment variables. To set it up, create the GOOGLE_API_KEY in the Google Cloud credential console (https://console.cloud.google.com/apis/credentials) and a GOOGLE_CSE_ID using the Programmable Search Enginge (https://programmablesearchengine.google.com/controlpanel/create). Next, it is good to follow the instructions found [here](https://stackoverflow.com/questions/37083058/programmatically-searching-google-in-python-using-custom-search).\n",
+    "First, you need to set up the proper API keys and environment variables. To set it up, create the GOOGLE_API_KEY in the Google Cloud credential console (https://console.cloud.google.com/apis/credentials) and a GOOGLE_CSE_ID using the Programmable Search Engine (https://programmablesearchengine.google.com/controlpanel/create). Next, it is good to follow the instructions found [here](https://stackoverflow.com/questions/37083058/programmatically-searching-google-in-python-using-custom-search).\n",
     "\n",
     "Then we will need to set some environment variables."
    ]

docs/extras/integrations/tools/nuclia.ipynb

@@ -0,0 +1,173 @@
+{
+ "cells": [
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Nuclia Understanding API tool\n",
+    "\n",
+    "[Nuclia](https://nuclia.com) automatically indexes your unstructured data from any internal and external source, providing optimized search results and generative answers. It can handle video and audio transcription, image content extraction, and document parsing.\n",
+    "\n",
+    "The Nuclia Understanding API supports the processing of unstructured data, including text, web pages, documents, and audio/video contents. It extracts all texts wherever it is (using speech-to-text or OCR when needed), it identifies entities, it aslo extracts metadata, embedded files (like images in a PDF), and web links. It also provides a summary of the content.\n",
+    "\n",
+    "To use the Nuclia Understanding API, you need to have a Nuclia account. You can create one for free at [https://nuclia.cloud](https://nuclia.cloud), and then [create a NUA key](https://docs.nuclia.dev/docs/docs/using/understanding/intro)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install --upgrade protobuf\n",
+    "#!pip install nucliadb-protos"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "\n",
+    "os.environ[\"NUCLIA_ZONE\"] = \"<YOUR_ZONE>\"  # e.g. europe-1\n",
+    "os.environ[\"NUCLIA_NUA_KEY\"] = \"<YOUR_API_KEY>\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.tools.nuclia import NucliaUnderstandingAPI\n",
+    "\n",
+    "nua = NucliaUnderstandingAPI(enable_ml=False)"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You can push files to the Nuclia Understanding API using the `push` action. As the processing is done asynchronously, the results might be returned in a different order than the files were pushed. That is why you need to provide an `id` to match the results with the corresponding file."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "nua.run({\"action\": \"push\", \"id\": \"1\", \"path\": \"./report.docx\"})\n",
+    "nua.run({\"action\": \"push\", \"id\": \"2\", \"path\": \"./interview.mp4\"})"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You can now call the `pull` action in a loop until you get the JSON-formatted result."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import time\n",
+    "\n",
+    "pending = True\n",
+    "data = None\n",
+    "while pending:\n",
+    "    time.sleep(15)\n",
+    "    data = nua.run({\"action\": \"pull\", \"id\": \"1\", \"path\": None})\n",
+    "    if data:\n",
+    "        print(data)\n",
+    "        pending = False\n",
+    "    else:\n",
+    "        print(\"waiting...\")"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You can also do it in one step in `async` mode, you only need to do a push, and it will wait until the results are pulled:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import asyncio\n",
+    "\n",
+    "\n",
+    "async def process():\n",
+    "    data = await nua.arun(\n",
+    "        {\"action\": \"push\", \"id\": \"1\", \"path\": \"./talk.mp4\", \"text\": None}\n",
+    "    )\n",
+    "    print(data)\n",
+    "\n",
+    "\n",
+    "asyncio.run(process())"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Retrieved information\n",
+    "\n",
+    "Nuclia returns the following information:\n",
+    "\n",
+    "- file metadata\n",
+    "- extracted text\n",
+    "- nested text (like text in an embedded image)\n",
+    "- a summary (only when `enable_ml` is set to `True`)\n",
+    "- paragraphs and sentences splitting (defined by the position of their first and last characters, plus start time and end time for a video or audio file)\n",
+    "- named entities: people, dates, places, organizations, etc. (only when `enable_ml` is set to `True`)\n",
+    "- links\n",
+    "- a thumbnail\n",
+    "- embedded files\n",
+    "- the vector representations of the text (only when `enable_ml` is set to `True`)\n",
+    "\n",
+    "Note:\n",
+    "\n",
+    "  Generated files (thumbnail, extracted embedded files, etc.) are provided as a token. You can download them with the [`/processing/download` endpoint](https://docs.nuclia.dev/docs/api#operation/Download_binary_file_processing_download_get).\n",
+    "\n",
+    "  Also at any level, if an attribute exceeds a certain size, it will be put in a downloadable file and will be replaced in the document by a file pointer. This will consist of `{\"file\": {\"uri\": \"JWT_TOKEN\"}}`. The rule is that if the size of the message is greater than 1000000 characters, the biggest parts will be moved to downloadable files. First, the compression process will target vectors. If that is not enough, it will target large field metadata, and finally it will target extracted text.\n"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "langchain",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.5"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

docs/extras/integrations/tools/sqlite.mdx

@@ -1,7 +1,7 @@
-# SQL
+# SQL Database Chain
 
 This example demonstrates the use of the `SQLDatabaseChain` for answering questions over a SQL database.
 
 import Example from "@snippets/modules/chains/popular/sqlite.mdx"
 
-<Example/>
+<Example/>

docs/extras/integrations/vectorstores/faiss.ipynb

@@ -80,7 +80,7 @@
    "source": [
     "from langchain.document_loaders import TextLoader\n",
     "\n",
-    "loader = TextLoader(\"../../../state_of_the_union.txt\")\n",
+    "loader = TextLoader(\"../../../extras/modules/state_of_the_union.txt\")\n",
     "documents = loader.load()\n",
     "text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)\n",
     "docs = text_splitter.split_documents(documents)\n",
@@ -90,7 +90,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 3,
    "id": "5eabdb75",
    "metadata": {
     "tags": []
@@ -249,6 +249,50 @@
     "docs[0]"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "30c8f57b",
+   "metadata": {},
+   "source": [
+    "# Serializing and De-Serializing to bytes\n",
+    "\n",
+    "you can pickle the FAISS Index by these functions. If you use embeddings model which is of 90 mb (sentence-transformers/all-MiniLM-L6-v2 or any other model), the resultant pickle size would be more than 90 mb. the size of the model is also included in the overall size. To overcome this, use the below functions. These functions only serializes FAISS index and size would be much lesser. this can be helpful if you wish to store the index in database like sql."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d8faead5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pkl = db.serialize_to_bytes() # serializes the faiss index"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "eb083247",
+   "metadata": {
+    "vscode": {
+     "languageId": "r"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "embeddings = HuggingFaceEmbeddings(model_name=\"all-MiniLM-L6-v2\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e36e220b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "db = FAISS.deserialize_from_bytes(embeddings = embeddings, serialized = pkl) # Load the index"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "57da60d4",
@@ -473,6 +517,67 @@
     "for doc in results:\n",
     "    print(f\"Content: {doc.page_content}, Metadata: {doc.metadata}\")"
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1becca53",
+   "metadata": {},
+   "source": [
+    "## Delete\n",
+    "\n",
+    "You can also delete ids. Note that the ids to delete should be the ids in the docstore."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "1408b870",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "db.delete([db.index_to_docstore_id[0]])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "d13daf33",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "False"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Is now missing\n",
+    "0 in db.index_to_docstore_id"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "30ace43e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {
@@ -491,7 +596,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.9"
+   "version": "3.9.17"
   }
  },
  "nbformat": 4,

docs/extras/integrations/vectorstores/rockset.ipynb

@@ -21,7 +21,10 @@
     "\n",
     "1. Leverage the `Rockset` console to create a [collection](https://rockset.com/docs/collections/) with the Write API as your source. In this walkthrough, we create a collection named `langchain_demo`. \n",
     "    \n",
-    "    Configure the following [ingest transformation](https://rockset.com/docs/ingest-transformation/) to mark your embeddings field and take advantage of performance and storage optimizations:"
+    "    Configure the following [ingest transformation](https://rockset.com/docs/ingest-transformation/) to mark your embeddings field and take advantage of performance and storage optimizations:\n",
+    "\n",
+    "\n",
+    "   (We used OpenAI `text-embedding-ada-002` for this examples, where #length_of_vector_embedding = 1536)"
    ]
   },
   {
@@ -75,23 +78,10 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": null,
    "id": "29505c1e",
    "metadata": {},
-   "outputs": [
-    {
-     "ename": "InitializationException",
-     "evalue": "The rockset client was initialized incorrectly: An api key must be provided as a parameter to the RocksetClient or the Configuration object.",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
-      "\u001b[0;31mInitializationException\u001b[0m                   Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn[5], line 6\u001b[0m\n\u001b[1;32m      4\u001b[0m ROCKSET_API_KEY \u001b[39m=\u001b[39m os\u001b[39m.\u001b[39menviron\u001b[39m.\u001b[39mget(\u001b[39m\"\u001b[39m\u001b[39mROCKSET_API_KEY\u001b[39m\u001b[39m\"\u001b[39m) \u001b[39m# Verify ROCKSET_API_KEY environment variable\u001b[39;00m\n\u001b[1;32m      5\u001b[0m ROCKSET_API_SERVER \u001b[39m=\u001b[39m rockset\u001b[39m.\u001b[39mRegions\u001b[39m.\u001b[39musw2a1 \u001b[39m# Verify Rockset region\u001b[39;00m\n\u001b[0;32m----> 6\u001b[0m rockset_client \u001b[39m=\u001b[39m rockset\u001b[39m.\u001b[39;49mRocksetClient(ROCKSET_API_SERVER, ROCKSET_API_KEY)\n\u001b[1;32m      8\u001b[0m COLLECTION_NAME\u001b[39m=\u001b[39m\u001b[39m'\u001b[39m\u001b[39mlangchain_demo\u001b[39m\u001b[39m'\u001b[39m\n\u001b[1;32m      9\u001b[0m TEXT_KEY\u001b[39m=\u001b[39m\u001b[39m'\u001b[39m\u001b[39mdescription\u001b[39m\u001b[39m'\u001b[39m\n",
-      "File \u001b[0;32m~/Library/Python/3.9/lib/python/site-packages/rockset/rockset_client.py:242\u001b[0m, in \u001b[0;36mRocksetClient.__init__\u001b[0;34m(self, host, api_key, max_workers, config)\u001b[0m\n\u001b[1;32m    239\u001b[0m     config\u001b[39m.\u001b[39mhost \u001b[39m=\u001b[39m host\n\u001b[1;32m    241\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mnot\u001b[39;00m config\u001b[39m.\u001b[39mapi_key:\n\u001b[0;32m--> 242\u001b[0m     \u001b[39mraise\u001b[39;00m InitializationException(\n\u001b[1;32m    243\u001b[0m         \u001b[39m\"\u001b[39m\u001b[39mAn api key must be provided as a parameter to the RocksetClient or the Configuration object.\u001b[39m\u001b[39m\"\u001b[39m\n\u001b[1;32m    244\u001b[0m     )\n\u001b[1;32m    246\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mapi_client \u001b[39m=\u001b[39m ApiClient(config, max_workers\u001b[39m=\u001b[39mmax_workers)\n\u001b[1;32m    248\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mAliases \u001b[39m=\u001b[39m AliasesApiWrapper(\u001b[39mself\u001b[39m\u001b[39m.\u001b[39mapi_client)\n",
-      "\u001b[0;31mInitializationException\u001b[0m: The rockset client was initialized incorrectly: An api key must be provided as a parameter to the RocksetClient or the Configuration object."
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "import os\n",
     "import rockset\n",
@@ -118,18 +108,7 @@
    "execution_count": null,
    "id": "9740d8c4",
    "metadata": {},
-   "outputs": [
-    {
-     "ename": "",
-     "evalue": "",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[1;31mRunning cells with '/opt/local/bin/python3.11' requires the ipykernel package.\n",
-      "\u001b[1;31mRun the following command to install 'ipykernel' into the Python environment. \n",
-      "\u001b[1;31mCommand: '/opt/local/bin/python3.11 -m pip install ipykernel -U --user --force-reinstall'"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "from langchain.embeddings.openai import OpenAIEmbeddings\n",
     "from langchain.text_splitter import CharacterTextSplitter\n",
@@ -155,20 +134,9 @@
    "execution_count": null,
    "id": "85b6a6c5",
    "metadata": {},
-   "outputs": [
-    {
-     "ename": "",
-     "evalue": "",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[1;31mRunning cells with '/opt/local/bin/python3.11' requires the ipykernel package.\n",
-      "\u001b[1;31mRun the following command to install 'ipykernel' into the Python environment. \n",
-      "\u001b[1;31mCommand: '/opt/local/bin/python3.11 -m pip install ipykernel -U --user --force-reinstall'"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
-    "embeddings = OpenAIEmbeddings() # Verify OPENAI_KEY environment variable\n",
+    "embeddings = OpenAIEmbeddings() # Verify OPENAI_API_KEY environment variable\n",
     "\n",
     "docsearch = Rockset(\n",
     "    client=rockset_client,\n",
@@ -194,22 +162,10 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "id": "0bbf3df0",
    "metadata": {},
-   "outputs": [
-    {
-     "ename": "NameError",
-     "evalue": "name 'docsearch' is not defined",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
-      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn[1], line 2\u001b[0m\n\u001b[1;32m      1\u001b[0m query \u001b[39m=\u001b[39m \u001b[39m\"\u001b[39m\u001b[39mWhat did the president say about Ketanji Brown Jackson?\u001b[39m\u001b[39m\"\u001b[39m\n\u001b[0;32m----> 2\u001b[0m output \u001b[39m=\u001b[39m docsearch\u001b[39m.\u001b[39msimilarity_search_with_relevance_scores(query, \u001b[39m4\u001b[39m, Rockset\u001b[39m.\u001b[39mDistanceFunction\u001b[39m.\u001b[39mCOSINE_SIM)\n\u001b[1;32m      4\u001b[0m \u001b[39mprint\u001b[39m(\u001b[39m\"\u001b[39m\u001b[39moutput length:\u001b[39m\u001b[39m\"\u001b[39m, \u001b[39mlen\u001b[39m(output))\n\u001b[1;32m      5\u001b[0m \u001b[39mfor\u001b[39;00m d, dist \u001b[39min\u001b[39;00m output:\n",
-      "\u001b[0;31mNameError\u001b[0m: name 'docsearch' is not defined"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "query = \"What did the president say about Ketanji Brown Jackson\"\n",
     "output = docsearch.similarity_search_with_relevance_scores(\n",
@@ -313,7 +269,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.9.6"
+   "version": "3.10.12"
   }
  },
  "nbformat": 4,

docs/extras/integrations/vectorstores/usearch.ipynb

@@ -0,0 +1,195 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "bb384510-d9b4-4fa1-84c2-f181eb28487d",
+   "metadata": {},
+   "source": [
+    "# USearch\n",
+    ">[USearch](https://unum-cloud.github.io/usearch/) is a Smaller & Faster Single-File Vector Search Engine\n",
+    "\n",
+    "USearch's base functionality is identical to FAISS, and the interface should look familiar if you have ever investigated Approximate Nearest Neigbors search. FAISS is a widely recognized standard for high-performance vector search engines. USearch and FAISS both employ the same HNSW algorithm, but they differ significantly in their design principles. USearch is compact and broadly compatible without sacrificing performance, with a primary focus on user-defined metrics and fewer dependencies."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "497fcd89-e832-46a7-a74a-c71199666206",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "!pip install usearch"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "38237514-b3fa-44a4-9cff-30cd6bf50073",
+   "metadata": {},
+   "source": [
+    "We want to use OpenAIEmbeddings so we have to get the OpenAI API Key. "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "47f9b495-88f1-4286-8d5d-1416103931a7",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import getpass\n",
+    "\n",
+    "os.environ[\"OPENAI_API_KEY\"] = getpass.getpass(\"OpenAI API Key:\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "aac9563e",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from langchain.embeddings.openai import OpenAIEmbeddings\n",
+    "from langchain.text_splitter import CharacterTextSplitter\n",
+    "from langchain.vectorstores import USearch\n",
+    "from langchain.document_loaders import TextLoader"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "a3c3999a",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from langchain.document_loaders import TextLoader\n",
+    "\n",
+    "loader = TextLoader(\"../../../extras/modules/state_of_the_union.txt\")\n",
+    "documents = loader.load()\n",
+    "text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)\n",
+    "docs = text_splitter.split_documents(documents)\n",
+    "\n",
+    "embeddings = OpenAIEmbeddings()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "5eabdb75",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "db = USearch.from_documents(docs, embeddings)\n",
+    "\n",
+    "query = \"What did the president say about Ketanji Brown Jackson\"\n",
+    "docs = db.similarity_search(query)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "4b172de8",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Tonight. I call on the Senate to: Pass the Freedom to Vote Act. Pass the John Lewis Voting Rights Act. And while you’re at it, pass the Disclose Act so Americans can know who is funding our elections. \n",
+      "\n",
+      "Tonight, I’d like to honor someone who has dedicated his life to serve this country: Justice Stephen Breyer—an Army veteran, Constitutional scholar, and retiring Justice of the United States Supreme Court. Justice Breyer, thank you for your service. \n",
+      "\n",
+      "One of the most serious constitutional responsibilities a President has is nominating someone to serve on the United States Supreme Court. \n",
+      "\n",
+      "And I did that 4 days ago, when I nominated Circuit Court of Appeals Judge Ketanji Brown Jackson. One of our nation’s top legal minds, who will continue Justice Breyer’s legacy of excellence.\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(docs[0].page_content)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f13473b5",
+   "metadata": {},
+   "source": [
+    "## Similarity Search with score\n",
+    "The `similarity_search_with_score` method allows you to return not only the documents but also the distance score of the query to them. The returned distance score is L2 distance. Therefore, a lower score is better."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "186ee1d8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "docs_and_scores = db.similarity_search_with_score(query)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "284e04b5",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(Document(page_content='Tonight. I call on the Senate to: Pass the Freedom to Vote Act. Pass the John Lewis Voting Rights Act. And while you’re at it, pass the Disclose Act so Americans can know who is funding our elections. \\n\\nTonight, I’d like to honor someone who has dedicated his life to serve this country: Justice Stephen Breyer—an Army veteran, Constitutional scholar, and retiring Justice of the United States Supreme Court. Justice Breyer, thank you for your service. \\n\\nOne of the most serious constitutional responsibilities a President has is nominating someone to serve on the United States Supreme Court. \\n\\nAnd I did that 4 days ago, when I nominated Circuit Court of Appeals Judge Ketanji Brown Jackson. One of our nation’s top legal minds, who will continue Justice Breyer’s legacy of excellence.', metadata={'source': '../../../extras/modules/state_of_the_union.txt'}),\n",
+       " 0.1845687)"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "docs_and_scores[0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "483f6013-fb32-4756-a9e2-3d529fb81f68",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/integrations/vectorstores/vectara.ipynb

