x

Add `batch` utility to langchain

.devcontainer/README.md

@@ -15,7 +15,11 @@ You may use the button above, or follow these steps to open this repo in a Codes
 For more info, check out the [GitHub documentation](https://docs.github.com/en/free-pro-team@latest/github/developing-online-with-codespaces/creating-a-codespace#creating-a-codespace).
   
 ## VS Code Dev Containers
-[![Open in Dev Containers](https://img.shields.io/static/v1?label=Dev%20Containers&message=Open&color=blue&logo=visualstudiocode)](https://vscode.dev/redirect?url=vscode://ms-vscode-remote.remote-containers/cloneInVolume?url=https://github.com/hwchase17/langchain)
+[![Open in Dev Containers](https://img.shields.io/static/v1?label=Dev%20Containers&message=Open&color=blue&logo=visualstudiocode)](https://vscode.dev/redirect?url=vscode://ms-vscode-remote.remote-containers/cloneInVolume?url=https://github.com/langchain-ai/langchain)
+
+Note: If you click this link you will open the main repo and not your local cloned repo, you can use this link and replace with your username and cloned repo name: 
+https://vscode.dev/redirect?url=vscode://ms-vscode-remote.remote-containers/cloneInVolume?url=https://github.com/<yourusername>/<yourclonedreponame>
+
 
 If you already have VS Code and Docker installed, you can use the button above to get started. This will cause VS Code to automatically install the Dev Containers extension if needed, clone the source code into a container volume, and spin up a dev container for use.
 
@@ -25,7 +29,7 @@ You can also follow these steps to open this repo in a container using the VS Co
 
 2. Open a locally cloned copy of the code:
 
-   - Clone this repository to your local filesystem.
+   - Fork and Clone this repository to your local filesystem.
    - Press <kbd>F1</kbd> and select the **Dev Containers: Open Folder in Container...** command.
    - Select the cloned copy of this folder, wait for the container to start, and try things out!
 

.github/CONTRIBUTING.md

@@ -33,7 +33,7 @@ best way to get our attention.
 ### 🚩GitHub Issues
 
 Our [issues](https://github.com/hwchase17/langchain/issues) page is kept up to date
-with bugs, improvements, and feature requests. 
+with bugs, improvements, and feature requests.
 
 There is a taxonomy of labels to help with sorting and discovery of issues of interest. Please use these to help
 organize issues.
@@ -61,11 +61,11 @@ we do not want these to get in the way of getting good code into the codebase.
 
 > **Note:** You can run this repository locally (which is described below) or in a [development container](https://containers.dev/) (which is described in the [.devcontainer folder](https://github.com/hwchase17/langchain/tree/master/.devcontainer)).
 
-This project uses [Poetry](https://python-poetry.org/) as a dependency manager. Check out Poetry's [documentation on how to install it](https://python-poetry.org/docs/#installation) on your system before proceeding.
+This project uses [Poetry](https://python-poetry.org/) v1.5.1 as a dependency manager. Check out Poetry's [documentation on how to install it](https://python-poetry.org/docs/#installation) on your system before proceeding.
 
 ❗Note: If you use `Conda` or `Pyenv` as your environment / package manager, avoid dependency conflicts by doing the following first:
 1. *Before installing Poetry*, create and activate a new Conda env (e.g. `conda create -n langchain python=3.9`)
-2. Install Poetry (see above)
+2. Install Poetry v1.5.1 (see above)
 3. Tell Poetry to use the virtualenv python environment (`poetry config virtualenvs.prefer-active-python true`)
 4. Continue with the following steps.
 
@@ -73,7 +73,7 @@ There are two separate projects in this repository:
 - `langchain`: core langchain code, abstractions, and use cases
 - `langchain.experimental`: more experimental code
 
-Each of these has their OWN development environment. 
+Each of these has their OWN development environment.
 In order to run any of the commands below, please move into their respective directories.
 For example, to contribute to `langchain` run `cd libs/langchain` before getting started with the below.
 
@@ -85,7 +85,7 @@ poetry install -E all
 
 This will install all requirements for running the package, examples, linting, formatting, tests, and coverage. Note the `-E all` flag will install all optional dependencies necessary for integration testing.
 
-❗Note: If you're running Poetry 1.4.1 and receive a `WheelFileValidationError` for `debugpy` during installation, you can try either downgrading to Poetry 1.4.0 or disabling "modern installation" (`poetry config installer.modern-installation false`) and re-install requirements. See [this `debugpy` issue](https://github.com/microsoft/debugpy/issues/1246) for more details.
+❗Note: If during installation you receive a `WheelFileValidationError` for `debugpy`, please make sure you are running Poetry v1.5.1. This bug was present in older versions of Poetry (e.g. 1.4.1) and has been resolved in newer releases. If you are still seeing this bug on v1.5.1, you may also try disabling "modern installation" (`poetry config installer.modern-installation false`) and re-installing requirements. See [this `debugpy` issue](https://github.com/microsoft/debugpy/issues/1246) for more details.
 