@@ -8,7 +8,7 @@
    "source": [
     "# Vectara\n",
     "\n",
-    ">[Vectara](https://vectara.com/) is a API platform for building GenAI applications. It provides an easy-to-use API for document indexing and query that is managed by Vectara and is optimized for performance and accuracy. \n",
+    ">[Vectara](https://vectara.com/) is a API platform for building GenAI applications. It provides an easy-to-use API for document indexing and querying that is managed by Vectara and is optimized for performance and accuracy. \n",
     "See the [Vectara API documentation ](https://docs.vectara.com/docs/) for more information on how to use the API.\n",
     "\n",
     "This notebook shows how to use functionality related to the `Vectara`'s integration with langchain.\n",

docs/extras/integrations/vectorstores/weaviate.ipynb

@@ -81,17 +81,18 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 21,
    "id": "53b7ce2d-3c09-4d1c-b66b-5769ce6746ae",
    "metadata": {},
    "outputs": [],
    "source": [
-    "os.environ[\"WEAVIATE_API_KEY\"] = getpass.getpass(\"WEAVIATE_API_KEY:\")"
+    "os.environ[\"WEAVIATE_API_KEY\"] = getpass.getpass(\"WEAVIATE_API_KEY:\")\n",
+    "WEAVIATE_API_KEY = os.environ[\"WEAVIATE_API_KEY\"]"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 7,
    "id": "aac9563e",
    "metadata": {
     "tags": []
@@ -106,7 +107,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 12,
    "id": "a3c3999a",
    "metadata": {},
    "outputs": [],
@@ -123,7 +124,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 14,
    "id": "21e9e528",
    "metadata": {},
    "outputs": [],
@@ -133,7 +134,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 15,
    "id": "b4170176",
    "metadata": {},
    "outputs": [],
@@ -144,7 +145,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 16,
    "id": "ecf3b890",
    "metadata": {},
    "outputs": [
@@ -166,6 +167,53 @@
     "print(docs[0].page_content)"
    ]
   },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "id": "7826d0ea",
+   "metadata": {},
+   "source": [
+    "## Authentication"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "13989a7c",
+   "metadata": {},
+   "source": [
+    "Weaviate instances have authentication enabled by default. You can use either a username/password combination or API key.  "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "f6604f1d",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<langchain.vectorstores.weaviate.Weaviate object at 0x107f46550>\n"
+     ]
+    }
+   ],
+   "source": [
+    "import weaviate\n",
+    "\n",
+    "client = weaviate.Client(url=WEAVIATE_URL, auth_client_secret=weaviate.AuthApiKey(WEAVIATE_API_KEY))\n",
+    "\n",
+    "# client = weaviate.Client(\n",
+    "#     url=WEAVIATE_URL,\n",
+    "#     auth_client_secret=weaviate.AuthClientPassword(\n",
+    "#         username = \"WCS_USERNAME\",  # Replace w/ your WCS username\n",
+    "#         password = \"WCS_PASSWORD\",  # Replace w/ your WCS password\n",
+    "#     ),\n",
+    "# )\n",
+    "\n",
+    "vectorstore = Weaviate.from_documents(documents, embeddings, client=client, by_text=False)"
+   ]
+  },
   {
    "attachments": {},
    "cell_type": "markdown",
@@ -187,7 +235,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 17,
    "id": "102105a1",
    "metadata": {},
    "outputs": [
@@ -213,7 +261,7 @@
    "id": "8fc3487b",
    "metadata": {},
    "source": [
-    "# Persistance"
+    "# Persistence"
    ]
   },
   {
@@ -249,7 +297,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": null,
    "id": "8b7df7ae",
    "metadata": {},
    "outputs": [
@@ -287,7 +335,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": null,
    "id": "5e824f3b",
    "metadata": {},
    "outputs": [],
@@ -298,7 +346,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 13,
+   "execution_count": null,
    "id": "61209cc3",
    "metadata": {},
    "outputs": [],
@@ -311,7 +359,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
+   "execution_count": null,
    "id": "4abc3d37",
    "metadata": {},
    "outputs": [],
@@ -327,7 +375,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
+   "execution_count": null,
    "id": "c7062393",
    "metadata": {},
    "outputs": [],
@@ -339,7 +387,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 16,
+   "execution_count": null,
    "id": "7e41b773",
    "metadata": {},
    "outputs": [

docs/extras/integrations/vectorstores/xata.ipynb

@@ -0,0 +1,240 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Xata\n",
+    "\n",
+    "> [Xata](https://xata.io) is a serverless data platform, based on PostgreSQL. It provides a Python SDK for interacting with your database, and a UI for managing your data.\n",
+    "> Xata has a native vector type, which can be added to any table, and supports similarity search. LangChain inserts vectors directly to Xata, and queries it for the nearest neighbors of a given vector, so that you can use all the LangChain Embeddings integrations with Xata."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "This notebook guides you how to use Xata as a VectorStore."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Setup\n",
+    "\n",
+    "### Create a database to use as a vector store\n",
+    "\n",
+    "In the [Xata UI](https://app.xata.io) create a new database. You can name it whatever you want, in this notepad we'll use `langchain`.\n",
+    "Create a table, again you can name it anything, but we will use `vectors`. Add the following columns via the UI:\n",
+    "\n",
+    "* `content` of type \"Text\". This is used to store the `Document.pageContent` values.\n",
+    "* `embedding` of type \"Vector\". Use the dimension used by the model you plan to use. In this notebook we use OpenAI embeddings, which have 1536 dimensions.\n",
+    "* `search` of type \"Text\". This is used as a metadata column by this example.\n",
+    "* any other columns you want to use as metadata. They are populated from the `Document.metadata` object. For example, if in the `Document.metadata` object you have a `title` property, you can create a `title` column in the table and it will be populated.\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's first install our dependencies:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "outputs": [],
+   "source": [
+    "!pip install xata==1.0.0a7 openai tiktoken langchain"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's load the OpenAI key to the environemnt. If you don't have one you can create an OpenAI account and create a key on this [page](https://platform.openai.com/account/api-keys)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import getpass\n",
+    "\n",
+    "os.environ[\"OPENAI_API_KEY\"] = getpass.getpass(\"OpenAI API Key:\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Similarly, we need to get the environment variables for Xata. You can create a new API key by visiting your [account settings](https://app.xata.io/settings). To find the database URL, go to the Settings page of the database that you have created. The database URL should look something like this: `https://demo-uni3q8.eu-west-1.xata.sh/db/langchain`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "api_key = getpass.getpass(\"Xata API key: \")\n",
+    "db_url = input(\"Xata database URL (copy it from your DB settings):\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "outputs": [],
+   "source": [
+    "from langchain.embeddings.openai import OpenAIEmbeddings\n",
+    "from langchain.text_splitter import CharacterTextSplitter\n",
+    "from langchain.document_loaders import TextLoader\n",
+    "from langchain.vectorstores.xata import XataVectorStore\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Create the Xata vector store\n",
+    "Let's import our test dataset:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "outputs": [],
+   "source": [
+    "loader = TextLoader(\"../../../state_of_the_union.txt\")\n",
+    "documents = loader.load()\n",
+    "text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)\n",
+    "docs = text_splitter.split_documents(documents)\n",
+    "\n",
+    "embeddings = OpenAIEmbeddings()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Now create the actual vector store, backed by the Xata table."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "outputs": [],
+   "source": [
+    "vector_store = XataVectorStore.from_documents(docs, embeddings, api_key=api_key, db_url=db_url, table_name=\"vectors\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "After running the above command, if you go to the Xata UI, you should see the documents loaded together with their embeddings."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Similarity Search"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "outputs": [],
+   "source": [
+    "query = \"What did the president say about Ketanji Brown Jackson\"\n",
+    "found_docs = vector_store.similarity_search(query)\n",
+    "print(found_docs)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Similarity Search with score (vector distance)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
+   },
+   "outputs": [],
+   "source": [
+    "query = \"What did the president say about Ketanji Brown Jackson\"\n",
+    "result = vector_store.similarity_search_with_score(query)\n",
+    "for doc, score in result:\n",
+    "    print(f\"document={doc}, score={score}\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}

docs/extras/modules/data_connection/document_transformers/post_retrieval/long_context_reorder.ipynb

@@ -109,11 +109,11 @@
    "source": [
     "# Reorder the documents:\n",
     "# Less relevant document will be at the middle of the list and more\n",
-    "# relevant elements at begining / end.\n",
+    "# relevant elements at beginning / end.\n",
     "reordering = LongContextReorder()\n",
     "reordered_docs = reordering.transform_documents(docs)\n",
     "\n",
-    "# Confirm that the 4 relevant documents are at begining and end.\n",
+    "# Confirm that the 4 relevant documents are at beginning and end.\n",
     "reordered_docs"
    ]
   },

docs/extras/modules/data_connection/retrievers/parent_document_retriever.ipynb

@@ -0,0 +1,440 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "34883374",
+   "metadata": {},
+   "source": [
+    "# Parent Document Retriever\n",
+    "\n",
+    "When splitting documents for retrieval, there are often conflicting desires:\n",
+    "\n",
+    "1. You may want to have small documents, so that their embeddings can most\n",
+    "    accurately reflect their meaning. If too long, then the embeddings can\n",
+    "    lose meaning.\n",
+    "2. You want to have long enough documents that the context of each chunk is\n",
+    "    retained.\n",
+    "\n",
+    "The ParentDocumentRetriever strikes that balance by splitting and storing\n",
+    "small chunks of data. During retrieval, it first fetches the small chunks\n",
+    "but then looks up the parent ids for those chunks and returns those larger\n",
+    "documents.\n",
+    "\n",
+    "Note that \"parent document\" refers to the document that a small chunk\n",
+    "originated from. This can either be the whole raw document OR a larger\n",
+    "chunk."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "8b6e74b2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.retrievers import ParentDocumentRetriever"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "1d17af96",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.vectorstores import Chroma\n",
+    "from langchain.embeddings import OpenAIEmbeddings\n",
+    "from langchain.text_splitter import RecursiveCharacterTextSplitter\n",
+    "from langchain.storage import InMemoryStore\n",
+    "from langchain.document_loaders import TextLoader"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "604ff981",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "loaders = [\n",
+    "    TextLoader('../../paul_graham_essay.txt'),\n",
+    "    TextLoader('../../state_of_the_union.txt'),\n",
+    "]\n",
+    "docs = []\n",
+    "for l in loaders:\n",
+    "    docs.extend(l.load())"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d3943f72",
+   "metadata": {},
+   "source": [
+    "## Retrieving Full Documents\n",
+    "\n",
+    "In this mode, we want to retrieve the full documents. Therefor, we only specify a child splitter."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "1a8b2e5f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# This text splitter is used to create the child documents\n",
+    "\n",
+    "child_splitter = RecursiveCharacterTextSplitter(chunk_size=400)\n",
+    "# The vectorstore to use to index the child chunks\n",
+    "vectorstore = Chroma(\n",
+    "    collection_name=\"full_documents\",\n",
+    "    embedding_function=OpenAIEmbeddings()\n",
+    ")\n",
+    "# The storage layer for the parent documents\n",
+    "store = InMemoryStore()\n",
+    "retriever = ParentDocumentRetriever(\n",
+    "    vectorstore=vectorstore, \n",
+    "    docstore=store, \n",
+    "    child_splitter=child_splitter,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "2b107935",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "retriever.add_documents(docs)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d05b97b7",
+   "metadata": {},
+   "source": [
+    "This should yield two keys, because we added two documents."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "30e3812b",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "['05fe8d8a-bf60-4f87-b576-4351b23df266',\n",
+       " '571cc9e5-9ef7-4f6c-b800-835c83a1858b']"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "list(store.yield_keys())"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f895d62b",
+   "metadata": {},
+   "source": [
+    "Let's now call the vectorstore search functionality - we should see that it returns small chunks (since we're storing the small chunks"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "b261c02c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "sub_docs = vectorstore.similarity_search(\"justice breyer\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "5108222f",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Tonight, I’d like to honor someone who has dedicated his life to serve this country: Justice Stephen Breyer—an Army veteran, Constitutional scholar, and retiring Justice of the United States Supreme Court. Justice Breyer, thank you for your service. \n",
+      "\n",
+      "One of the most serious constitutional responsibilities a President has is nominating someone to serve on the United States Supreme Court.\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(sub_docs[0].page_content)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bda8ed5a",
+   "metadata": {},
+   "source": [
+    "Let's now retrieve from the overall retriever. This should return large documents - since it returns the documents where the smaller chunks are located."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "419a91c4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "retrieved_docs = retriever.get_relevant_documents(\"justice breyer\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "cf10d250",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "38539"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(retrieved_docs[0].page_content)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "14f813a5",
+   "metadata": {},
+   "source": [
+    "## Retrieving Larger Chunks\n",
+    "\n",
+    "Sometimes, the full documents can be too big to want to retrieve them as is. In that case, what we really want to do is to first split the raw documents into larger chunks, and then split it into smaller chunks. We then index the smaller chunks, but on retrieval we retrieve the larger chunks (but still not the full documents)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "b6f9a4f0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# This text splitter is used to create the parent documents\n",
+    "parent_splitter = RecursiveCharacterTextSplitter(chunk_size=2000)\n",
+    "# This text splitter is used to create the child documents\n",
+    "# It should create documents smaller than the parent\n",
+    "child_splitter = RecursiveCharacterTextSplitter(chunk_size=400)\n",
+    "# The vectorstore to use to index the child chunks\n",
+    "vectorstore = Chroma(collection_name=\"split_parents\", embedding_function=OpenAIEmbeddings())\n",
+    "# The storage layer for the parent documents\n",
+    "store = InMemoryStore()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "19478ff3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "retriever = ParentDocumentRetriever(\n",
+    "    vectorstore=vectorstore, \n",
+    "    docstore=store, \n",
+    "    child_splitter=child_splitter,\n",
+    "    parent_splitter=parent_splitter,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "fe16e620",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "retriever.add_documents(docs)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "64ad3c8c",
+   "metadata": {},
+   "source": [
+    "We can see that there are much more than two documents now - these are the larger chunks"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "24d81886",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "66"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(list(store.yield_keys()))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "baaef673",
+   "metadata": {},
+   "source": [
+    "Let's make sure the underlying vectorstore still retrieves the small chunks."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "b1c859de",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "sub_docs = vectorstore.similarity_search(\"justice breyer\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "6fffa2eb",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Tonight, I’d like to honor someone who has dedicated his life to serve this country: Justice Stephen Breyer—an Army veteran, Constitutional scholar, and retiring Justice of the United States Supreme Court. Justice Breyer, thank you for your service. \n",
+      "\n",
+      "One of the most serious constitutional responsibilities a President has is nominating someone to serve on the United States Supreme Court.\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(sub_docs[0].page_content)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "3a3202df",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "retrieved_docs = retriever.get_relevant_documents(\"justice breyer\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "684fdb2c",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "1849"
+      ]
+     },
+     "execution_count": 18,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(retrieved_docs[0].page_content)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "9f17f662",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "In state after state, new laws have been passed, not only to suppress the vote, but to subvert entire elections. \n",
+      "\n",
+      "We cannot let this happen. \n",
+      "\n",
+      "Tonight. I call on the Senate to: Pass the Freedom to Vote Act. Pass the John Lewis Voting Rights Act. And while you’re at it, pass the Disclose Act so Americans can know who is funding our elections. \n",
+      "\n",
+      "Tonight, I’d like to honor someone who has dedicated his life to serve this country: Justice Stephen Breyer—an Army veteran, Constitutional scholar, and retiring Justice of the United States Supreme Court. Justice Breyer, thank you for your service. \n",
+      "\n",
+      "One of the most serious constitutional responsibilities a President has is nominating someone to serve on the United States Supreme Court. \n",
+      "\n",
+      "And I did that 4 days ago, when I nominated Circuit Court of Appeals Judge Ketanji Brown Jackson. One of our nation’s top legal minds, who will continue Justice Breyer’s legacy of excellence. \n",
+      "\n",
+      "A former top litigator in private practice. A former federal public defender. And from a family of public school educators and police officers. A consensus builder. Since she’s been nominated, she’s received a broad range of support—from the Fraternal Order of Police to former judges appointed by Democrats and Republicans. \n",
+      "\n",
+      "And if we are to advance liberty and justice, we need to secure the Border and fix the immigration system. \n",
+      "\n",
+      "We can do both. At our border, we’ve installed new technology like cutting-edge scanners to better detect drug smuggling.  \n",
+      "\n",
+      "We’ve set up joint patrols with Mexico and Guatemala to catch more human traffickers.  \n",
+      "\n",
+      "We’re putting in place dedicated immigration judges so families fleeing persecution and violence can have their cases heard faster. \n",
+      "\n",
+      "We’re securing commitments and supporting partners in South and Central America to host more refugees and secure their own borders.\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(retrieved_docs[0].page_content)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "facfdacb",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/use_cases/extraction.ipynb