 Now, you should be able to run the common tasks in the following section. To double check, run `make test`, all tests should pass. If they don't you may need to pip install additional dependencies, such as `numexpr` and `openapi_schema_pydantic`.
 
@@ -175,9 +175,9 @@ If you're adding a new dependency to Langchain, assume that it will be an option
 that most users won't have it installed.
 
 Users that do not have the dependency installed should be able to **import** your code without
-any side effects (no warnings, no errors, no exceptions). 
+any side effects (no warnings, no errors, no exceptions).
 
-To introduce the dependency to the pyproject.toml file correctly, please do the following: 
+To introduce the dependency to the pyproject.toml file correctly, please do the following:
 
 1. Add the dependency to the main group as an optional dependency
   ```bash
@@ -220,7 +220,7 @@ If you add new logic, please add a unit test.
 
 Integration tests cover logic that requires making calls to outside APIs (often integration with other services).
 
-**warning** Almost no tests should be integration tests. 
+**warning** Almost no tests should be integration tests.
 
   Tests that require making network connections make it difficult for other
   developers to test the code.
@@ -307,4 +307,3 @@ even patch releases may contain [non-backwards-compatible changes](https://semve
 
 If your contribution has made its way into a release, we will want to give you credit on Twitter (only if you want though)!
 If you have a Twitter account you would like us to mention, please let us know in the PR or in another manner.
-

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -1,5 +1,5 @@
 name: "\U0001F41B Bug Report"
-description: Submit a bug report to help us improve LangChain
+description: Submit a bug report to help us improve LangChain. To report a security issue, please instead use the security option below.
 labels: ["02 Bug Report"]
 body:
   - type: markdown

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,28 +1,20 @@
 <!-- Thank you for contributing to LangChain!
 
-Replace this comment with:
+Replace this entire comment with:
   - Description: a description of the change, 
   - Issue: the issue # it fixes (if applicable),
   - Dependencies: any dependencies required for this change,
   - Tag maintainer: for a quicker response, tag the relevant maintainer (see below),
   - Twitter handle: we announce bigger features on Twitter. If your PR gets announced and you'd like a mention, we'll gladly shout you out!
 
-Please make sure you're PR is passing linting and testing before submitting. Run `make format`, `make lint` and `make test` to check this locally.
+Please make sure your PR is passing linting and testing before submitting. Run `make format`, `make lint` and `make test` to check this locally.
+
+See contribution guidelines for more information on how to write/run tests, lint, etc: 
+https://github.com/hwchase17/langchain/blob/master/.github/CONTRIBUTING.md
 
 If you're adding a new integration, please include:
   1. a test for the integration, preferably unit tests that do not rely on network access,
-  2. an example notebook showing its use.
+  2. an example notebook showing its use. These live is docs/extras directory.
 
-Maintainer responsibilities:
-  - General / Misc / if you don't know who to tag: @baskaryan
-  - DataLoaders / VectorStores / Retrievers: @rlancemartin, @eyurtsev
-  - Models / Prompts: @hwchase17, @baskaryan
-  - Memory: @hwchase17
-  - Agents / Tools / Toolkits: @hinthornw
-  - Tracing / Callbacks: @agola11
-  - Async: @agola11
-
-If no one reviews your PR within a few days, feel free to @-mention the same people again.
-
-See contribution guidelines for more information on how to write/run tests, lint, etc: https://github.com/hwchase17/langchain/blob/master/.github/CONTRIBUTING.md
+If no one reviews your PR within a few days, please @-mention one of @baskaryan, @eyurtsev, @hwchase17, @rlancemartin.
  -->

.github/actions/poetry_setup/action.yml

@@ -41,14 +41,18 @@ runs:
       id: cache-pip
       name: Cache Pip ${{ inputs.python-version }}
       env:
-        SEGMENT_DOWNLOAD_TIMEOUT_MIN: "15"
+        SEGMENT_DOWNLOAD_TIMEOUT_MIN: "4"
       with:
         path: |
           ~/.cache/pip
         key: pip-${{ runner.os }}-${{ runner.arch }}-py-${{ inputs.python-version }}
 
-    - run: pipx install poetry==${{ inputs.poetry-version }} --python python${{ inputs.python-version }}
+    - name: Install poetry
       shell: bash
+      env:
+        POETRY_VERSION: ${{ inputs.poetry-version }}
+        PYTHON_VERSION: ${{ inputs.python-version }}