@@ -0,0 +1,597 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "b84edb4e",
+   "metadata": {},
+   "source": [
+    "# Extraction\n",
+    "\n",
+    "[![Open In Collab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/langchain-ai/langchain/blob/master/docs/extras/use_cases/extraction.ipynb)\n",
+    "\n",
+    "## Use case\n",
+    "\n",
+    "Getting structured output from raw LLM generations is hard.\n",
+    "\n",
+    "For example, suppose you need the model output formatted with a specific schema for:\n",
+    "\n",
+    "- Extracting a structured row to insert into a database \n",
+    "- Extracting API parameters\n",
+    "- Extracting different parts of a user query (e.g., for semantic vs keyword search)\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "178dbc59",
+   "metadata": {},
+   "source": [
+    "![Image description](/img/extraction.png)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "97f474d4",
+   "metadata": {},
+   "source": [
+    "## Overview \n",
+    "\n",
+    "There are two primary approaches for this:\n",
+    "\n",
+    "- `Functions`: Some LLMs can call [functions](https://openai.com/blog/function-calling-and-other-api-updates) to extract arbitrary entities from LLM responses.\n",
+    "\n",
+    "- `Parsing`: [Output parsers](/docs/modules/model_io/output_parsers/) are classes that structure LLM responses. \n",
+    "\n",
+    "Only some LLMs support functions (e.g., OpenAI), and they are more general than parsers. \n",
+    "\n",
+    "Parsers extract precisely what is enumerated in a provided schema (e.g., specific attributes of a person).\n",
+    "\n",
+    "Functions can infer things beyond of a provided schema (e.g., attributes about a person that you did not ask for)."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "25d89f21",
+   "metadata": {},
+   "source": [
+    "## Quickstart\n",
+    "\n",
+    "OpenAI funtions are one way to get started with extraction.\n",
+    "\n",
+    "Define a schema that specifies the properties we want to extract from the LLM output.\n",
+    "\n",
+    "Then, we can use `create_extraction_chain` to extract our desired schema using an OpenAI function call."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3f5ec7a3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pip install langchain openai \n",
+    "\n",
+    "# Set env var OPENAI_API_KEY or load from a .env file:\n",
+    "# import dotenv\n",
+    "# dotenv.load_env()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "3e017ba0",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[{'name': 'Alex', 'height': 5, 'hair_color': 'blonde'},\n",
+       " {'name': 'Claudia', 'height': 6, 'hair_color': 'brunette'}]"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from langchain.chat_models import ChatOpenAI\n",
+    "from langchain.chains import create_extraction_chain\n",
+    "\n",
+    "# Schema\n",
+    "schema = {\n",
+    "    \"properties\": {\n",
+    "        \"name\": {\"type\": \"string\"},\n",
+    "        \"height\": {\"type\": \"integer\"},\n",
+    "        \"hair_color\": {\"type\": \"string\"},\n",
+    "    },\n",
+    "    \"required\": [\"name\", \"height\"],\n",
+    "}\n",
+    "\n",
+    "# Input \n",
+    "inp = \"\"\"Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\"\"\"\n",
+    "\n",
+    "# Run chain\n",
+    "llm = ChatOpenAI(temperature=0, model=\"gpt-3.5-turbo\")\n",
+    "chain = create_extraction_chain(schema, llm)\n",
+    "chain.run(inp)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6f7eb826",
+   "metadata": {},
+   "source": [
+    "## Option 1: OpenAI funtions\n",
+    "\n",
+    "### Looking under the hood\n",
+    "\n",
+    "Let's dig into what is happening when we call `create_extraction_chain`.\n",
+    "\n",
+    "The [LangSmith trace](https://smith.langchain.com/public/72bc3205-7743-4ca6-929a-966a9d4c2a77/r) shows that we call the function `information_extraction` on the input string, `inp`.\n",
+    "\n",
+    "![Image description](/img/extraction_trace_function.png)\n",
+    "\n",
+    "This `information_extraction` function is defined [here](https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/chains/openai_functions/extraction.py) and returns a dict.\n",
+    "\n",
+    "We can see the `dict` in the model output:\n",
+    "```\n",
+    " {\n",
+    "      \"info\": [\n",
+    "        {\n",
+    "          \"name\": \"Alex\",\n",
+    "          \"height\": 5,\n",
+    "          \"hair_color\": \"blonde\"\n",
+    "        },\n",
+    "        {\n",
+    "          \"name\": \"Claudia\",\n",
+    "          \"height\": 6,\n",
+    "          \"hair_color\": \"brunette\"\n",
+    "        }\n",
+    "      ]\n",
+    "    }\n",
+    "```\n",
+    "\n",
+    "The `create_extraction_chain` then parses the raw LLM output for us using [`JsonKeyOutputFunctionsParser`](https://github.com/langchain-ai/langchain/blob/f81e613086d211327b67b0fb591fd4d5f9a85860/libs/langchain/langchain/chains/openai_functions/extraction.py#L62).\n",
+    "\n",
+    "This results in the list of JSON objects returned by the chain above:\n",
+    "```\n",
+    "[{'name': 'Alex', 'height': 5, 'hair_color': 'blonde'},\n",
+    " {'name': 'Claudia', 'height': 6, 'hair_color': 'brunette'}]\n",
+    " ```"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "dcb03138",
+   "metadata": {},
+   "source": [
+    "### Multiple entity types\n",
+    "\n",
+    "We can extend this further.\n",
+    "\n",
+    "Let's say we want to differentiate between dogs and people.\n",
+    "\n",
+    "We can add `person_` and `dog_` prefixes for each property"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "01eae733",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[{'person_name': 'Alex',\n",
+       "  'person_height': 5,\n",
+       "  'person_hair_color': 'blonde',\n",
+       "  'dog_name': 'Frosty',\n",
+       "  'dog_breed': 'labrador'},\n",
+       " {'person_name': 'Claudia',\n",
+       "  'person_height': 6,\n",
+       "  'person_hair_color': 'brunette'}]"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "schema = {\n",
+    "    \"properties\": {\n",
+    "        \"person_name\": {\"type\": \"string\"},\n",
+    "        \"person_height\": {\"type\": \"integer\"},\n",
+    "        \"person_hair_color\": {\"type\": \"string\"},\n",
+    "        \"dog_name\": {\"type\": \"string\"},\n",
+    "        \"dog_breed\": {\"type\": \"string\"},\n",
+    "    },\n",
+    "    \"required\": [\"person_name\", \"person_height\"],\n",
+    "}\n",
+    "\n",
+    "chain = create_extraction_chain(schema, llm)\n",
+    "\n",
+    "inp = \"\"\"Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\n",
+    "Alex's dog Frosty is a labrador and likes to play hide and seek.\"\"\"\n",
+    "\n",
+    "chain.run(inp)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f205905c",
+   "metadata": {},
+   "source": [
+    "### Unrelated entities\n",
+    "\n",
+    "If we use `required: []`, we allow the model to return **only** person attributes or **only** dog attributes for a single entity (person or dog)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "6ff4ac7e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[{'person_name': 'Alex', 'person_height': 5, 'person_hair_color': 'blonde'},\n",
+       " {'person_name': 'Claudia',\n",
+       "  'person_height': 6,\n",
+       "  'person_hair_color': 'brunette'},\n",
+       " {'dog_name': 'Willow', 'dog_breed': 'German Shepherd'},\n",
+       " {'dog_name': 'Milo', 'dog_breed': 'border collie'}]"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "schema = {\n",
+    "    \"properties\": {\n",
+    "        \"person_name\": {\"type\": \"string\"},\n",
+    "        \"person_height\": {\"type\": \"integer\"},\n",
+    "        \"person_hair_color\": {\"type\": \"string\"},\n",
+    "        \"dog_name\": {\"type\": \"string\"},\n",
+    "        \"dog_breed\": {\"type\": \"string\"},\n",
+    "    },\n",
+    "    \"required\": [],\n",
+    "}\n",
+    "\n",
+    "chain = create_extraction_chain(schema, llm)\n",
+    "\n",
+    "inp = \"\"\"Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\n",
+    "Willow is a German Shepherd that likes to play with other dogs and can always be found playing with Milo, a border collie that lives close by.\"\"\"\n",
+    "\n",
+    "chain.run(inp)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "34f3b958",
+   "metadata": {},
+   "source": [
+    "### Extra information\n",
+    "\n",
+    "The power of functions (relative to using parsers alone) lies in the ability to perform sematic extraction.\n",
+    "\n",
+    "In particular, `we can ask for things that are not explictly enumerated in the schema`.\n",
+    "\n",
+    "Suppose we want unspecified additional information about dogs. \n",
+    "\n",
+    "We can use add a placeholder for unstructured extraction, `dog_extra_info`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "40c7b26f",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[{'person_name': 'Alex', 'person_height': 5, 'person_hair_color': 'blonde'},\n",
+       " {'person_name': 'Claudia',\n",
+       "  'person_height': 6,\n",
+       "  'person_hair_color': 'brunette'},\n",
+       " {'dog_name': 'Willow',\n",
+       "  'dog_breed': 'German Shepherd',\n",
+       "  'dog_extra_info': 'likes to play with other dogs'},\n",
+       " {'dog_name': 'Milo',\n",
+       "  'dog_breed': 'border collie',\n",
+       "  'dog_extra_info': 'lives close by'}]"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "schema = {\n",
+    "    \"properties\": {\n",
+    "        \"person_name\": {\"type\": \"string\"},\n",
+    "        \"person_height\": {\"type\": \"integer\"},\n",
+    "        \"person_hair_color\": {\"type\": \"string\"},\n",
+    "        \"dog_name\": {\"type\": \"string\"},\n",
+    "        \"dog_breed\": {\"type\": \"string\"},\n",
+    "        \"dog_extra_info\": {\"type\": \"string\"},\n",
+    "    },\n",
+    "}\n",
+    "\n",
+    "chain = create_extraction_chain(schema, llm)\n",
+    "chain.run(inp)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3a949c60",
+   "metadata": {},
+   "source": [
+    "This gives us additional information about the dogs."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bf71ddce",
+   "metadata": {},
+   "source": [
+    "### Pydantic \n",
+    "\n",
+    "Pydantic is a data validation and settings management library for Python. \n",
+    "\n",
+    "It allows you to create data classes with attributes that are automatically validated when you instantiate an object.\n",
+    "\n",
+    "Lets define a class with attributes annotated with types."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "d36a743b",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[Properties(person_name='Alex', person_height=5, person_hair_color='blonde', dog_breed=None, dog_name=None),\n",
+       " Properties(person_name='Claudia', person_height=6, person_hair_color='brunette', dog_breed=None, dog_name=None)]"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from typing import Optional, List\n",
+    "from pydantic import BaseModel, Field\n",
+    "from langchain.chains import create_extraction_chain_pydantic\n",
+    "\n",
+    "# Pydantic data class\n",
+    "class Properties(BaseModel):\n",
+    "    person_name: str\n",
+    "    person_height: int\n",
+    "    person_hair_color: str\n",
+    "    dog_breed: Optional[str]\n",
+    "    dog_name: Optional[str]\n",
+    "        \n",
+    "# Extraction\n",
+    "chain = create_extraction_chain_pydantic(pydantic_schema=Properties, llm=llm)\n",
+    "\n",
+    "# Run \n",
+    "inp = \"\"\"Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\"\"\"\n",
+    "chain.run(inp)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "07a0351a",
+   "metadata": {},
+   "source": [
+    "As we can see from the [trace](https://smith.langchain.com/public/fed50ae6-26bb-4235-a254-e0b7a229d10f/r), we use the function `information_extraction`, as above, with the Pydantic schema. "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cbd9f121",
+   "metadata": {},
+   "source": [
+    "## Option 2: Parsing\n",
+    "\n",
+    "[Output parsers](/docs/modules/model_io/output_parsers/) are classes that help structure language model responses. \n",
+    "\n",
+    "As shown above, they are used to parse the output of the OpenAI function calls in `create_extraction_chain`.\n",
+    "\n",
+    "But, they can be used independent of functions.\n",
+    "\n",
+    "### Pydantic\n",
+    "\n",
+    "Just as a above, let's parse a generation based on a Pydantic data class."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "64650362",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "People(people=[Person(person_name='Alex', person_height=5, person_hair_color='blonde', dog_breed=None, dog_name=None), Person(person_name='Claudia', person_height=6, person_hair_color='brunette', dog_breed=None, dog_name=None)])"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from typing import Sequence\n",
+    "from langchain.prompts import (\n",
+    "    PromptTemplate,\n",
+    "    ChatPromptTemplate,\n",
+    "    HumanMessagePromptTemplate,\n",
+    ")\n",
+    "from langchain.llms import OpenAI\n",
+    "from pydantic import BaseModel, Field, validator\n",
+    "from langchain.output_parsers import PydanticOutputParser\n",
+    "\n",
+    "class Person(BaseModel):\n",
+    "    person_name: str\n",
+    "    person_height: int\n",
+    "    person_hair_color: str\n",
+    "    dog_breed: Optional[str]\n",
+    "    dog_name: Optional[str]\n",
+    "\n",
+    "class People(BaseModel):\n",
+    "    \"\"\"Identifying information about all people in a text.\"\"\"\n",
+    "    people: Sequence[Person]\n",
+    "\n",
+    "        \n",
+    "# Run \n",
+    "query = \"\"\"Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\"\"\"\n",
+    "\n",
+    "# Set up a parser + inject instructions into the prompt template.\n",
+    "parser = PydanticOutputParser(pydantic_object=People)\n",
+    "\n",
+    "# Prompt\n",
+    "prompt = PromptTemplate(\n",
+    "    template=\"Answer the user query.\\n{format_instructions}\\n{query}\\n\",\n",
+    "    input_variables=[\"query\"],\n",
+    "    partial_variables={\"format_instructions\": parser.get_format_instructions()},\n",
+    ")\n",
+    "\n",
+    "# Run\n",
+    "_input = prompt.format_prompt(query=query)\n",
+    "model = OpenAI(temperature=0)\n",
+    "output = model(_input.to_string())\n",
+    "parser.parse(output)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "826899df",
+   "metadata": {},
+   "source": [
+    "We can see from the [LangSmith trace](https://smith.langchain.com/public/8e3aa858-467e-46a5-aa49-5db65f0a2b9a/r) that we get the same output as above.\n",
+    "\n",
+    "![Image description](/img/extraction_trace_function_2.png)\n",
+    "\n",
+    "We can see that we provide a two-shot prompt in order to instruct the LLM to output in our desired format.\n",
+    "\n",
+    "And, we need to do a bit more work:\n",
+    "\n",
+    "* Define a class that holds multiple instances of `Person`\n",
+    "* Explicty parse the output of the LLM to the Pydantic class\n",
+    "\n",
+    "We can see this for other cases, too."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "837c350e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "Joke(setup='Why did the chicken cross the road?', punchline='To get to the other side!')"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from langchain.prompts import (\n",
+    "    PromptTemplate,\n",
+    "    ChatPromptTemplate,\n",
+    "    HumanMessagePromptTemplate,\n",
+    ")\n",
+    "from langchain.llms import OpenAI\n",
+    "from pydantic import BaseModel, Field, validator\n",
+    "from langchain.output_parsers import PydanticOutputParser\n",
+    "\n",
+    "# Define your desired data structure.\n",
+    "class Joke(BaseModel):\n",
+    "    setup: str = Field(description=\"question to set up a joke\")\n",
+    "    punchline: str = Field(description=\"answer to resolve the joke\")\n",
+    "\n",
+    "    # You can add custom validation logic easily with Pydantic.\n",
+    "    @validator(\"setup\")\n",
+    "    def question_ends_with_question_mark(cls, field):\n",
+    "        if field[-1] != \"?\":\n",
+    "            raise ValueError(\"Badly formed question!\")\n",
+    "        return field\n",
+    "\n",
+    "# And a query intented to prompt a language model to populate the data structure.\n",
+    "joke_query = \"Tell me a joke.\"\n",
+    "\n",
+    "# Set up a parser + inject instructions into the prompt template.\n",
+    "parser = PydanticOutputParser(pydantic_object=Joke)\n",
+    "\n",
+    "# Prompt\n",
+    "prompt = PromptTemplate(\n",
+    "    template=\"Answer the user query.\\n{format_instructions}\\n{query}\\n\",\n",
+    "    input_variables=[\"query\"],\n",
+    "    partial_variables={\"format_instructions\": parser.get_format_instructions()},\n",
+    ")\n",
+    "\n",
+    "# Run\n",
+    "_input = prompt.format_prompt(query=joke_query)\n",
+    "model = OpenAI(temperature=0)\n",
+    "output = model(_input.to_string())\n",
+    "parser.parse(output)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d3601bde",
+   "metadata": {},
+   "source": [
+    "As we can see, we get an output of the `Joke` class, which respects our originally desired schema: 'setup' and 'punchline'.\n",
+    "\n",
+    "We can look at the [LangSmith trace](https://smith.langchain.com/public/69f11d41-41be-4319-93b0-6d0eda66e969/r) to see exactly what is going on under the hood.\n",
+    "\n",
+    "![Image description](/img/extraction_trace_joke.png)\n",
+    "\n",
+    "### Going deeper\n",
+    "\n",
+    "* The [output parser](/docs/modules/model_io/output_parsers/) documentation includes various parser examples for specific types (e.g., lists, datetimne, enum, etc).  \n",
+    "* [JSONFormer](/docs/integrations/llms/jsonformer_experimental) offers another way for structured decoding of a subset of the JSON Schema.\n",
+    "* [Kor](https://eyurtsev.github.io/kor/) is another library for extraction where schema and examples can be provided to the LLM."
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/extras/use_cases/extraction/index.mdx