+      run: pipx install "poetry==$POETRY_VERSION" --python "python$PYTHON_VERSION" --verbose
 
     - name: Check Poetry File
       shell: bash
@@ -65,13 +69,15 @@ runs:
     - uses: actions/cache@v3
       id: cache-poetry
       env:
-        SEGMENT_DOWNLOAD_TIMEOUT_MIN: "15"
+        SEGMENT_DOWNLOAD_TIMEOUT_MIN: "4"
+        WORKDIR: ${{ inputs.working-directory == '' && '.' || inputs.working-directory }}
       with:
         path: |
           ~/.cache/pypoetry/virtualenvs
           ~/.cache/pypoetry/cache
           ~/.cache/pypoetry/artifacts
-        key: poetry-${{ runner.os }}-${{ runner.arch }}-py-${{ inputs.python-version }}-poetry-${{ inputs.poetry-version }}-${{ inputs.cache-key }}-${{ hashFiles('poetry.lock') }}
+          ${{ env.WORKDIR }}/.venv
+        key: poetry-${{ runner.os }}-${{ runner.arch }}-py-${{ inputs.python-version }}-poetry-${{ inputs.poetry-version }}-${{ inputs.cache-key }}-${{ hashFiles(format('{0}/poetry.lock', env.WORKDIR)) }}
 
     - run: ${{ inputs.install-command }}
       working-directory: ${{ inputs.working-directory }}

.github/tools/git-restore-mtime

@@ -0,0 +1,606 @@
+#!/usr/bin/env python3
+#
+# git-restore-mtime - Change mtime of files based on commit date of last change
+#
+#    Copyright (C) 2012 Rodrigo Silva (MestreLion) <linux@rodrigosilva.com>
+#
+#    This program is free software: you can redistribute it and/or modify
+#    it under the terms of the GNU General Public License as published by
+#    the Free Software Foundation, either version 3 of the License, or
+#    (at your option) any later version.
+#
+#    This program is distributed in the hope that it will be useful,
+#    but WITHOUT ANY WARRANTY; without even the implied warranty of
+#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+#    GNU General Public License for more details.
+#
+#    You should have received a copy of the GNU General Public License
+#    along with this program. See <http://www.gnu.org/licenses/gpl.html>
+#
+# Source: https://github.com/MestreLion/git-tools
+# Version: July 13, 2023 (commit hash 5f832e72453e035fccae9d63a5056918d64476a2)
+"""
+Change the modification time (mtime) of files in work tree, based on the
+date of the most recent commit that modified the file, including renames.
+
+Ignores untracked files and uncommitted deletions, additions and renames, and
+by default modifications too.
+---
+Useful prior to generating release tarballs, so each file is archived with a
+date that is similar to the date when the file was actually last modified,
+assuming the actual modification date and its commit date are close.
+"""
+
+# TODO:
+# - Add -z on git whatchanged/ls-files, so we don't deal with filename decoding
+# - When Python is bumped to 3.7, use text instead of universal_newlines on subprocess
+# - Update "Statistics for some large projects" with modern hardware and repositories.
+# - Create a README.md for git-restore-mtime alone. It deserves extensive documentation
+#   - Move Statistics there
+# - See git-extras as a good example on project structure and documentation
+
+# FIXME:
+# - When current dir is outside the worktree, e.g. using --work-tree, `git ls-files`
+#   assume any relative pathspecs are to worktree root, not the current dir. As such,
+#   relative pathspecs may not work.
+# - Renames are tricky:
+#   - R100 should not change mtime, but original name is not on filelist. Should
+#     track renames until a valid (A, M) mtime found and then set on current name.
+#   - Should set mtime for both current and original directories.
+#   - Check mode changes with unchanged blobs?
+# - Check file (A, D) for the directory mtime is not sufficient:
+#   - Renames also change dir mtime, unless rename was on a parent dir
+#   - If most recent change of all files in a dir was a Modification (M),
+#     dir might not be touched at all.
+#   - Dirs containing only subdirectories but no direct files will also
+#     not be touched. They're files' [grand]parent dir, but never their dirname().
+#   - Some solutions:
+#     - After files done, perform some dir processing for missing dirs, finding latest
+#       file (A, D, R)
+#     - Simple approach: dir mtime is the most recent child (dir or file) mtime
+#     - Use a virtual concept of "created at most at" to fill missing info, bubble up
+#       to parents and grandparents
+#   - When handling [grand]parent dirs, stay inside <pathspec>
+# - Better handling of merge commits. `-m` is plain *wrong*. `-c/--cc` is perfect, but
+#   painfully slow. First pass without merge commits is not accurate. Maybe add a new
+#   `--accurate` mode for `--cc`?
+
+if __name__ != "__main__":
+    raise ImportError("{} should not be used as a module.".format(__name__))