@@ -1,24 +0,0 @@
----
-sidebar_position: 2
----
-
-# Extraction
-
-Most APIs and databases still deal with structured information.
-Therefore, in order to better work with those, it can be useful to extract structured information from text.
-Examples of this include:
-
-- Extracting a structured row to insert into a database from a sentence
-- Extracting multiple rows to insert into a database from a long document
-- Extracting the correct API parameters from a user query
-
-This work is extremely related to [output parsing](/docs/modules/model_io/output_parsers/).
-Output parsers are responsible for instructing the LLM to respond in a specific format.
-In this case, the output parsers specify the format of the data you would like to extract from the document.
-Then, in addition to the output format instructions, the prompt should also contain the data you would like to extract information from.
-
-While normal output parsers are good enough for basic structuring of response data,
-when doing extraction you often want to extract more complicated or nested structures.
-For a deep dive on extraction, we recommend checking out [`kor`](https://eyurtsev.github.io/kor/),
-a library that uses the existing LangChain chain and OutputParser abstractions
-but deep dives on allowing extraction of more complicated schemas.

docs/extras/use_cases/extraction/openai_extraction.ipynb

@@ -1,566 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "id": "6605e7f7",
-   "metadata": {},
-   "source": [
-    "# Extraction with OpenAI Functions\n",
-    "\n",
-    "The extraction chain uses the OpenAI `functions` parameter to specify a schema to extract entities from a document. This helps us make sure that the model outputs exactly the schema of entities and properties that we want, with their appropriate types.\n",
-    "\n",
-    "The extraction chain is to be used when we want to extract several entities with their properties from the same passage (i.e. what people were mentioned in this passage?)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "id": "34f04daf",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "/Users/harrisonchase/.pyenv/versions/3.9.1/envs/langchain/lib/python3.9/site-packages/deeplake/util/check_latest_version.py:32: UserWarning: A newer version of deeplake (3.6.4) is available. It's recommended that you update to the latest version using `pip install -U deeplake`.\n",
-      "  warnings.warn(\n"
-     ]
-    }
-   ],
-   "source": [
-    "from langchain.chat_models import ChatOpenAI\n",
-    "from langchain.chains import create_extraction_chain, create_extraction_chain_pydantic\n",
-    "from langchain.prompts import ChatPromptTemplate"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "id": "a2648974",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "llm = ChatOpenAI(temperature=0, model=\"gpt-3.5-turbo-0613\")"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "5ef034ce",
-   "metadata": {},
-   "source": [
-    "## Extracting entities"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "78ff9df9",
-   "metadata": {},
-   "source": [
-    "To extract entities, we need to create a schema where we specify all the properties we want to find and the type we expect them to have. We can also specify which of these properties are required and which are optional."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "id": "4ac43eba",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "schema = {\n",
-    "    \"properties\": {\n",
-    "        \"name\": {\"type\": \"string\"},\n",
-    "        \"height\": {\"type\": \"integer\"},\n",
-    "        \"hair_color\": {\"type\": \"string\"},\n",
-    "    },\n",
-    "    \"required\": [\"name\", \"height\"],\n",
-    "}"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "id": "640bd005",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "inp = \"\"\"\n",
-    "Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\n",
-    "        \"\"\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 6,
-   "id": "64313214",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "chain = create_extraction_chain(schema, llm)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "17c48adb",
-   "metadata": {},
-   "source": [
-    "As we can see, we extracted the required entities and their properties in the required format (it even calculated Claudia's height before returning!)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "id": "cc5436ed",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "[{'name': 'Alex', 'height': 5, 'hair_color': 'blonde'},\n",
-       " {'name': 'Claudia', 'height': 6, 'hair_color': 'brunette'}]"
-      ]
-     },
-     "execution_count": 7,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "chain.run(inp)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8d51fcdc",
-   "metadata": {},
-   "source": [
-    "## Several entity types"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "5813affe",
-   "metadata": {},
-   "source": [
-    "Notice that we are using OpenAI functions under the hood and thus the model can only call one function per request (with one, unique schema)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "511b9838",
-   "metadata": {},
-   "source": [
-    "If we want to extract more than one entity type, we need to introduce a little hack - we will define our properties with an included entity type. \n",
-    "\n",
-    "Following we have an example where we also want to extract dog attributes from the passage. Notice the 'person_' and 'dog_' prefixes we use for each property; this tells the model which entity type the property refers to. In this way, the model can return properties from several entity types in one single call."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "id": "cf243a26",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "schema = {\n",
-    "    \"properties\": {\n",
-    "        \"person_name\": {\"type\": \"string\"},\n",
-    "        \"person_height\": {\"type\": \"integer\"},\n",
-    "        \"person_hair_color\": {\"type\": \"string\"},\n",
-    "        \"dog_name\": {\"type\": \"string\"},\n",
-    "        \"dog_breed\": {\"type\": \"string\"},\n",
-    "    },\n",
-    "    \"required\": [\"person_name\", \"person_height\"],\n",
-    "}"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "id": "52841fb3",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "inp = \"\"\"\n",
-    "Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\n",
-    "Alex's dog Frosty is a labrador and likes to play hide and seek.\n",
-    "        \"\"\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "id": "93f904ab",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "chain = create_extraction_chain(schema, llm)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "eb074f7b",
-   "metadata": {},
-   "source": [
-    "People attributes and dog attributes were correctly extracted from the text in the same call"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 6,
-   "id": "db3e9e17",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "[{'person_name': 'Alex',\n",
-       "  'person_height': 5,\n",
-       "  'person_hair_color': 'blonde',\n",
-       "  'dog_name': 'Frosty',\n",
-       "  'dog_breed': 'labrador'},\n",
-       " {'person_name': 'Claudia',\n",
-       "  'person_height': 6,\n",
-       "  'person_hair_color': 'brunette'}]"
-      ]
-     },
-     "execution_count": 11,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "chain.run(inp)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "0273e0e2",
-   "metadata": {},
-   "source": [
-    "## Unrelated entities"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "c07b3480",
-   "metadata": {},
-   "source": [
-    "What if our entities are unrelated? In that case, the model will return the unrelated entities in different dictionaries, allowing us to successfully extract several unrelated entity types in the same call."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "01d98af0",
-   "metadata": {},
-   "source": [
-    "Notice that we use `required: []`: we need to allow the model to return **only** person attributes or **only** dog attributes for a single entity (person or dog)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 48,
-   "id": "e584c993",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "schema = {\n",
-    "    \"properties\": {\n",
-    "        \"person_name\": {\"type\": \"string\"},\n",
-    "        \"person_height\": {\"type\": \"integer\"},\n",
-    "        \"person_hair_color\": {\"type\": \"string\"},\n",
-    "        \"dog_name\": {\"type\": \"string\"},\n",
-    "        \"dog_breed\": {\"type\": \"string\"},\n",
-    "    },\n",
-    "    \"required\": [],\n",
-    "}"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 49,
-   "id": "ad6b105f",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "inp = \"\"\"\n",
-    "Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\n",
-    "\n",
-    "Willow is a German Shepherd that likes to play with other dogs and can always be found playing with Milo, a border collie that lives close by.\n",
-    "\"\"\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 50,
-   "id": "6bfe5a33",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "chain = create_extraction_chain(schema, llm)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "24fe09af",
-   "metadata": {},
-   "source": [
-    "We have each entity in its own separate dictionary, with only the appropriate attributes being returned"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 51,
-   "id": "f6e1fd89",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "[{'person_name': 'Alex', 'person_height': 5, 'person_hair_color': 'blonde'},\n",
-       " {'person_name': 'Claudia',\n",
-       "  'person_height': 6,\n",
-       "  'person_hair_color': 'brunette'},\n",
-       " {'dog_name': 'Willow', 'dog_breed': 'German Shepherd'},\n",
-       " {'dog_name': 'Milo', 'dog_breed': 'border collie'}]"
-      ]
-     },
-     "execution_count": 51,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "chain.run(inp)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "0ac466d1",
-   "metadata": {},
-   "source": [
-    "## Extra info for an entity"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "d240ffc1",
-   "metadata": {},
-   "source": [
-    "What if.. _we don't know what we want?_ More specifically, say we know a few properties we want to extract for a given entity but we also want to know if there's any extra information in the passage. Fortunately, we don't need to structure everything - we can have unstructured extraction as well. \n",
-    "\n",
-    "We can do this by introducing another hack, namely the *extra_info* attribute - let's see an example."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 68,
-   "id": "f19685f6",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "schema = {\n",
-    "    \"properties\": {\n",
-    "        \"person_name\": {\"type\": \"string\"},\n",
-    "        \"person_height\": {\"type\": \"integer\"},\n",
-    "        \"person_hair_color\": {\"type\": \"string\"},\n",
-    "        \"dog_name\": {\"type\": \"string\"},\n",
-    "        \"dog_breed\": {\"type\": \"string\"},\n",
-    "        \"dog_extra_info\": {\"type\": \"string\"},\n",
-    "    },\n",
-    "}"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 81,
-   "id": "200c3477",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "inp = \"\"\"\n",
-    "Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\n",
-    "\n",
-    "Willow is a German Shepherd that likes to play with other dogs and can always be found playing with Milo, a border collie that lives close by.\n",
-    "\"\"\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 82,
-   "id": "ddad7dc6",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "chain = create_extraction_chain(schema, llm)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "e5c0dbbc",
-   "metadata": {},
-   "source": [
-    "It is nice to know more about Willow and Milo!"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 83,
-   "id": "c22cfd30",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "[{'person_name': 'Alex', 'person_height': 5, 'person_hair_color': 'blonde'},\n",
-       " {'person_name': 'Claudia',\n",
-       "  'person_height': 6,\n",
-       "  'person_hair_color': 'brunette'},\n",
-       " {'dog_name': 'Willow',\n",
-       "  'dog_breed': 'German Shepherd',\n",
-       "  'dog_extra_information': 'likes to play with other dogs'},\n",
-       " {'dog_name': 'Milo',\n",
-       "  'dog_breed': 'border collie',\n",
-       "  'dog_extra_information': 'lives close by'}]"
-      ]
-     },
-     "execution_count": 83,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "chain.run(inp)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "698b4c4d",
-   "metadata": {},
-   "source": [
-    "## Pydantic example"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "6504a6d9",
-   "metadata": {},
-   "source": [
-    "We can also use a Pydantic schema to choose the required properties and types and we will set as 'Optional' those that are not strictly required.\n",
-    "\n",
-    "By using the `create_extraction_chain_pydantic` function, we can send a Pydantic schema as input and the output will be an instantiated object that respects our desired schema. \n",
-    "\n",
-    "In this way, we can specify our schema in the same manner that we would a new class or function in Python - with purely Pythonic types."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "id": "6792866b",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from typing import Optional, List\n",
-    "from pydantic import BaseModel, Field"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "id": "36a63761",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "class Properties(BaseModel):\n",
-    "    person_name: str\n",
-    "    person_height: int\n",
-    "    person_hair_color: str\n",
-    "    dog_breed: Optional[str]\n",
-    "    dog_name: Optional[str]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "id": "8ffd1e57",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "chain = create_extraction_chain_pydantic(pydantic_schema=Properties, llm=llm)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 10,
-   "id": "24baa954",
-   "metadata": {
-    "scrolled": false
-   },
-   "outputs": [],
-   "source": [
-    "inp = \"\"\"\n",
-    "Alex is 5 feet tall. Claudia is 1 feet taller Alex and jumps higher than him. Claudia is a brunette and Alex is blonde.\n",
-    "Alex's dog Frosty is a labrador and likes to play hide and seek.\n",
-    "        \"\"\""
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "84e0a241",
-   "metadata": {},
-   "source": [
-    "As we can see, we extracted the required entities and their properties in the required format:"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 11,
-   "id": "f771df58",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "[Properties(person_name='Alex', person_height=5, person_hair_color='blonde', dog_breed='labrador', dog_name='Frosty'),\n",
-       " Properties(person_name='Claudia', person_height=6, person_hair_color='brunette', dog_breed=None, dog_name=None)]"
-      ]
-     },
-     "execution_count": 11,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "chain.run(inp)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "0df61283",
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.11.3"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 5
-}

docs/extras/use_cases/question_answering/index.mdx

@@ -138,9 +138,9 @@ len(docs_svm)
     4
 
 Some common ways to improve on vector similarity search include:
-- `MultiQueryRetriever` [generates variants of the input question](/docs/modules/data_connection/retrievers/how_to/MultiQueryRetriever) to improve retrieval.
+- `MultiQueryRetriever` [generates variants of the input question](/docs/modules/data_connection/retrievers/MultiQueryRetriever) to improve retrieval.
 - `Max marginal relevance` selects for [relevance and diversity](https://www.cs.cmu.edu/~jgc/publication/The_Use_MMR_Diversity_Based_LTMIR_1998.pdf) among the retrieved documents.
-- Documents can be filtered during retrieval using [`metadata` filters](/docs/use_cases/question_answering/document-context-aware-QA).
+- Documents can be filtered during retrieval using [`metadata` filters](/docs/use_cases/question_answering/how_to/document-context-aware-QA).
 
 
 ```python

docs/extras/use_cases/sql.ipynb

@@ -0,0 +1,858 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# SQL\n",
+    "\n",
+    "[![Open In Collab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/langchain-ai/langchain/blob/master/docs/extras/use_cases/sql.ipynb)\n",
+    "\n",
+    "## Use case\n",
+    "\n",
+    "Enterprise data is often stored in SQL databases.\n",
+    "\n",
+    "LLMs make it possible to interact with SQL databases using natural langugae.\n",
+    "\n",
+    "LangChain offers SQL Chains and Agents to build and run SQL queries based on natural language prompts. \n",
+    "\n",
+    "These are compatible with any SQL dialect supported by SQLAlchemy (e.g., MySQL, PostgreSQL, Oracle SQL, Databricks, SQLite).\n",
+    "\n",
+    "They enable use cases such as:\n",
+    "\n",
+    "- Generating queries that will be run based on natural language questions\n",
+    "- Creating chatbots that can answer questions based on database data\n",
+    "- Building custom dashboards based on insights a user wants to analyze\n",
+    "\n",
+    "## Overview\n",
+    "\n",
+    "LangChain provides tools to interact with SQL Databases:\n",
+    "\n",
+    "1. `Build SQL queries` based on natural language user questions\n",
+    "2. `Query a SQL database` using chains for query creation and execution\n",
+    "3. `Interact with a SQL database` using agents for robust and flexible querying \n",
+    "\n",
+    "![sql_usecase.png](/img/sql_usecase.png)\n",
+    "\n",
+    "## Quickstart\n",
+    "\n",
+    "First, get required packages and set environment variables:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "! pip install langchain langchain-experimental openai\n",
+    "\n",
+    "# Set env var OPENAI_API_KEY or load from a .env file\n",
+    "# import dotenv\n",
+    "\n",
+    "# dotenv.load_env()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The below example will use a SQLite connection with Chinook database. \n",
+    " \n",
+    "Follow [installation steps](https://database.guide/2-sample-databases-sqlite/) to create `Chinook.db` in the same directory as this notebook:\n",
+    "\n",
+    "* Save [this file](https://raw.githubusercontent.com/lerocha/chinook-database/master/ChinookDatabase/DataSources/Chinook_Sqlite.sql) to the directory as `Chinook_Sqlite.sql`\n",
+    "* Run `sqlite3 Chinook.db`\n",
+    "* Run `.read Chinook_Sqlite.sql`\n",
+    "* Test `SELECT * FROM Artist LIMIT 10;`\n",
+    "\n",
+    "Now, `Chinhook.db` is in our directory.\n",
+    "\n",
+    "Let's create a `SQLDatabaseChain` to create and execute SQL queries."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.utilities import SQLDatabase\n",
+    "from langchain.llms import OpenAI\n",
+    "from langchain_experimental.sql import SQLDatabaseChain\n",
+    "\n",
+    "db = SQLDatabase.from_uri(\"sqlite:///Chinook.db\")\n",
+    "llm = OpenAI(temperature=0, verbose=True)\n",
+    "db_chain = SQLDatabaseChain.from_llm(llm, db, verbose=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "\u001b[1m> Entering new SQLDatabaseChain chain...\u001b[0m\n",
+      "How many employees are there?\n",
+      "SQLQuery:\u001b[32;1m\u001b[1;3mSELECT COUNT(*) FROM \"Employee\";\u001b[0m\n",
+      "SQLResult: \u001b[33;1m\u001b[1;3m[(8,)]\u001b[0m\n",
+      "Answer:\u001b[32;1m\u001b[1;3mThere are 8 employees.\u001b[0m\n",
+      "\u001b[1m> Finished chain.\u001b[0m\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "'There are 8 employees.'"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "db_chain.run(\"How many employees are there?\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Note that this both creates and executes the query. \n",
+    "\n",
+    "In the following sections, we will cover the 3 different use cases mentioned in the overview.\n",
+    "\n",
+    "### Go deeper\n",
+    "\n",
+    "You can load tabular data from other sources other than SQL Databases.\n",
+    "For example:\n",
+    "- [Loading a CSV file](/docs/integrations/document_loaders/csv)\n",
+    "- [Loading a Pandas DataFrame](/docs/integrations/document_loaders/pandas_dataframe)\n",
+    "Here you can [check full list of Document Loaders](/docs/integrations/document_loaders/)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Case 1: Text-to-SQL query\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {
+    "scrolled": false
+   },
+   "outputs": [],
+   "source": [
+    "from langchain.chat_models import ChatOpenAI\n",
+    "from langchain.chains import create_sql_query_chain"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's create the chain that will build the SQL Query:\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "SELECT COUNT(*) FROM Employee\n"
+     ]
+    }
+   ],
+   "source": [
+    "chain = create_sql_query_chain(ChatOpenAI(temperature=0), db)\n",
+    "response = chain.invoke({\"question\":\"How many employees are there\"})\n",
+    "print(response)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "After building the SQL query based on a user question, we can execute the query:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'[(8,)]'"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "db.run(response)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "As we can see, the SQL Query Builder chain **only created** the query, and we handled the **query execution separately**."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Go deeper\n",
+    "\n",
+    "**Looking under the hood**\n",
+    "\n",
+    "We can look at the [LangSmith trace](https://smith.langchain.com/public/c8fa52ea-be46-4829-bde2-52894970b830/r) to unpack this:\n",
+    "\n",
+    "[Some papers](https://arxiv.org/pdf/2204.00498.pdf) have reported good performance when prompting with:\n",
+    " \n",
+    "* A `CREATE TABLE` description for each table, which include column names, their types, etc\n",
+    "* Followed by three example rows in a `SELECT` statement\n",
+    "\n",
+    "`create_sql_query_chain` adopts this the best practice (see more in this [blog](https://blog.langchain.dev/llms-and-sql/)).  \n",
+    "![sql_usecase.png](/img/create_sql_query_chain.png)\n",
+    "\n",
+    "**Improvements**\n",
+    "\n",
+    "The query builder can be improved in several ways, such as (but not limited to):\n",
+    "\n",
+    "- Customizing database description to your specific use case\n",
+    "- Hardcoding a few examples of questions and their corresponding SQL query in the prompt\n",
+    "- Using a vector database to include dynamic examples that are relevant to the specific user question\n",
+    "\n",
+    "All these examples involve customizing the chain's prompt. \n",
+    "\n",
+    "For example, we can include a few examples in our prompt like so:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.prompts import PromptTemplate\n",
+    "\n",
+    "TEMPLATE = \"\"\"Given an input question, first create a syntactically correct {dialect} query to run, then look at the results of the query and return the answer.\n",
+    "Use the following format:\n",
+    "\n",
+    "Question: \"Question here\"\n",
+    "SQLQuery: \"SQL Query to run\"\n",
+    "SQLResult: \"Result of the SQLQuery\"\n",
+    "Answer: \"Final answer here\"\n",
+    "\n",
+    "Only use the following tables:\n",
+    "\n",
+    "{table_info}.\n",
+    "\n",
+    "Some examples of SQL queries that corrsespond to questions are:\n",
+    "\n",
+    "{few_shot_examples}\n",
+    "\n",
+    "Question: {input}\"\"\"\n",
+    "\n",
+    "CUSTOM_PROMPT = PromptTemplate(\n",
+    "    input_variables=[\"input\", \"few_shot_examples\", \"table_info\", \"dialect\"], template=TEMPLATE\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Case 2: Text-to-SQL query and execution\n",
+    "\n",
+    "We can use `SQLDatabaseChain` from `langchain_experimental` to create and run SQL queries."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.llms import OpenAI\n",
+    "from langchain_experimental.sql import SQLDatabaseChain\n",
+    "\n",
+    "llm = OpenAI(temperature=0, verbose=True)\n",
+    "db_chain = SQLDatabaseChain.from_llm(llm, db, verbose=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "\u001b[1m> Entering new SQLDatabaseChain chain...\u001b[0m\n",
+      "How many employees are there?\n",
+      "SQLQuery:\u001b[32;1m\u001b[1;3mSELECT COUNT(*) FROM \"Employee\";\u001b[0m\n",
+      "SQLResult: \u001b[33;1m\u001b[1;3m[(8,)]\u001b[0m\n",
+      "Answer:\u001b[32;1m\u001b[1;3mThere are 8 employees.\u001b[0m\n",
+      "\u001b[1m> Finished chain.\u001b[0m\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "'There are 8 employees.'"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "db_chain.run(\"How many employees are there?\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "As we can see, we get the same result as the previous case.\n",
+    "\n",
+    "Here, the chain **also handles the query execution** and provides a final answer based on the user question and the query result.\n",
+    "\n",
+    "**Be careful** while using this approach as it is susceptible to `SQL Injection`:\n",
+    "\n",
+    "* The chain is executing queries that are created by an LLM, and weren't validated\n",
+    "* e.g. records may be created, modified or deleted unintentionally_\n",
+    "\n",
+    "This is why we see the `SQLDatabaseChain` is inside `langchain_experimental`."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Go deeper\n",
+    "\n",
+    "**Looking under the hood**\n",
+    "\n",
+    "We can use the [LangSmith trace](https://smith.langchain.com/public/7f202a0c-1e35-42d6-a84a-6c2a58f697ef/r) to see what is happening under the hood:\n",
+    "\n",
+    "* As discussed above, first we create the query:\n",
+    "\n",
+    "```\n",
+    "text: ' SELECT COUNT(*) FROM \"Employee\";'\n",
+    "```\n",
+    "\n",
+    "* Then, it executes the query and passes the results to an LLM for synthesis.\n",
+    "\n",
+    "![sql_usecase.png](/img/sqldbchain_trace.png)\n",
+    "\n",
+    "**Improvements**\n",
+    "\n",
+    "The performance of the `SQLDatabaseChain` can be enhanced in several ways:\n",
+    "\n",
+    "- [Adding sample rows](#adding-sample-rows)\n",
+    "- [Specifying custom table information](/docs/integrations/tools/sqlite#custom-table-info)\n",
+    "- [Using Query Checker](/docs/integrations/tools/sqlite#use-query-checker) self-correct invalid SQL using parameter `use_query_checker=True`\n",
+    "- [Customizing the LLM Prompt](/docs/integrations/tools/sqlite#customize-prompt) include specific instructions or relevant information, using parameter `prompt=CUSTOM_PROMPT`\n",
+    "- [Get intermediate steps](/docs/integrations/tools/sqlite#return-intermediate-steps) access the SQL statement as well as the final result using parameter `return_intermediate_steps=True`\n",
+    "- [Limit the number of rows](/docs/integrations/tools/sqlite#choosing-how-to-limit-the-number-of-rows-returned) a query will return using parameter `top_k=5`\n",
+    "\n",
+    "You might find [SQLDatabaseSequentialChain](/docs/integrations/tools/sqlite#sqldatabasesequentialchain)\n",
+    "useful for cases in which the number of tables in the database is large.\n",
+    "\n",
+    "This `Sequential Chain` handles the process of:\n",
+    "\n",
+    "1. Determining which tables to use based on the user question\n",
+    "2. Calling the normal SQL database chain using only relevant tables\n",
+    "\n",
+    "**Adding Sample Rows**\n",
+    "\n",
+    "Providing sample data can help the LLM construct correct queries when the data format is not obvious. \n",
+    "\n",
+    "For example, we can tell LLM that artists are saved with their full names by providing two rows from the Track table.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "db = SQLDatabase.from_uri(\n",
+    "    \"sqlite:///Chinook.db\",\n",
+    "    include_tables=['Track'], # we include only one table to save tokens in the prompt :)\n",
+    "    sample_rows_in_table_info=2)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The sample rows are added to the prompt after each corresponding table's column information.\n",
+    "\n",
+    "We can use `db.table_info` and check which sample rows are included:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "CREATE TABLE \"Track\" (\n",
+      "\t\"TrackId\" INTEGER NOT NULL, \n",
+      "\t\"Name\" NVARCHAR(200) NOT NULL, \n",
+      "\t\"AlbumId\" INTEGER, \n",
+      "\t\"MediaTypeId\" INTEGER NOT NULL, \n",
+      "\t\"GenreId\" INTEGER, \n",
+      "\t\"Composer\" NVARCHAR(220), \n",
+      "\t\"Milliseconds\" INTEGER NOT NULL, \n",
+      "\t\"Bytes\" INTEGER, \n",
+      "\t\"UnitPrice\" NUMERIC(10, 2) NOT NULL, \n",
+      "\tPRIMARY KEY (\"TrackId\"), \n",
+      "\tFOREIGN KEY(\"MediaTypeId\") REFERENCES \"MediaType\" (\"MediaTypeId\"), \n",
+      "\tFOREIGN KEY(\"GenreId\") REFERENCES \"Genre\" (\"GenreId\"), \n",
+      "\tFOREIGN KEY(\"AlbumId\") REFERENCES \"Album\" (\"AlbumId\")\n",
+      ")\n",
+      "\n",
+      "/*\n",
+      "2 rows from Track table:\n",
+      "TrackId\tName\tAlbumId\tMediaTypeId\tGenreId\tComposer\tMilliseconds\tBytes\tUnitPrice\n",
+      "1\tFor Those About To Rock (We Salute You)\t1\t1\t1\tAngus Young, Malcolm Young, Brian Johnson\t343719\t11170334\t0.99\n",
+      "2\tBalls to the Wall\t2\t2\t1\tNone\t342562\t5510424\t0.99\n",
+      "*/\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(db.table_info)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Case 3: SQL agents\n",
+    "\n",
+    "LangChain has an SQL Agent which provides a more flexible way of interacting with SQL Databases than the `SQLDatabaseChain`.\n",
+    "\n",
+    "The main advantages of using the SQL Agent are:\n",
+    "\n",
+    "- It can answer questions based on the databases' schema as well as on the databases' content (like describing a specific table)\n",
+    "- It can recover from errors by running a generated query, catching the traceback and regenerating it correctly\n",
+    "\n",
+    "To initialize the agent, we use `create_sql_agent` function. \n",
+    "\n",
+    "This agent contains the `SQLDatabaseToolkit` which contains tools to: \n",
+    "\n",
+    "* Create and execute queries\n",
+    "* Check query syntax\n",
+    "* Retrieve table descriptions\n",
+    "* ... and more"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.agents import create_sql_agent\n",
+    "from langchain.agents.agent_toolkits import SQLDatabaseToolkit\n",
+    "# from langchain.agents import AgentExecutor\n",
+    "from langchain.agents.agent_types import AgentType\n",
+    "\n",
+    "db = SQLDatabase.from_uri(\"sqlite:///Chinook.db\")\n",
+    "llm = OpenAI(temperature=0, verbose=True)\n",
+    "\n",
+    "agent_executor = create_sql_agent(\n",
+    "    llm=OpenAI(temperature=0),\n",
+    "    toolkit=SQLDatabaseToolkit(db=db, llm=OpenAI(temperature=0)),\n",
+    "    verbose=True,\n",
+    "    agent_type=AgentType.ZERO_SHOT_REACT_DESCRIPTION,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Agent task example #1 - Running queries\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "\u001b[1m> Entering new AgentExecutor chain...\u001b[0m\n",
+      "\u001b[32;1m\u001b[1;3mAction: sql_db_list_tables\n",
+      "Action Input: \u001b[0m\n",
+      "Observation: \u001b[38;5;200m\u001b[1;3mAlbum, Artist, Customer, Employee, Genre, Invoice, InvoiceLine, MediaType, Playlist, PlaylistTrack, Track\u001b[0m\n",
+      "Thought:\u001b[32;1m\u001b[1;3m I should query the schema of the Invoice and Customer tables.\n",
+      "Action: sql_db_schema\n",
+      "Action Input: Invoice, Customer\u001b[0m\n",
+      "Observation: \u001b[33;1m\u001b[1;3m\n",
+      "CREATE TABLE \"Customer\" (\n",
+      "\t\"CustomerId\" INTEGER NOT NULL, \n",
+      "\t\"FirstName\" NVARCHAR(40) NOT NULL, \n",
+      "\t\"LastName\" NVARCHAR(20) NOT NULL, \n",
+      "\t\"Company\" NVARCHAR(80), \n",
+      "\t\"Address\" NVARCHAR(70), \n",
+      "\t\"City\" NVARCHAR(40), \n",
+      "\t\"State\" NVARCHAR(40), \n",
+      "\t\"Country\" NVARCHAR(40), \n",
+      "\t\"PostalCode\" NVARCHAR(10), \n",
+      "\t\"Phone\" NVARCHAR(24), \n",
+      "\t\"Fax\" NVARCHAR(24), \n",
+      "\t\"Email\" NVARCHAR(60) NOT NULL, \n",
+      "\t\"SupportRepId\" INTEGER, \n",
+      "\tPRIMARY KEY (\"CustomerId\"), \n",
+      "\tFOREIGN KEY(\"SupportRepId\") REFERENCES \"Employee\" (\"EmployeeId\")\n",
+      ")\n",
+      "\n",
+      "/*\n",
+      "3 rows from Customer table:\n",
+      "CustomerId\tFirstName\tLastName\tCompany\tAddress\tCity\tState\tCountry\tPostalCode\tPhone\tFax\tEmail\tSupportRepId\n",
+      "1\tLuís\tGonçalves\tEmbraer - Empresa Brasileira de Aeronáutica S.A.\tAv. Brigadeiro Faria Lima, 2170\tSão José dos Campos\tSP\tBrazil\t12227-000\t+55 (12) 3923-5555\t+55 (12) 3923-5566\tluisg@embraer.com.br\t3\n",
+      "2\tLeonie\tKöhler\tNone\tTheodor-Heuss-Straße 34\tStuttgart\tNone\tGermany\t70174\t+49 0711 2842222\tNone\tleonekohler@surfeu.de\t5\n",
+      "3\tFrançois\tTremblay\tNone\t1498 rue Bélanger\tMontréal\tQC\tCanada\tH2G 1A7\t+1 (514) 721-4711\tNone\tftremblay@gmail.com\t3\n",
+      "*/\n",
+      "\n",
+      "\n",
+      "CREATE TABLE \"Invoice\" (\n",
+      "\t\"InvoiceId\" INTEGER NOT NULL, \n",
+      "\t\"CustomerId\" INTEGER NOT NULL, \n",
+      "\t\"InvoiceDate\" DATETIME NOT NULL, \n",
+      "\t\"BillingAddress\" NVARCHAR(70), \n",
+      "\t\"BillingCity\" NVARCHAR(40), \n",
+      "\t\"BillingState\" NVARCHAR(40), \n",
+      "\t\"BillingCountry\" NVARCHAR(40), \n",
+      "\t\"BillingPostalCode\" NVARCHAR(10), \n",
+      "\t\"Total\" NUMERIC(10, 2) NOT NULL, \n",
+      "\tPRIMARY KEY (\"InvoiceId\"), \n",
+      "\tFOREIGN KEY(\"CustomerId\") REFERENCES \"Customer\" (\"CustomerId\")\n",
+      ")\n",
+      "\n",
+      "/*\n",
+      "3 rows from Invoice table:\n",
+      "InvoiceId\tCustomerId\tInvoiceDate\tBillingAddress\tBillingCity\tBillingState\tBillingCountry\tBillingPostalCode\tTotal\n",
+      "1\t2\t2009-01-01 00:00:00\tTheodor-Heuss-Straße 34\tStuttgart\tNone\tGermany\t70174\t1.98\n",
+      "2\t4\t2009-01-02 00:00:00\tUllevålsveien 14\tOslo\tNone\tNorway\t0171\t3.96\n",
+      "3\t8\t2009-01-03 00:00:00\tGrétrystraat 63\tBrussels\tNone\tBelgium\t1000\t5.94\n",
+      "*/\u001b[0m\n",
+      "Thought:\u001b[32;1m\u001b[1;3m I should query the total sales per country.\n",
+      "Action: sql_db_query\n",
+      "Action Input: SELECT Country, SUM(Total) AS TotalSales FROM Invoice INNER JOIN Customer ON Invoice.CustomerId = Customer.CustomerId GROUP BY Country ORDER BY TotalSales DESC LIMIT 10\u001b[0m\n",
+      "Observation: \u001b[36;1m\u001b[1;3m[('USA', 523.0600000000003), ('Canada', 303.9599999999999), ('France', 195.09999999999994), ('Brazil', 190.09999999999997), ('Germany', 156.48), ('United Kingdom', 112.85999999999999), ('Czech Republic', 90.24000000000001), ('Portugal', 77.23999999999998), ('India', 75.25999999999999), ('Chile', 46.62)]\u001b[0m\n",
+      "Thought:\u001b[32;1m\u001b[1;3m I now know the final answer\n",
+      "Final Answer: The country with the highest total sales is the USA, with a total of $523.06.\u001b[0m\n",
+      "\n",
+      "\u001b[1m> Finished chain.\u001b[0m\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "'The country with the highest total sales is the USA, with a total of $523.06.'"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agent_executor.run(\n",
+    "    \"List the total sales per country. Which country's customers spent the most?\"\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Looking at the [LangSmith trace](https://smith.langchain.com/public/a86dbe17-5782-4020-bce6-2de85343168a/r), we can see:\n",
+    "\n",
+    "* The agent is using a ReAct style prompt\n",
+    "* First, it will look at the tables: `Action: sql_db_list_tables` using tool `sql_db_list_tables`\n",
+    "* Given the tables as an observation, it `thinks` and then determinates the next `action`:\n",
+    "\n",
+    "```\n",
+    "Observation: Album, Artist, Customer, Employee, Genre, Invoice, InvoiceLine, MediaType, Playlist, PlaylistTrack, Track\n",
+    "Thought: I should query the schema of the Invoice and Customer tables.\n",
+    "Action: sql_db_schema\n",
+    "Action Input: Invoice, Customer\n",
+    "```\n",
+    "\n",
+    "* It then formulates the query using the schema from tool `sql_db_schema`\n",
+    "\n",
+    "```\n",
+    "Thought: I should query the total sales per country.\n",
+    "Action: sql_db_query\n",
+    "Action Input: SELECT Country, SUM(Total) AS TotalSales FROM Invoice INNER JOIN Customer ON Invoice.CustomerId = Customer.CustomerId GROUP BY Country ORDER BY TotalSales DESC LIMIT 10\n",
+    "```\n",
+    "\n",
+    "* It finally executes the generated query using tool `sql_db_query`\n",
+    "\n",
+    "![sql_usecase.png](/img/SQLDatabaseToolkit.png)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Agent task example #2 - Describing a Table"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "\u001b[1m> Entering new AgentExecutor chain...\u001b[0m\n",
+      "\u001b[32;1m\u001b[1;3mAction: sql_db_list_tables\n",
+      "Action Input: \u001b[0m\n",
+      "Observation: \u001b[38;5;200m\u001b[1;3mAlbum, Artist, Customer, Employee, Genre, Invoice, InvoiceLine, MediaType, Playlist, PlaylistTrack, Track\u001b[0m\n",
+      "Thought:\u001b[32;1m\u001b[1;3m The PlaylistTrack table is the most relevant to the question.\n",
+      "Action: sql_db_schema\n",
+      "Action Input: PlaylistTrack\u001b[0m\n",
+      "Observation: \u001b[33;1m\u001b[1;3m\n",
+      "CREATE TABLE \"PlaylistTrack\" (\n",
+      "\t\"PlaylistId\" INTEGER NOT NULL, \n",
+      "\t\"TrackId\" INTEGER NOT NULL, \n",
+      "\tPRIMARY KEY (\"PlaylistId\", \"TrackId\"), \n",
+      "\tFOREIGN KEY(\"TrackId\") REFERENCES \"Track\" (\"TrackId\"), \n",
+      "\tFOREIGN KEY(\"PlaylistId\") REFERENCES \"Playlist\" (\"PlaylistId\")\n",
+      ")\n",
+      "\n",
+      "/*\n",
+      "3 rows from PlaylistTrack table:\n",
+      "PlaylistId\tTrackId\n",
+      "1\t3402\n",
+      "1\t3389\n",
+      "1\t3390\n",
+      "*/\u001b[0m\n",
+      "Thought:\u001b[32;1m\u001b[1;3m I now know the final answer\n",
+      "Final Answer: The PlaylistTrack table contains two columns, PlaylistId and TrackId, which are both integers and form a primary key. It also has two foreign keys, one to the Track table and one to the Playlist table.\u001b[0m\n",
+      "\n",
+      "\u001b[1m> Finished chain.\u001b[0m\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "'The PlaylistTrack table contains two columns, PlaylistId and TrackId, which are both integers and form a primary key. It also has two foreign keys, one to the Track table and one to the Playlist table.'"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agent_executor.run(\"Describe the playlisttrack table\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Go deeper\n",
+    "\n",
+    "To learn more about the SQL Agent and how it works we refer to the [SQL Agent Toolkit](/docs/integrations/toolkits/sql_database) documentation.\n",
+    "\n",
+    "You can also check Agents for other document types:\n",
+    "- [Pandas Agent](/docs/integrations/toolkits/pandas.html)\n",
+    "- [CSV Agent](/docs/integrations/toolkits/csv.html)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Elastic Search\n",
+    "\n",
+    "Going beyond the above use-case, there are integrations with other databases.\n",
+    "\n",
+    "For example, we can interact with Elasticsearch analytics database. \n",
+    "\n",
+    "This chain builds search queries via the Elasticsearch DSL API (filters and aggregations).\n",
+    "\n",
+    "The Elasticsearch client must have permissions for index listing, mapping description and search queries.\n",
+    "\n",
+    "See [here](https://www.elastic.co/guide/en/elasticsearch/reference/current/docker.html) for instructions on how to run Elasticsearch locally.\n",
+    "\n",
+    "Make sure to install the Elasticsearch Python client before:\n",
+    "\n",
+    "```sh\n",
+    "pip install elasticsearch\n",
+    "```"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from elasticsearch import Elasticsearch\n",
+    "from langchain.chat_models import ChatOpenAI\n",
+    "from langchain.chains.elasticsearch_database import ElasticsearchDatabaseChain"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Initialize Elasticsearch python client.\n",
+    "# See https://elasticsearch-py.readthedocs.io/en/v8.8.2/api.html#elasticsearch.Elasticsearch\n",
+    "ELASTIC_SEARCH_SERVER = \"https://elastic:pass@localhost:9200\"\n",
+    "db = Elasticsearch(ELASTIC_SEARCH_SERVER)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Uncomment the next cell to initially populate your db."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# customers = [\n",
+    "#     {\"firstname\": \"Jennifer\", \"lastname\": \"Walters\"},\n",
+    "#     {\"firstname\": \"Monica\",\"lastname\":\"Rambeau\"},\n",
+    "#     {\"firstname\": \"Carol\",\"lastname\":\"Danvers\"},\n",
+    "#     {\"firstname\": \"Wanda\",\"lastname\":\"Maximoff\"},\n",
+    "#     {\"firstname\": \"Jennifer\",\"lastname\":\"Takeda\"},\n",
+    "# ]\n",
+    "# for i, customer in enumerate(customers):\n",
+    "#     db.create(index=\"customers\", document=customer, id=i)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "llm = ChatOpenAI(model_name=\"gpt-4\", temperature=0)\n",
+    "chain = ElasticsearchDatabaseChain.from_llm(llm=llm, database=db, verbose=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "question = \"What are the first names of all the customers?\"\n",
+    "chain.run(question)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "We can customize the prompt."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.chains.elasticsearch_database.prompts import DEFAULT_DSL_TEMPLATE\n",
+    "from langchain.prompts.prompt import PromptTemplate\n",
+    "\n",
+    "PROMPT_TEMPLATE = \"\"\"Given an input question, create a syntactically correct Elasticsearch query to run. Unless the user specifies in their question a specific number of examples they wish to obtain, always limit your query to at most {top_k} results. You can order the results by a relevant column to return the most interesting examples in the database.\n",
+    "\n",
+    "Unless told to do not query for all the columns from a specific index, only ask for a the few relevant columns given the question.\n",
+    "\n",
+    "Pay attention to use only the column names that you can see in the mapping description. Be careful to not query for columns that do not exist. Also, pay attention to which column is in which index. Return the query as valid json.\n",
+    "\n",
+    "Use the following format:\n",
+    "\n",
+    "Question: Question here\n",
+    "ESQuery: Elasticsearch Query formatted as json\n",
+    "\"\"\"\n",
+    "\n",
+    "PROMPT = PromptTemplate.from_template(\n",
+    "    PROMPT_TEMPLATE,\n",
+    ")\n",
+    "chain = ElasticsearchDatabaseChain.from_llm(llm=llm, database=db, query_prompt=PROMPT)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

docs/extras/use_cases/tabular/elasticsearch_database.ipynb

@@ -1,206 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "id": "dd7ec7af",
-   "metadata": {},
-   "source": [
-    "# Elasticsearch database\n",
-    "\n",
-    "Interact with Elasticsearch analytics database via Langchain. This chain builds search queries via the Elasticsearch DSL API (filters and aggregations).\n",
-    "\n",
-    "The Elasticsearch client must have permissions for index listing, mapping description and search queries.\n",
-    "\n",
-    "See [here](https://www.elastic.co/guide/en/elasticsearch/reference/current/docker.html) for instructions on how to run Elasticsearch locally.\n",
-    "\n",
-    "Make sure to install the Elasticsearch Python client before:\n",
-    "\n",
-    "```sh\n",
-    "pip install elasticsearch\n",
-    "```"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 11,
-   "id": "dd8eae75",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from elasticsearch import Elasticsearch\n",
-    "\n",
-    "from langchain.chains.elasticsearch_database import ElasticsearchDatabaseChain\n",
-    "from langchain.chat_models import ChatOpenAI"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "id": "5cde03bc",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Initialize Elasticsearch python client.\n",
-    "# See https://elasticsearch-py.readthedocs.io/en/v8.8.2/api.html#elasticsearch.Elasticsearch\n",
-    "ELASTIC_SEARCH_SERVER = \"https://elastic:pass@localhost:9200\"\n",
-    "db = Elasticsearch(ELASTIC_SEARCH_SERVER)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "74a41374",
-   "metadata": {},
-   "source": [
-    "Uncomment the next cell to initially populate your db."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 10,
-   "id": "430ada0f",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# customers = [\n",
-    "#     {\"firstname\": \"Jennifer\", \"lastname\": \"Walters\"},\n",
-    "#     {\"firstname\": \"Monica\",\"lastname\":\"Rambeau\"},\n",
-    "#     {\"firstname\": \"Carol\",\"lastname\":\"Danvers\"},\n",
-    "#     {\"firstname\": \"Wanda\",\"lastname\":\"Maximoff\"},\n",
-    "#     {\"firstname\": \"Jennifer\",\"lastname\":\"Takeda\"},\n",
-    "# ]\n",
-    "# for i, customer in enumerate(customers):\n",
-    "#     db.create(index=\"customers\", document=customer, id=i)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 12,
-   "id": "f36ae0d8",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "llm = ChatOpenAI(model_name=\"gpt-4\", temperature=0)\n",
-    "chain = ElasticsearchDatabaseChain.from_llm(llm=llm, database=db, verbose=True)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 13,
-   "id": "b5d22d9d",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "\n",
-      "\n",
-      "\u001b[1m> Entering new ElasticsearchDatabaseChain chain...\u001b[0m\n",
-      "What are the first names of all the customers?\n",
-      "ESQuery:\u001b[32;1m\u001b[1;3m{'size': 10, 'query': {'match_all': {}}, '_source': ['firstname']}\u001b[0m\n",
-      "ESResult: \u001b[33;1m\u001b[1;3m{'took': 5, 'timed_out': False, '_shards': {'total': 1, 'successful': 1, 'skipped': 0, 'failed': 0}, 'hits': {'total': {'value': 6, 'relation': 'eq'}, 'max_score': 1.0, 'hits': [{'_index': 'customers', '_id': '0', '_score': 1.0, '_source': {'firstname': 'Jennifer'}}, {'_index': 'customers', '_id': '1', '_score': 1.0, '_source': {'firstname': 'Monica'}}, {'_index': 'customers', '_id': '2', '_score': 1.0, '_source': {'firstname': 'Carol'}}, {'_index': 'customers', '_id': '3', '_score': 1.0, '_source': {'firstname': 'Wanda'}}, {'_index': 'customers', '_id': '4', '_score': 1.0, '_source': {'firstname': 'Jennifer'}}, {'_index': 'customers', '_id': 'firstname', '_score': 1.0, '_source': {'firstname': 'Jennifer'}}]}}\u001b[0m\n",
-      "Answer:\u001b[32;1m\u001b[1;3mThe first names of all the customers are Jennifer, Monica, Carol, Wanda, and Jennifer.\u001b[0m\n",
-      "\u001b[1m> Finished chain.\u001b[0m\n"
-     ]
-    },
-    {
-     "data": {
-      "text/plain": [
-       "'The first names of all the customers are Jennifer, Monica, Carol, Wanda, and Jennifer.'"
-      ]
-     },
-     "execution_count": 13,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "question = \"What are the first names of all the customers?\"\n",
-    "chain.run(question)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9b4bfada",
-   "metadata": {},
-   "source": [
-    "## Custom prompt\n",
-    "\n",
-    "For best results you'll likely need to customize the prompt."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "id": "0a494f5b",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from langchain.chains.elasticsearch_database.prompts import DEFAULT_DSL_TEMPLATE\n",
-    "from langchain.prompts.prompt import PromptTemplate\n",
-    "\n",
-    "PROMPT_TEMPLATE = \"\"\"Given an input question, create a syntactically correct Elasticsearch query to run. Unless the user specifies in their question a specific number of examples they wish to obtain, always limit your query to at most {top_k} results. You can order the results by a relevant column to return the most interesting examples in the database.\n",
-    "\n",
-    "Unless told to do not query for all the columns from a specific index, only ask for a the few relevant columns given the question.\n",
-    "\n",
-    "Pay attention to use only the column names that you can see in the mapping description. Be careful to not query for columns that do not exist. Also, pay attention to which column is in which index. Return the query as valid json.\n",
-    "\n",
-    "Use the following format:\n",
-    "\n",
-    "Question: Question here\n",
-    "ESQuery: Elasticsearch Query formatted as json\n",
-    "\"\"\"\n",
-    "\n",
-    "PROMPT = PromptTemplate.from_template(\n",
-    "    PROMPT_TEMPLATE,\n",
-    ")\n",
-    "chain = ElasticsearchDatabaseChain.from_llm(llm=llm, database=db, query_prompt=PROMPT)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "372b8f93",
-   "metadata": {},
-   "source": [
-    "## Adding example rows from each index\n",
-    "\n",
-    "Sometimes, the format of the data is not obvious and it is optimal to include a sample of rows from the indices in the prompt to allow the LLM to understand the data before providing a final query. Here we will use this feature to let the LLM know that artists are saved with their full names by providing ten rows from the index."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "eef818de",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "chain = ElasticsearchDatabaseChain.from_llm(\n",
-    "    llm=ChatOpenAI(temperature=0),\n",
-    "    database=db,\n",
-    "    sample_documents_in_index_info=2,  # 2 rows from each index will be included in the prompt as sample data\n",
-    ")"
-   ]
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "venv",
-   "language": "python",
-   "name": "venv"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.11.3"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 5
-}

docs/extras/use_cases/tabular/index.mdx

@@ -1,35 +0,0 @@
----
-sidebar_position: 1
----
-
-# Analyzing structured data
-
-Lots of data and information is stored in tabular data, whether it be csvs, excel sheets, or SQL tables.
-This page covers all resources available in LangChain for working with data in this format.
-
-## Document loading
-If you have text data stored in a tabular format, you may want to load the data into a Document and then index it as you would
-other text/unstructured data. For this, you should use a document loader like the [CSVLoader](/docs/modules/data_connection/document_loaders/how_to/csv.html)
-and then you should [create an index](/docs/modules/data_connection) over that data, and [query it that way](/docs/use_cases/question_answering/how_to/vector_db_qa.html).
-
-## Querying
-If you have more numeric tabular data, or have a large amount of data and don't want to index it, you should get started
-by looking at various chains and agents we have for dealing with this data.
-
-### Chains
-
-If you are just getting started, and you have relatively small/simple tabular data, you should get started with chains.
-Chains are a sequence of predetermined steps, so they are good to get started with as they give you more control and let you 
-understand what is happening better.
-
-- [SQL Database Chain](/docs/use_cases/tabular/sqlite.html)
-
-### Agents
-
-Agents are more complex, and involve multiple queries to the LLM to understand what to do.
-The downside of agents are that you have less control. The upside is that they are more powerful,
-which allows you to use them on larger databases and more complex schemas. 
-
-- [SQL Agent](/docs/integrations/toolkits/sql_database.html)
-- [Pandas Agent](/docs/integrations/toolkits/pandas.html)
-- [CSV Agent](/docs/integrations/toolkits/csv.html)

docs/extras/use_cases/tabular/sql_query.ipynb

@@ -1,125 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "id": "c04293ac",
-   "metadata": {},
-   "source": [
-    "# SQL Query\n",
-    "\n",
-    "This notebook walks through how to load and run a chain that constructs SQL queries that can be run against your database to answer a question. Note that this ONLY constructs the query and does not run it."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "id": "e9063a93",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from langchain.chains import create_sql_query_chain\n",
-    "\n",
-    "from langchain.chat_models import ChatOpenAI\n",
-    "from langchain.utilities import SQLDatabase"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "id": "a1ff5cee",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "db = SQLDatabase.from_uri(\"sqlite:///../../../../notebooks/Chinook.db\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "id": "cb04579f",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "chain = create_sql_query_chain(ChatOpenAI(temperature=0), db)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 12,
-   "id": "744e6210",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "response = chain.invoke({\"question\":\"How many employees are there\"})"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 14,
-   "id": "28f984f1",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "SELECT COUNT(*) FROM Employee\n"
-     ]
-    }
-   ],
-   "source": [
-    "print(response)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 15,
-   "id": "08de511c",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "'[(8,)]'"
-      ]
-     },
-     "execution_count": 15,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "db.run(response)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "e3a006a7",
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.9.1"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 5
-}

docs/snippets/modules/agents/get_started.mdx

@@ -44,6 +44,7 @@ This allows for a few different ways to customize, including passing in a custom
 