+
+import argparse
+import datetime
+import logging
+import os.path
+import shlex
+import signal
+import subprocess
+import sys
+import time
+
+__version__ = "2022.12+dev"
+
+# Update symlinks only if the platform supports not following them
+UPDATE_SYMLINKS = bool(os.utime in getattr(os, 'supports_follow_symlinks', []))
+
+# Call os.path.normpath() only if not in a POSIX platform (Windows)
+NORMALIZE_PATHS = (os.path.sep != '/')
+
+# How many files to process in each batch when re-trying merge commits
+STEPMISSING = 100
+
+# (Extra) keywords for the os.utime() call performed by touch()
+UTIME_KWS = {} if not UPDATE_SYMLINKS else {'follow_symlinks': False}
+
+
+# Command-line interface ######################################################
+
+def parse_args():
+    parser = argparse.ArgumentParser(
+        description=__doc__.split('\n---')[0])
+
+    group = parser.add_mutually_exclusive_group()
+    group.add_argument('--quiet', '-q', dest='loglevel',
+        action="store_const", const=logging.WARNING, default=logging.INFO,
+        help="Suppress informative messages and summary statistics.")
+    group.add_argument('--verbose', '-v', action="count", help="""
+        Print additional information for each processed file.
+        Specify twice to further increase verbosity.
+        """)
+
+    parser.add_argument('--cwd', '-C', metavar="DIRECTORY", help="""
+        Run as if %(prog)s was started in directory %(metavar)s.
+        This affects how --work-tree, --git-dir and PATHSPEC arguments are handled.
+        See 'man 1 git' or 'git --help' for more information.
+        """)
+
+    parser.add_argument('--git-dir', dest='gitdir', metavar="GITDIR", help="""
+        Path to the git repository, by default auto-discovered by searching
+        the current directory and its parents for a .git/ subdirectory.
+        """)
+
+    parser.add_argument('--work-tree', dest='workdir', metavar="WORKTREE", help="""
+        Path to the work tree root, by default the parent of GITDIR if it's
+        automatically discovered, or the current directory if GITDIR is set.
+        """)
+
+    parser.add_argument('--force', '-f', default=False, action="store_true", help="""
+        Force updating files with uncommitted modifications.
+        Untracked files and uncommitted deletions, renames and additions are
+        always ignored.
+        """)
+
+    parser.add_argument('--merge', '-m', default=False, action="store_true", help="""
+        Include merge commits.
+        Leads to more recent times and more files per commit, thus with the same
+        time, which may or may not be what you want.
+        Including merge commits may lead to fewer commits being evaluated as files
+        are found sooner, which can improve performance, sometimes substantially.
+        But as merge commits are usually huge, processing them may also take longer.
+        By default, merge commits are only used for files missing from regular commits.
+        """)
+
+    parser.add_argument('--first-parent', default=False, action="store_true", help="""
+        Consider only the first parent, the "main branch", when evaluating merge commits.
+        Only effective when merge commits are processed, either when --merge is
+        used or when finding missing files after the first regular log search.
+        See --skip-missing.
+        """)
+
+    parser.add_argument('--skip-missing', '-s', dest="missing", default=True,
+        action="store_false", help="""
+        Do not try to find missing files.
+        If merge commits were not evaluated with --merge and some files were
+        not found in regular commits, by default %(prog)s searches for these
+        files again in the merge commits.
+        This option disables this retry, so files found only in merge commits
+        will not have their timestamp updated.
+        """)
+
+    parser.add_argument('--no-directories', '-D', dest='dirs', default=True,
+        action="store_false", help="""
+        Do not update directory timestamps.
+        By default, use the time of its most recently created, renamed or deleted file.
+        Note that just modifying a file will NOT update its directory time.
+        """)
+
+    parser.add_argument('--test', '-t', default=False, action="store_true",
+        help="Test run: do not actually update any file timestamp.")
+
+    parser.add_argument('--commit-time', '-c', dest='commit_time', default=False,
+        action='store_true', help="Use commit time instead of author time.")
+
+    parser.add_argument('--oldest-time', '-o', dest='reverse_order', default=False,
+        action='store_true', help="""