 ```python
 from langchain.schema import SystemMessage
+from langchain.agents import OpenAIFunctionsAgent
 system_message = SystemMessage(content="You are very powerful assistant, but bad at calculating lengths of words.")
 prompt = OpenAIFunctionsAgent.create_prompt(system_message=system_message)
 ```
@@ -51,7 +52,6 @@ prompt = OpenAIFunctionsAgent.create_prompt(system_message=system_message)
 Putting those pieces together, we can now create the agent.
 
 ```python
-from langchain.agents import OpenAIFunctionsAgent
 agent = OpenAIFunctionsAgent(llm=llm, tools=tools, prompt=prompt)
 ```
 

docs/snippets/modules/chains/popular/chat_vector_db.mdx

@@ -8,7 +8,6 @@ from langchain.chains import ConversationalRetrievalChain
 
 Load in documents. You can replace this with a loader for whatever type of data you want
 
-
 ```python
 from langchain.document_loaders import TextLoader
 loader = TextLoader("../../state_of_the_union.txt")
@@ -17,7 +16,6 @@ documents = loader.load()
 
 If you had multiple loaders that you wanted to combine, you do something like:
 
-
 ```python
 # loaders = [....]
 # docs = []
@@ -27,7 +25,6 @@ If you had multiple loaders that you wanted to combine, you do something like:
 
 We now split the documents, create embeddings for them, and put them in a vectorstore. This allows us to do semantic search over them.
 