+        Update times based on the oldest, instead of the most recent commit of a file.
+        This reverses the order in which the git log is processed to emulate a
+        file "creation" date. Note this will be inaccurate for files deleted and
+        re-created at later dates.
+        """)
+
+    parser.add_argument('--skip-older-than', metavar='SECONDS', type=int, help="""
+        Ignore files that are currently older than %(metavar)s.
+        Useful in workflows that assume such files already have a correct timestamp,
+        as it may improve performance by processing fewer files.
+        """)
+
+    parser.add_argument('--skip-older-than-commit', '-N', default=False,
+        action='store_true', help="""
+        Ignore files older than the timestamp it would be updated to.
+        Such files may be considered "original", likely in the author's repository.
+        """)
+
+    parser.add_argument('--unique-times', default=False, action="store_true", help="""
+        Set the microseconds to a unique value per commit.
+        Allows telling apart changes that would otherwise have identical timestamps,
+        as git's time accuracy is in seconds.
+        """)
+
+    parser.add_argument('pathspec', nargs='*', metavar='PATHSPEC', help="""
+        Only modify paths matching %(metavar)s, relative to current directory.
+        By default, update all but untracked files and submodules.
+        """)
+
+    parser.add_argument('--version', '-V', action='version',
+        version='%(prog)s version {version}'.format(version=get_version()))
+
+    args_ = parser.parse_args()
+    if args_.verbose:
+        args_.loglevel = max(logging.TRACE, logging.DEBUG // args_.verbose)
+    args_.debug = args_.loglevel <= logging.DEBUG
+    return args_
+
+
+def get_version(version=__version__):
+    if not version.endswith('+dev'):
+        return version
+    try:
+        cwd = os.path.dirname(os.path.realpath(__file__))
+        return Git(cwd=cwd, errors=False).describe().lstrip('v')
+    except Git.Error:
+        return '-'.join((version, "unknown"))
+
+
+# Helper functions ############################################################
+
+def setup_logging():
+    """Add TRACE logging level and corresponding method, return the root logger"""
+    logging.TRACE = TRACE = logging.DEBUG // 2
+    logging.Logger.trace = lambda _, m, *a, **k: _.log(TRACE, m, *a, **k)
+    return logging.getLogger()
+
+
+def normalize(path):
+    r"""Normalize paths from git, handling non-ASCII characters.
+
+    Git stores paths as UTF-8 normalization form C.
+    If path contains non-ASCII or non-printable characters, git outputs the UTF-8
+    in octal-escaped notation, escaping double-quotes and backslashes, and then
+    double-quoting the whole path.
+    https://git-scm.com/docs/git-config#Documentation/git-config.txt-corequotePath
+
+    This function reverts this encoding, so:
+    normalize(r'"Back\\slash_double\"quote_a\303\247a\303\255"') =>
+        r'Back\slash_double"quote_açaí')
+
+    Paths with invalid UTF-8 encoding, such as single 0x80-0xFF bytes (e.g, from
+    Latin1/Windows-1251 encoding) are decoded using surrogate escape, the same
+    method used by Python for filesystem paths. So 0xE6 ("æ" in Latin1, r'\\346'
+    from Git) is decoded as "\udce6". See https://peps.python.org/pep-0383/ and
+    https://vstinner.github.io/painful-history-python-filesystem-encoding.html
+
+    Also see notes on `windows/non-ascii-paths.txt` about path encodings on
+    non-UTF-8 platforms and filesystems.
+    """
+    if path and path[0] == '"':
+        # Python 2: path = path[1:-1].decode("string-escape")
+        # Python 3: https://stackoverflow.com/a/46650050/624066
+        path = (path[1:-1]                 # Remove enclosing double quotes
+                .encode('latin1')          # Convert to bytes, required by 'unicode-escape'
+                .decode('unicode-escape')  # Perform the actual octal-escaping decode
+                .encode('latin1')          # 1:1 mapping to bytes, UTF-8 encoded
+                .decode('utf8', 'surrogateescape'))  # Decode from UTF-8
+    if NORMALIZE_PATHS:
+        # Make sure the slash matches the OS; for Windows we need a backslash
+        path = os.path.normpath(path)
+    return path
+
+
+def dummy(*_args, **_kwargs):
+    """No-op function used in dry-run tests"""
+
+
+def touch(path, mtime):
+    """The actual mtime update"""
+    os.utime(path, (mtime, mtime), **UTIME_KWS)
+
+
+def touch_ns(path, mtime_ns):
+    """The actual mtime update, using nanoseconds for unique timestamps"""
+    os.utime(path, None, ns=(mtime_ns, mtime_ns), **UTIME_KWS)
+
+
+def isodate(secs: int):
+    # time.localtime() accepts floats, but discards fractional part
+    return time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(secs))
+
+
+def isodate_ns(ns: int):
+    # for integers fromtimestamp() is equivalent and ~16% slower than isodate()
+    return datetime.datetime.fromtimestamp(ns / 1000000000).isoformat(sep=' ')
+
+
+def get_mtime_ns(secs: int, idx: int):
+    # Time resolution for filesystems and functions:
+    # ext-4 and other POSIX filesystems: 1 nanosecond
+    # NTFS (Windows default): 100 nanoseconds
+    # datetime.datetime() (due to 64-bit float epoch): 1 microsecond
+    us = idx % 1000000  # 10**6
+    return 1000 * (1000000 * secs + us)
+
+
+def get_mtime_path(path):
+    return os.path.getmtime(path)
+
+
+# Git class and parse_log(), the heart of the script ##########################
+
+class Git:
+    def __init__(self, workdir=None, gitdir=None, cwd=None, errors=True):
+        self.gitcmd = ['git']
+        self.errors = errors
+        self._proc = None
+        if workdir: self.gitcmd.extend(('--work-tree', workdir))
+        if gitdir:  self.gitcmd.extend(('--git-dir',   gitdir))
+        if cwd:     self.gitcmd.extend(('-C',          cwd))
+        self.workdir, self.gitdir = self._get_repo_dirs()
+
+    def ls_files(self, paths: list = None):
+        return (normalize(_) for _ in self._run('ls-files --full-name', paths))
+
+    def ls_dirty(self, force=False):
+        return (normalize(_[3:].split(' -> ', 1)[-1])
+                for _ in self._run('status --porcelain')
+                if _[:2] != '??' and (not force or (_[0] in ('R', 'A')
+                                                    or _[1] == 'D')))
+
+    def log(self, merge=False, first_parent=False, commit_time=False,
+            reverse_order=False, paths: list = None):
+        cmd = 'whatchanged --pretty={}'.format('%ct' if commit_time else '%at')
+        if merge:         cmd += ' -m'
+        if first_parent:  cmd += ' --first-parent'
+        if reverse_order: cmd += ' --reverse'
+        return self._run(cmd, paths)
+
+    def describe(self):
+        return self._run('describe --tags', check=True)[0]
+
+    def terminate(self):
+        if self._proc is None:
+            return
+        try:
+            self._proc.terminate()
+        except OSError:
+            # Avoid errors on OpenBSD
+            pass
+
+    def _get_repo_dirs(self):
+        return (os.path.normpath(_) for _ in
+            self._run('rev-parse --show-toplevel --absolute-git-dir', check=True))
+
+    def _run(self, cmdstr: str, paths: list = None, output=True, check=False):
+        cmdlist = self.gitcmd + shlex.split(cmdstr)
+        if paths:
+            cmdlist.append('--')
+            cmdlist.extend(paths)
+        popen_args = dict(universal_newlines=True, encoding='utf8')
+        if not self.errors:
+            popen_args['stderr'] = subprocess.DEVNULL
+        log.trace("Executing: %s", ' '.join(cmdlist))
+        if not output:
+            return subprocess.call(cmdlist, **popen_args)
+        if check:
+            try:
+                stdout: str = subprocess.check_output(cmdlist, **popen_args)
+                return stdout.splitlines()
+            except subprocess.CalledProcessError as e:
+                raise self.Error(e.returncode, e.cmd, e.output, e.stderr)
+        self._proc = subprocess.Popen(cmdlist, stdout=subprocess.PIPE, **popen_args)
+        return (_.rstrip() for _ in self._proc.stdout)
+
+    def __del__(self):
+        self.terminate()
+
+    class Error(subprocess.CalledProcessError):
+        """Error from git executable"""
+
+
+def parse_log(filelist, dirlist, stats, git, merge=False, filterlist=None):
+    mtime = 0
+    datestr = isodate(0)
+    for line in git.log(
+            merge,
+            args.first_parent,
+            args.commit_time,
+            args.reverse_order,
+            filterlist
+    ):
+        stats['loglines'] += 1
+
+        # Blank line between Date and list of files
+        if not line:
+            continue
+
+        # Date line
+        if line[0] != ':':  # Faster than `not line.startswith(':')`
+            stats['commits'] += 1
+            mtime = int(line)
+            if args.unique_times:
+                mtime = get_mtime_ns(mtime, stats['commits'])
+            if args.debug:
+                datestr = isodate(mtime)
+            continue
+
+        # File line: three tokens if it describes a renaming, otherwise two
+        tokens = line.split('\t')
+
+        # Possible statuses:
+        # M: Modified (content changed)
+        # A: Added (created)
+        # D: Deleted
+        # T: Type changed: to/from regular file, symlinks, submodules