-
 ```python
 text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
 documents = text_splitter.split_documents(documents)
@@ -46,7 +43,6 @@ vectorstore = Chroma.from_documents(documents, embeddings)
 
 We can now create a memory object, which is necessary to track the inputs/outputs and hold a conversation.
 
-
 ```python
 from langchain.memory import ConversationBufferMemory
 memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
@@ -54,18 +50,15 @@ memory = ConversationBufferMemory(memory_key="chat_history", return_messages=Tru
 
 We now initialize the `ConversationalRetrievalChain`
 
-
 ```python
 qa = ConversationalRetrievalChain.from_llm(OpenAI(temperature=0), vectorstore.as_retriever(), memory=memory)
 ```
 
-
 ```python
 query = "What did the president say about Ketanji Brown Jackson"
 result = qa({"question": query})
 ```
 
-
 ```python
 result["answer"]
 ```
@@ -78,13 +71,11 @@ result["answer"]
 
 </CodeOutputBlock>
 
-
 ```python
 query = "Did he mention who she succeeded"
 result = qa({"question": query})
 ```
 
-
 ```python
 result['answer']
 ```
@@ -101,21 +92,18 @@ result['answer']
 
 In the above example, we used a Memory object to track chat history. We can also just pass it in explicitly. In order to do this, we need to initialize a chain without any memory object.
 
-
 ```python
 qa = ConversationalRetrievalChain.from_llm(OpenAI(temperature=0), vectorstore.as_retriever())
 ```
 
 Here's an example of asking a question with no chat history
 
-
 ```python
 chat_history = []
 query = "What did the president say about Ketanji Brown Jackson"
 result = qa({"question": query, "chat_history": chat_history})
 ```
 
-
 ```python
 result["answer"]
 ```
@@ -130,14 +118,12 @@ result["answer"]
 
 Here's an example of asking a question with some chat history
 
-
 ```python
 chat_history = [(query, result["answer"])]
 query = "Did he mention who she succeeded"
 result = qa({"question": query, "chat_history": chat_history})
 ```
 
-
 ```python
 result['answer']
 ```
@@ -154,12 +140,10 @@ result['answer']
 
 This chain has two steps. First, it condenses the current question and the chat history into a standalone question. This is necessary to create a standanlone vector to use for retrieval. After that, it does retrieval and then answers the question using retrieval augmented generation with a separate model. Part of the power of the declarative nature of LangChain is that you can easily use a separate language model for each call. This can be useful to use a cheaper and faster model for the simpler task of condensing the question, and then a more expensive model for answering the question. Here is an example of doing so.
 
-
 ```python
 from langchain.chat_models import ChatOpenAI
 ```
 
-
 ```python
 qa = ConversationalRetrievalChain.from_llm(
     ChatOpenAI(temperature=0, model="gpt-4"),
@@ -168,36 +152,90 @@ qa = ConversationalRetrievalChain.from_llm(
 )
 ```
 
-
 ```python
 chat_history = []
 query = "What did the president say about Ketanji Brown Jackson"
 result = qa({"question": query, "chat_history": chat_history})
 ```
 
-
 ```python
 chat_history = [(query, result["answer"])]
 query = "Did he mention who she succeeded"
 result = qa({"question": query, "chat_history": chat_history})
 ```
 
-## Return Source Documents
-You can also easily return source documents from the ConversationalRetrievalChain. This is useful for when you want to inspect what documents were returned.
+## Using a custom prompt for condensing the question
 
+By default, ConversationalRetrievalQA uses CONDENSE_QUESTION_PROMPT to condense a question. Here is the implementation of this in the docs
+
+```python
+from langchain.prompts.prompt import PromptTemplate
+
+_template = """Given the following conversation and a follow up question, rephrase the follow up question to be a standalone question, in its original language.
+
+Chat History:
+{chat_history}
+Follow Up Input: {question}
+Standalone question:"""
+CONDENSE_QUESTION_PROMPT = PromptTemplate.from_template(_template)
+
+```
+
+But instead of this any custom template can be used to further augment information in the question or instruct the LLM to do something. Here is an example
+
+```python
+from langchain.prompts.prompt import PromptTemplate
+```
+
+```python
+custom_template = """Given the following conversation and a follow up question, rephrase the follow up question to be a standalone question. At the end of standalone question add this 'Answer the question in German language.' If you do not know the answer reply with 'I am sorry'.
+Chat History:
+{chat_history}
+Follow Up Input: {question}
+Standalone question:"""
+```
+
+```python
+CUSTOM_QUESTION_PROMPT = PromptTemplate.from_template(custom_template)
+```
+
+```python
+model = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0.3)
+embeddings = OpenAIEmbeddings()
+vectordb = Chroma(embedding_function=embeddings, persist_directory=directory)
+memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
+qa = ConversationalRetrievalChain.from_llm(
+    model,
+    vectordb.as_retriever(),
+    condense_question_prompt=CUSTOM_QUESTION_PROMPT,
+    memory=memory
+)
+```
+
+```python
+query = "What did the president say about Ketanji Brown Jackson"
+result = qa({"question": query})
+```
+
+```python
+query = "Did he mention who she succeeded"
+result = qa({"question": query})
+```
+
+## Return Source Documents
+
+You can also easily return source documents from the ConversationalRetrievalChain. This is useful for when you want to inspect what documents were returned.
 
 ```python
 qa = ConversationalRetrievalChain.from_llm(OpenAI(temperature=0), vectorstore.as_retriever(), return_source_documents=True)
 ```
 
-
 ```python
 chat_history = []
 query = "What did the president say about Ketanji Brown Jackson"
 result = qa({"question": query, "chat_history": chat_history})
 ```
 
-
 ```python
 result['source_documents'][0]
 ```
@@ -211,14 +249,13 @@ result['source_documents'][0]
 </CodeOutputBlock>
 
 ## ConversationalRetrievalChain with `search_distance`
-If you are using a vector store that supports filtering by search distance, you can add a threshold value parameter.
 
+If you are using a vector store that supports filtering by search distance, you can add a threshold value parameter.
 
 ```python
 vectordbkwargs = {"search_distance": 0.9}
 ```
 
-
 ```python
 qa = ConversationalRetrievalChain.from_llm(OpenAI(temperature=0), vectorstore.as_retriever(), return_source_documents=True)
 chat_history = []
@@ -227,8 +264,8 @@ result = qa({"question": query, "chat_history": chat_history, "vectordbkwargs":
 ```
 
 ## ConversationalRetrievalChain with `map_reduce`
-We can also use different types of combine document chains with the ConversationalRetrievalChain chain.
 
+We can also use different types of combine document chains with the ConversationalRetrievalChain chain.
 
 ```python
 from langchain.chains import LLMChain
@@ -236,7 +273,6 @@ from langchain.chains.question_answering import load_qa_chain
 from langchain.chains.conversational_retrieval.prompts import CONDENSE_QUESTION_PROMPT
 ```
 
-
 ```python
 llm = OpenAI(temperature=0)
 question_generator = LLMChain(llm=llm, prompt=CONDENSE_QUESTION_PROMPT)
@@ -249,14 +285,12 @@ chain = ConversationalRetrievalChain(
 )
 ```
 
-
 ```python
 chat_history = []
 query = "What did the president say about Ketanji Brown Jackson"
 result = chain({"question": query, "chat_history": chat_history})
 ```
 
-
 ```python
 result['answer']
 ```
@@ -273,12 +307,10 @@ result['answer']
 
 You can also use this chain with the question answering with sources chain.
 
-
 ```python
 from langchain.chains.qa_with_sources import load_qa_with_sources_chain
 ```
 
-
 ```python
 llm = OpenAI(temperature=0)
 question_generator = LLMChain(llm=llm, prompt=CONDENSE_QUESTION_PROMPT)
@@ -291,14 +323,12 @@ chain = ConversationalRetrievalChain(
 )
 ```
 
-
 ```python
 chat_history = []
 query = "What did the president say about Ketanji Brown Jackson"
 result = chain({"question": query, "chat_history": chat_history})
 ```
 
-
 ```python
 result['answer']
 ```
@@ -315,7 +345,6 @@ result['answer']
 
 Output from the chain will be streamed to `stdout` token by token in this example.
 
-
 ```python
 from langchain.chains.llm import LLMChain
 from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
@@ -334,7 +363,6 @@ qa = ConversationalRetrievalChain(
     retriever=vectorstore.as_retriever(), combine_docs_chain=doc_chain, question_generator=question_generator)
 ```
 
-
 ```python
 chat_history = []
 query = "What did the president say about Ketanji Brown Jackson"
@@ -349,7 +377,6 @@ result = qa({"question": query, "chat_history": chat_history})
 
 </CodeOutputBlock>
 
-
 ```python
 chat_history = [(query, result["answer"])]
 query = "Did he mention who she succeeded"
@@ -365,8 +392,8 @@ result = qa({"question": query, "chat_history": chat_history})
 </CodeOutputBlock>
 
 ## get_chat_history Function
-You can also specify a `get_chat_history` function, which can be used to format the chat_history string.
 
+You can also specify a `get_chat_history` function, which can be used to format the chat_history string.
 
 ```python
 def get_chat_history(inputs) -> str:
@@ -377,14 +404,12 @@ def get_chat_history(inputs) -> str:
 qa = ConversationalRetrievalChain.from_llm(OpenAI(temperature=0), vectorstore.as_retriever(), get_chat_history=get_chat_history)
 ```
 
-
 ```python
 chat_history = []
 query = "What did the president say about Ketanji Brown Jackson"
 result = qa({"question": query, "chat_history": chat_history})
 ```
 
-
 ```python
 result['answer']
 ```

docs/snippets/modules/chains/popular/vector_db_qa.mdx

@@ -117,3 +117,38 @@ qa.run(query)
 ```
 
 </CodeOutputBlock>
+
+## Vectorstore Retriever Options
+You can adjust how documents are retrieved from your vectorstore depending on the specific task.
+
+There are two main ways to retrieve documents relevant to a query- Similarity Search and Max Marginal Relevance Search (MMR Search). Similarity Search is the default, but you can use MMR by adding the `search_type` parameter:
+
+```python
+docsearch.as_retriever(search_type="mmr")
+```
+
+You can also modify the search by passing specific search arguments through the retriever to the search function, using the `search_kwargs` keyword argument.
+
+- `k` defines how many documents are returned; defaults to 4.
+- `score_threshold` allows you to set a minimum relevance for documents returned by the retriever, if you are using the "similarity_score_threshold" search type.
+- `fetch_k` determines the amount of documents to pass to the MMR algorithm; defaults to 20. 
+- `lambda_mult` controls the diversity of results returned by the MMR algorithm, with 1 being minimum diversity and 0 being maximum. Defaults to 0.5.
+- `filter` allows you to define a filter on what documents should be retrieved, based on the documents' metadata. This has no effect if the Vectorstore doesn't store any metadata.
+
+Some examples for how these parameters can be used:
+```python
+# Retrieve more documents with higher diversity- useful if your dataset has many similar documents
+docsearch.as_retriever(search_type="mmr", search_kwargs={'k': 6, 'lambda_mult': 0.25})
+
+# Fetch more documents for the MMR algorithm to consider, but only return the top 5
+docsearch.as_retriever(search_type="mmr", search_kwargs={'k': 5, 'fetch_k': 50})
+
+# Only retrieve documents that have a relevance score above a certain threshold
+docsearch.as_retriever(search_type="similarity_score_threshold", search_kwargs={'score_threshold': 0.8})
+
+# Only get the single most similar document from the dataset
+docsearch.as_retriever(search_kwargs={'k': 1})
+
+# Use a filter to only retrieve documents from a specific paper 
+docsearch.as_retriever(search_kwargs={'filter': {'paper_title':'GPT-4 Technical Report'}})
+```

docs/snippets/modules/chains/popular/vector_db_qa_with_sources.mdx

@@ -3,7 +3,7 @@ Additionally, we can return the source documents used to answer the question by
 
 
 ```python
-qa = RetrievalQA.from_chain_type(llm=OpenAI(), chain_type="stuff", retriever=docsearch.as_retriever(), return_source_documents=True)
+qa = RetrievalQA.from_chain_type(llm=OpenAI(), chain_type="stuff", retriever=docsearch.as_retriever(search_type="mmr", search_kwargs={'fetch_k': 30}), return_source_documents=True)
 ```
 
 

docs/snippets/modules/data_connection/retrievers/get_started.mdx

@@ -159,6 +159,22 @@ index.vectorstore.as_retriever()
 
 </CodeOutputBlock>
 
+It can also be convenient to filter the vectorstore by the metadata associated with documents, particularly when your vectorstore has multiple sources.  This can be done using the `query` method like so:
+
+
+```python
+index.query("Summarize the general content of this document.", retriever_kwargs={"search_kwargs": {"filter": {"source": "../state_of_the_union.txt"}}})
+```
+
+<CodeOutputBlock lang="python">
+
+```
+    " The document is a speech given by President Trump to the nation on the occasion of his 245th birthday. The speech highlights the importance of American values and the challenges facing the country, including the ongoing conflict in Ukraine, the ongoing trade war with China, and the ongoing conflict in Syria. The speech also discusses the importance of investing in emerging technologies and American manufacturing, and calls on Congress to pass the Bipartisan Innovation Act and other important legislation."
+```
+
+</CodeOutputBlock>
+
+
 ## Walkthrough
 
 Okay, so what's actually going on? How is this index getting created?

docs/snippets/modules/data_connection/retrievers/self_query/get_started.mdx

@@ -3,7 +3,7 @@ We'll use a Pinecone vector store in this example.
 
 First we'll want to create a `Pinecone` VectorStore and seed it with some data. We've created a small demo set of documents that contain summaries of movies.
 
-To use Pinecone, you to have `pinecone` package installed and you must have an API key and an Environment. Here are the [installation instructions](https://docs.pinecone.io/docs/quickstart).
+To use Pinecone, you need to have `pinecone` package installed and you must have an API key and an Environment. Here are the [installation instructions](https://docs.pinecone.io/docs/quickstart).
 
 NOTE: The self-query retriever requires you to have `lark` package installed.
 
@@ -85,7 +85,7 @@ retriever = SelfQueryRetriever.from_llm(llm, vectorstore, document_content_descr
 ```
 
 ## Testing it out
-And now we can try actually using our retriever!
+And now we can actually try using our retriever!
 
 
 ```python
@@ -198,4 +198,4 @@ retriever = SelfQueryRetriever.from_llm(
 ```python
 # This example only specifies a relevant query
 retriever.get_relevant_documents("What are two movies about dinosaurs")
-```
+```

docs/snippets/modules/model_io/prompts/prompt_templates/get_started.mdx

@@ -1,140 +1,115 @@
-Here's the simplest example:
+Typically, language models expect the prompt to either be a string or else a list of chat messages.
+
+## Prompt template
+
+Use `PromptTemplate` to create a template for a string prompt.
+
+By default, `PromptTemplate` uses [Python's str.format](https://docs.python.org/3/library/stdtypes.html#str.format)
+syntax for templating; however other templating syntax is available (e.g., `jinja2`).
 
 ```python
 from langchain import PromptTemplate
 
-
-template = """\
-You are a naming consultant for new companies.
-What is a good name for a company that makes {product}?
-"""
-
-prompt = PromptTemplate.from_template(template)
-prompt.format(product="colorful socks")
+prompt_template = PromptTemplate.from_template(
+    "Tell me a {adjective} joke about {content}."
+)
+prompt_template.format(adjective="funny", content="chickens")
 ```
 
 <CodeOutputBlock lang="python">
 
 ```
-You are a naming consultant for new companies.
-What is a good name for a company that makes colorful socks?
+"Tell me a funny joke about chickens."
 ```
 
 </CodeOutputBlock>
 
-
-## Create a prompt template
-
-You can create simple hardcoded prompts using the `PromptTemplate` class. Prompt templates can take any number of input variables, and can be formatted to generate a prompt.
-
+The template supports any number of variables, including no variables:
 
 ```python
 from langchain import PromptTemplate
 
-# An example prompt with no input variables
-no_input_prompt = PromptTemplate(input_variables=[], template="Tell me a joke.")
-no_input_prompt.format()
-# -> "Tell me a joke."
-
-# An example prompt with one input variable
-one_input_prompt = PromptTemplate(input_variables=["adjective"], template="Tell me a {adjective} joke.")
-one_input_prompt.format(adjective="funny")
-# -> "Tell me a funny joke."
-
-# An example prompt with multiple input variables
-multiple_input_prompt = PromptTemplate(
-    input_variables=["adjective", "content"], 
-    template="Tell me a {adjective} joke about {content}."
+prompt_template = PromptTemplate.from_template(
+"Tell me a joke"
 )
-multiple_input_prompt.format(adjective="funny", content="chickens")
-# -> "Tell me a funny joke about chickens."
+prompt_template.format()
 ```
 
-If you do not wish to specify `input_variables` manually, you can also create a `PromptTemplate` using `from_template` class method. `langchain` will automatically infer the `input_variables` based on the `template` passed.
+For additional validation, specify `input_variables` explicitly. These variables
+will be compared against the variables present in the template string during instantiation, raising an exception if
+there is a mismatch; for example,
 
 ```python
-template = "Tell me a {adjective} joke about {content}."
+from langchain import PromptTemplate
 
-prompt_template = PromptTemplate.from_template(template)
-prompt_template.input_variables
-# -> ['adjective', 'content']
-prompt_template.format(adjective="funny", content="chickens")
-# -> Tell me a funny joke about chickens.
+invalid_prompt = PromptTemplate(
+    input_variables=["adjective"],
+    template="Tell me a {adjective} joke about {content}."
+)
 ```
 
-You can create custom prompt templates that format the prompt in any way you want. For more information, see [Custom Prompt Templates](./custom_prompt_template.html).
-
+You can create custom prompt templates that format the prompt in any way you want.
+For more information, see [Custom Prompt Templates](./custom_prompt_template.html).
 
 <!-- TODO(shreya): Add link to Jinja -->
 
 ## Chat prompt template
 
-[Chat Models](../models/chat) take a list of chat messages as input - this list commonly referred to as a `prompt`.
-These chat messages differ from raw string (which you would pass into a [LLM](/docs/modules/model_io/models/llms) model) in that every message is associated with a `role`.
-
-For example, in OpenAI [Chat Completion API](https://platform.openai.com/docs/guides/chat/introduction), a chat message can be associated with the AI, human or system role. The model is supposed to follow instruction from system chat message more closely.
-
-LangChain provides several prompt templates to make constructing and working with prompts easy. You are encouraged to use these chat related prompt templates instead of `PromptTemplate` when querying chat models to fully utilize the potential of the underlying chat model.
-
-
+The prompt to [Chat Models](../models/chat) is a list of chat messages.
 
+Each chat message is associated with content, and an additional parameter called `role`.
+For example, in the OpenAI [Chat Completions API](https://platform.openai.com/docs/guides/chat/introduction), a chat message can be associated with an AI assistant, a human or a system role.
 
+Create a chat prompt template like this:
 
 ```python
-from langchain.prompts import (
-    ChatPromptTemplate,
-    PromptTemplate,
-    SystemMessagePromptTemplate,
-    AIMessagePromptTemplate,
-    HumanMessagePromptTemplate,
-)
-from langchain.schema import (
-    AIMessage,
-    HumanMessage,
-    SystemMessage
+from langchain.prompts import ChatPromptTemplate
+
+template = ChatPromptTemplate.from_messages([
+    ("system", "You are a helpful AI bot. Your name is {name}."),
+    ("human", "Hello, how are you doing?"),
+    ("ai", "I'm doing well, thanks!"),
+    ("human", "{user_input}"),
+])
+
+messages = template.format_messages(
+    name="Bob",
+    user_input="What is your name?"
 )
 ```
 
-To create a message template associated with a role, you use `MessagePromptTemplate`.
-
-For convenience, there is a `from_template` method exposed on the template. If you were to use this template, this is what it would look like:
+`ChatPromptTemplate.from_messages` accepts a variety of message representations.
 
+For example, in addition to using the 2-tuple representation of (type, content) used
+above, you could pass in an instance of `MessagePromptTemplate` or `BaseMessage`.
 
 ```python
-template="You are a helpful assistant that translates {input_language} to {output_language}."
-system_message_prompt = SystemMessagePromptTemplate.from_template(template)
-human_template="{text}"
-human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
-```
+from langchain.prompts import ChatPromptTemplate
+from langchain.prompts.chat import SystemMessage, HumanMessagePromptTemplate
 
-If you wanted to construct the `MessagePromptTemplate` more directly, you could create a PromptTemplate outside and then pass it in, eg:
-
-
-```python
-prompt=PromptTemplate(
-    template="You are a helpful assistant that translates {input_language} to {output_language}.",
-    input_variables=["input_language", "output_language"],
+template = ChatPromptTemplate.from_messages(
+    [
+        SystemMessage(
+            content=(
+                "You are a helpful assistant that re-writes the user's text to "
+                "sound more upbeat."
+            )
+        ),
+        HumanMessagePromptTemplate.from_template("{text}"),
+    ]
 )
-system_message_prompt_2 = SystemMessagePromptTemplate(prompt=prompt)
 
-assert system_message_prompt == system_message_prompt_2
-```
+from langchain.chat_models import ChatOpenAI
 
-After that, you can build a `ChatPromptTemplate` from one or more `MessagePromptTemplates`. You can use `ChatPromptTemplate`'s `format_prompt` -- this returns a `PromptValue`, which you can convert to a string or Message object, depending on whether you want to use the formatted value as input to an llm or chat model.
-
-
-```python
-chat_prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
-
-# get a chat completion from the formatted messages
-chat_prompt.format_prompt(input_language="English", output_language="French", text="I love programming.").to_messages()
+llm = ChatOpenAI()
+llm(template.format_messages(text='i dont like eating tasty things.'))
 ```
 
 <CodeOutputBlock lang="python">
-
 ```
-    [SystemMessage(content='You are a helpful assistant that translates English to French.', additional_kwargs={}),
-     HumanMessage(content='I love programming.', additional_kwargs={})]
+AIMessage(content='I absolutely adore indulging in delicious treats!', additional_kwargs={}, example=False)
 ```
-
 </CodeOutputBlock>
+
+This provides you with a lot of flexibility in how you construct your chat prompts.
+

libs/experimental/tests/integration_tests/chains/test_cpal.py

@@ -7,6 +7,10 @@ from unittest import mock
 
 import pydantic
 import pytest
+from langchain import OpenAI
+from langchain.output_parsers import PydanticOutputParser
+from langchain.prompts.prompt import PromptTemplate
+
 from langchain_experimental.cpal.base import (
     CausalChain,
     CPALChain,
@@ -35,10 +39,6 @@ from langchain_experimental.cpal.templates.univariate.narrative import (
 from langchain_experimental.cpal.templates.univariate.query import (
     template as query_template,
 )
-
-from langchain import OpenAI
-from langchain.output_parsers import PydanticOutputParser
-from langchain.prompts.prompt import PromptTemplate
 from tests.unit_tests.llms.fake_llm import FakeLLM
 
 

libs/experimental/tests/integration_tests/chains/test_sql_database.py

@@ -1,12 +1,11 @@
 """Test SQL Database Chain."""
-from sqlalchemy import Column, Integer, MetaData, String, Table, create_engine, insert
-
-from langchain.chains.sql_database.base import (
+from langchain.llms.openai import OpenAI
+from langchain.utilities.sql_database import SQLDatabase
+from libs.experimental.langchain_experimental.sql.base import (
     SQLDatabaseChain,
     SQLDatabaseSequentialChain,
 )
-from langchain.llms.openai import OpenAI
-from langchain.utilities.sql_database import SQLDatabase
+from sqlalchemy import Column, Integer, MetaData, String, Table, create_engine, insert
 
 metadata_obj = MetaData()
 

libs/langchain/Makefile

@@ -59,6 +59,9 @@ test_watch:
 integration_tests:
 	poetry run pytest tests/integration_tests
 
+scheduled_tests:
+	poetry run pytest -m scheduled tests/integration_tests
+
 docker_tests:
 	docker build -t my-langchain-image:test .
 	docker run --rm my-langchain-image:test
@@ -92,15 +95,24 @@ spell_fix:
 ######################
 
 help:
-	@echo '----'
-	@echo 'coverage                     - run unit tests and generate coverage report'
+	@echo '===================='
+	@echo '-- DOCUMENTATION --'
+	@echo 'clean                        - run docs_clean and api_docs_clean'
 	@echo 'docs_build                   - build the documentation'
 	@echo 'docs_clean                   - clean the documentation build artifacts'
 	@echo 'docs_linkcheck               - run linkchecker on the documentation'
+	@echo 'api_docs_build               - build the API Reference documentation'
+	@echo 'api_docs_clean               - clean the API Reference documentation build artifacts'
+	@echo 'api_docs_linkcheck           - run linkchecker on the API Reference documentation'
+	@echo '-- LINTING --'
 	@echo 'format                       - run code formatters'
 	@echo 'lint                         - run linters'
+	@echo 'spell_check               	- run codespell on the project'
+	@echo 'spell_fix               		- run codespell on the project and fix the errors'
+	@echo '-- TESTS --'
+	@echo 'coverage                     - run unit tests and generate coverage report'
 	@echo 'test                         - run unit tests'
-	@echo 'tests                        - run unit tests'
+	@echo 'tests                        - run unit tests (alias for "make test")'
 	@echo 'test TEST_FILE=<test_file>   - run all tests in file'
 	@echo 'extended_tests               - run only extended unit tests'
 	@echo 'test_watch                   - run unit tests in watch mode'

libs/langchain/langchain/_api/__init__.py

@@ -0,0 +1,22 @@
+"""Helper functions for managing the LangChain API.
+
+This module is only relevant for LangChain developers, not for users.
+
+.. warning::
+
+    This module and its submodules are for internal use only.  Do not use them
+    in your own code.  We may change the API at any time with no warning.
+
+"""
+
+from .deprecation import (
+    LangChainDeprecationWarning,
+    deprecated,
+    suppress_langchain_deprecation_warning,
+)
+
+__all__ = [
+    "deprecated",
+    "LangChainDeprecationWarning",
+    "suppress_langchain_deprecation_warning",
+]

libs/langchain/langchain/_api/deprecation.py

@@ -0,0 +1,306 @@
+"""Helper functions for deprecating parts of the LangChain API.
+
+This module was adapted from matplotlibs _api/deprecation.py module:
+
+https://github.com/matplotlib/matplotlib/blob/main/lib/matplotlib/_api/deprecation.py
+
+.. warning::
+
+    This module is for internal use only.  Do not use it in your own code.
+    We may change the API at any time with no warning.
+"""
+
+import contextlib
+import functools
+import inspect
+import warnings
+from typing import Any, Callable, Generator, Type, TypeVar
+
+
+class LangChainDeprecationWarning(DeprecationWarning):
+    """A class for issuing deprecation warnings for LangChain users."""
+
+
+def _warn_deprecated(
+    since: str,
+    *,
+    message: str = "",
+    name: str = "",
+    alternative: str = "",
+    pending: bool = False,
+    obj_type: str = "",
+    addendum: str = "",
+    removal: str = "",
+) -> None:
+    """Display a standardized deprecation.
+
+    Arguments:
+        since : str
+            The release at which this API became deprecated.
+        message : str, optional
+            Override the default deprecation message. The %(since)s,
+            %(name)s, %(alternative)s, %(obj_type)s, %(addendum)s,
+            and %(removal)s format specifiers will be replaced by the
+            values of the respective arguments passed to this function.
+        name : str, optional
+            The name of the deprecated object.
+        alternative : str, optional
+            An alternative API that the user may use in place of the
+            deprecated API. The deprecation warning will tell the user
+            about this alternative if provided.
+        pending : bool, optional
+            If True, uses a PendingDeprecationWarning instead of a
+            DeprecationWarning. Cannot be used together with removal.
+        obj_type : str, optional
+            The object type being deprecated.
+        addendum : str, optional
+            Additional text appended directly to the final message.
+        removal : str, optional
+            The expected removal version. With the default (an empty
+            string), a removal version is automatically computed from
+            since. Set to other Falsy values to not schedule a removal
+            date. Cannot be used together with pending.
+    """
+    if pending and removal:
+        raise ValueError("A pending deprecation cannot have a scheduled removal")
+
+    if not pending:
+        if not removal:
+            removal = f"in {removal}" if removal else "within ?? minor releases"
+            raise NotImplementedError(
+                f"Need to determine which default deprecation schedule to use. "
+                f"{removal}"
+            )
+        else:
+            removal = f"in {removal}"
+
+    if not message:
+        message = ""
+
+        if obj_type:
+            message += f"The {obj_type} `{name}`"
+        else:
+            message += f"`{name}`"
+
+        if pending:
+            message += " will be deprecated in a future version"
+        else:
+            message += f" was deprecated in LangChain {since}"
+
+            if removal:
+                message += f" and will be removed {removal}"
+
+        if alternative:
+            message += f". Use {alternative} instead."
+
+        if addendum:
+            message += f" {addendum}"
+
+    warning_cls = PendingDeprecationWarning if pending else LangChainDeprecationWarning
+    warning = warning_cls(message)
+    warnings.warn(warning, category=LangChainDeprecationWarning, stacklevel=2)
+
+
+# PUBLIC API
+
+
+T = TypeVar("T", Type, Callable)
+
+
+def deprecated(
+    since: str,
+    *,
+    message: str = "",
+    name: str = "",
+    alternative: str = "",
+    pending: bool = False,
+    obj_type: str = "",
+    addendum: str = "",
+    removal: str = "",
+) -> Callable[[T], T]:
+    """Decorator to mark a function, a class, or a property as deprecated.
+
+    When deprecating a classmethod, a staticmethod, or a property, the
+    ``@deprecated`` decorator should go *under* ``@classmethod`` and
+    ``@staticmethod`` (i.e., `deprecated` should directly decorate the
+    underlying callable), but *over* ``@property``.
+
+    When deprecating a class ``C`` intended to be used as a base class in a
+    multiple inheritance hierarchy, ``C`` *must* define an ``__init__`` method
+    (if ``C`` instead inherited its ``__init__`` from its own base class, then
+    ``@deprecated`` would mess up ``__init__`` inheritance when installing its
+    own (deprecation-emitting) ``C.__init__``).
+
+    Parameters are the same as for `warn_deprecated`, except that *obj_type*
+    defaults to 'class' if decorating a class, 'attribute' if decorating a
+    property, and 'function' otherwise.
+
+    Arguments:
+        since : str
+            The release at which this API became deprecated.
+        message : str, optional
+            Override the default deprecation message. The %(since)s,
+            %(name)s, %(alternative)s, %(obj_type)s, %(addendum)s,
+            and %(removal)s format specifiers will be replaced by the
+            values of the respective arguments passed to this function.
+        name : str, optional
+            The name of the deprecated object.
+        alternative : str, optional
+            An alternative API that the user may use in place of the
+            deprecated API. The deprecation warning will tell the user
+            about this alternative if provided.
+        pending : bool, optional
+            If True, uses a PendingDeprecationWarning instead of a
+            DeprecationWarning. Cannot be used together with removal.
+        obj_type : str, optional
+            The object type being deprecated.
+        addendum : str, optional
+            Additional text appended directly to the final message.
+        removal : str, optional
+            The expected removal version. With the default (an empty
+            string), a removal version is automatically computed from
+            since. Set to other Falsy values to not schedule a removal
+            date. Cannot be used together with pending.
+
+    Examples
+    --------
+
+        .. code-block:: python
+
+            @deprecated('1.4.0')
+            def the_function_to_deprecate():
+                pass
+    """
+
+    def deprecate(
+        obj: T,
+        *,
+        _obj_type: str = obj_type,
+        _name: str = name,
+        _message: str = message,
+        _alternative: str = alternative,
+        _pending: bool = pending,
+        _addendum: str = addendum,
+    ) -> T:
+        """Implementation of the decorator returned by `deprecated`."""
+        if isinstance(obj, type):
+            if not _obj_type:
+                _obj_type = "class"
+            wrapped = obj.__init__  # type: ignore
+            _name = _name or obj.__name__
+            old_doc = obj.__doc__
+
+            def finalize(wrapper: Callable[..., Any], new_doc: str) -> T:
+                """Finalize the deprecation of a class."""
+                try:
+                    obj.__doc__ = new_doc
+                except AttributeError:  # Can't set on some extension objects.
+                    pass
+                obj.__init__ = functools.wraps(obj.__init__)(  # type: ignore[misc]
+                    wrapper
+                )
+                return obj
+
+        elif isinstance(obj, property):
+            if not _obj_type:
+                _obj_type = "attribute"
+            wrapped = None
+            _name = _name or obj.fget.__name__
+            old_doc = obj.__doc__
+
+            class _deprecated_property(type(obj)):  # type: ignore
+                """A deprecated property."""
+
+                def __get__(self, instance, owner=None):  # type: ignore
+                    if instance is not None or owner is not None:
+                        emit_warning()
+                    return super().__get__(instance, owner)
+
+                def __set__(self, instance, value):  # type: ignore
+                    if instance is not None:
+                        emit_warning()
+                    return super().__set__(instance, value)
+
+                def __delete__(self, instance):  # type: ignore
+                    if instance is not None:
+                        emit_warning()
+                    return super().__delete__(instance)
+
+                def __set_name__(self, owner, set_name):  # type: ignore
+                    nonlocal _name
+                    if _name == "<lambda>":
+                        _name = set_name
+
+            def finalize(_: Any, new_doc: str) -> Any:  # type: ignore
+                """Finalize the property."""
+                return _deprecated_property(
+                    fget=obj.fget, fset=obj.fset, fdel=obj.fdel, doc=new_doc
+                )
+
+        else:
+            if not _obj_type:
+                _obj_type = "function"
+            wrapped = obj
+            _name = _name or obj.__name__  # type: ignore
+            old_doc = wrapped.__doc__
+
+            def finalize(  # type: ignore
+                wrapper: Callable[..., Any], new_doc: str
+            ) -> T:
+                """Wrap the wrapped function using the wrapper and update the docstring.
+
+                Args:
+                    wrapper: The wrapper function.
+                    new_doc: The new docstring.
+
+                Returns:
+                    The wrapped function.
+                """
+                wrapper = functools.wraps(wrapped)(wrapper)
+                wrapper.__doc__ = new_doc
+                return wrapper
+
+        def emit_warning() -> None:
+            """Emit the warning."""
+            _warn_deprecated(
+                since,
+                message=_message,
+                name=_name,
+                alternative=_alternative,
+                pending=_pending,
+                obj_type=_obj_type,
+                addendum=_addendum,
+                removal=removal,
+            )
+
+        def warning_emitting_wrapper(*args: Any, **kwargs: Any) -> Any:
+            """Wrapper for the original wrapped callable that emits a warning.
+
+            Args:
+                *args: The positional arguments to the function.
+                **kwargs: The keyword arguments to the function.
+
+            Returns:
+                The return value of the function being wrapped.
+            """
+            emit_warning()
+            return wrapped(*args, **kwargs)
+
+        old_doc = inspect.cleandoc(old_doc or "").strip("\n")
+
+        if not old_doc:
+            new_doc = "[*Deprecated*]"
+        else:
+            new_doc = f"[*Deprecated*]  {old_doc}"
+
+        return finalize(warning_emitting_wrapper, new_doc)
+
+    return deprecate
+
+
+@contextlib.contextmanager
+def suppress_langchain_deprecation_warning() -> Generator[None, None, None]:
+    """Context manager to suppress LangChainDeprecationWarning."""
+    with warnings.catch_warnings():
+        warnings.simplefilter("ignore", LangChainDeprecationWarning)
+        yield