+        # R099: Renamed (moved), with % of unchanged content. 100 = pure rename
+        # Not possible in log: C=Copied, U=Unmerged, X=Unknown, B=pairing Broken
+        status = tokens[0].split(' ')[-1]
+        file = tokens[-1]
+
+        # Handles non-ASCII chars and OS path separator
+        file = normalize(file)
+
+        def do_file():
+            if args.skip_older_than_commit and get_mtime_path(file) <= mtime:
+                stats['skip'] += 1
+                return
+            if args.debug:
+                log.debug("%d\t%d\t%d\t%s\t%s",
+                          stats['loglines'], stats['commits'], stats['files'],
+                          datestr, file)
+            try:
+                touch(os.path.join(git.workdir, file), mtime)
+                stats['touches'] += 1
+            except Exception as e:
+                log.error("ERROR: %s: %s", e, file)
+                stats['errors'] += 1
+
+        def do_dir():
+            if args.debug:
+                log.debug("%d\t%d\t-\t%s\t%s",
+                          stats['loglines'], stats['commits'],
+                          datestr, "{}/".format(dirname or '.'))
+            try:
+                touch(os.path.join(git.workdir, dirname), mtime)
+                stats['dirtouches'] += 1
+            except Exception as e:
+                log.error("ERROR: %s: %s", e, dirname)
+                stats['direrrors'] += 1
+
+        if file in filelist:
+            stats['files'] -= 1
+            filelist.remove(file)
+            do_file()
+
+        if args.dirs and status in ('A', 'D'):
+            dirname = os.path.dirname(file)
+            if dirname in dirlist:
+                dirlist.remove(dirname)
+                do_dir()
+
+        # All files done?
+        if not stats['files']:
+            git.terminate()
+            return
+
+
+# Main Logic ##################################################################
+
+def main():
+    start = time.time()  # yes, Wall time. CPU time is not realistic for users.
+    stats = {_: 0 for _ in ('loglines', 'commits', 'touches', 'skip', 'errors',
+                            'dirtouches', 'direrrors')}
+
+    logging.basicConfig(level=args.loglevel, format='%(message)s')
+    log.trace("Arguments: %s", args)
+
+    # First things first: Where and Who are we?
+    if args.cwd:
+        log.debug("Changing directory: %s", args.cwd)
+        try:
+            os.chdir(args.cwd)
+        except OSError as e:
+            log.critical(e)
+            return e.errno
+    # Using both os.chdir() and `git -C` is redundant, but might prevent side effects
+    # `git -C` alone could be enough if we make sure that:
+    # - all paths, including args.pathspec, are processed by git: ls-files, rev-parse
+    # - touch() / os.utime() path argument is always prepended with git.workdir
+    try:
+        git = Git(workdir=args.workdir, gitdir=args.gitdir, cwd=args.cwd)
+    except Git.Error as e:
+        # Not in a git repository, and git already informed user on stderr. So we just...
+        return e.returncode
+
+    # Get the files managed by git and build file list to be processed
+    if UPDATE_SYMLINKS and not args.skip_older_than:
+        filelist = set(git.ls_files(args.pathspec))
+    else:
+        filelist = set()
+        for path in git.ls_files(args.pathspec):
+            fullpath = os.path.join(git.workdir, path)
+
+            # Symlink (to file, to dir or broken - git handles the same way)
+            if not UPDATE_SYMLINKS and os.path.islink(fullpath):
+                log.warning("WARNING: Skipping symlink, no OS support for updates: %s",
+                            path)
+                continue
+
+            # skip files which are older than given threshold
+            if (args.skip_older_than
+                    and start - get_mtime_path(fullpath) > args.skip_older_than):
+                continue
+
+            # Always add files relative to worktree root
+            filelist.add(path)
+
+    # If --force, silently ignore uncommitted deletions (not in the filesystem)
+    # and renames / additions (will not be found in log anyway)
+    if args.force:
+        filelist -= set(git.ls_dirty(force=True))
+    # Otherwise, ignore any dirty files
+    else:
+        dirty = set(git.ls_dirty())
+        if dirty:
+            log.warning("WARNING: Modified files in the working directory were ignored."
+                "\nTo include such files, commit your changes or use --force.")
+            filelist -= dirty
+
+    # Build dir list to be processed
+    dirlist = set(os.path.dirname(_) for _ in filelist) if args.dirs else set()
+
+    stats['totalfiles'] = stats['files'] = len(filelist)
+    log.info("{0:,} files to be processed in work dir".format(stats['totalfiles']))
+
+    if not filelist:
+        # Nothing to do. Exit silently and without errors, just like git does
+        return
+
+    # Process the log until all files are 'touched'
+    log.debug("Line #\tLog #\tF.Left\tModification Time\tFile Name")
+    parse_log(filelist, dirlist, stats, git, args.merge, args.pathspec)
+
+    # Missing files
+    if filelist:
+        # Try to find them in merge logs, if not done already
+        # (usually HUGE, thus MUCH slower!)
+        if args.missing and not args.merge:
+            filterlist = list(filelist)
+            missing = len(filterlist)
+            log.info("{0:,} files not found in log, trying merge commits".format(missing))
+            for i in range(0, missing, STEPMISSING):
+                parse_log(filelist, dirlist, stats, git,
+                          merge=True, filterlist=filterlist[i:i + STEPMISSING])
+
+        # Still missing some?
+        for file in filelist:
+            log.warning("WARNING: not found in the log: %s", file)
+
+    # Final statistics
+    # Suggestion: use git-log --before=mtime to brag about skipped log entries
+    def log_info(msg, *a, width=13):
+        ifmt = '{:%d,}'    % (width,)  # not using 'n' for consistency with ffmt
+        ffmt = '{:%d,.2f}' % (width,)
+        # %-formatting lacks a thousand separator, must pre-render with .format()
+        log.info(msg.replace('%d', ifmt).replace('%f', ffmt).format(*a))
+
+    log_info(
+        "Statistics:\n"
+        "%f seconds\n"
+        "%d log lines processed\n"
+        "%d commits evaluated",
+        time.time() - start, stats['loglines'], stats['commits'])
+
+    if args.dirs:
+        if stats['direrrors']: log_info("%d directory update errors", stats['direrrors'])
+        log_info("%d directories updated", stats['dirtouches'])
+
+    if stats['touches'] != stats['totalfiles']:
+                        log_info("%d files",              stats['totalfiles'])
+    if stats['skip']:   log_info("%d files skipped",      stats['skip'])
+    if stats['files']:  log_info("%d files missing",      stats['files'])
+    if stats['errors']: log_info("%d file update errors", stats['errors'])
+
+    log_info("%d files updated", stats['touches'])
+
+    if args.test:
+        log.info("TEST RUN - No files modified!")
+
+
+# Keep only essential, global assignments here. Any other logic must be in main()
+log = setup_logging()
+args = parse_args()
+
+# Set the actual touch() and other functions based on command-line arguments
+if args.unique_times:
+    touch = touch_ns
+    isodate = isodate_ns
+
+# Make sure this is always set last to ensure --test behaves as intended
+if args.test:
+    touch = dummy
+
+# UI done, it's showtime!
+try:
+    sys.exit(main())
+except KeyboardInterrupt:
+    log.info("\nAborting")
+    signal.signal(signal.SIGINT, signal.SIG_DFL)
+    os.kill(os.getpid(), signal.SIGINT)

.github/workflows/_lint.yml

@@ -9,38 +9,133 @@ on:
         description: "From which folder this pipeline executes"
 
 env:
-  POETRY_VERSION: "1.4.2"
+  POETRY_VERSION: "1.5.1"
+  WORKDIR: ${{ inputs.working-directory == '' && '.' || inputs.working-directory }}
 
 jobs:
   build:
-    defaults:
-      run:
-        working-directory: ${{ inputs.working-directory }}
     runs-on: ubuntu-latest
+    env:
+      # This number is set "by eye": we want it to be big enough
+      # so that it's bigger than the number of commits in any reasonable PR,
+      # and also as small as possible since increasing the number makes
+      # the initial `git fetch` slower.
+      FETCH_DEPTH: 50
     strategy:
       matrix:
+        # Only lint on the min and max supported Python versions.
+        # It's extremely unlikely that there's a lint issue on any version in between
+        # that doesn't show up on the min or max versions.
+        #
+        # GitHub rate-limits how many jobs can be running at any one time.
+        # Starting new jobs is also relatively slow,
+        # so linting on fewer versions makes CI faster.
         python-version:
           - "3.8"
-          - "3.9"
-          - "3.10"
           - "3.11"
     steps:
       - uses: actions/checkout@v3
+        with:
+          # Fetch the last FETCH_DEPTH commits, so the mtime-changing script
+          # can accurately set the mtimes of files modified in the last FETCH_DEPTH commits.
+          fetch-depth: ${{ env.FETCH_DEPTH }}
+      - name: Restore workdir file mtimes to last-edited commit date
+        id: restore-mtimes
+        # This is needed to make black caching work.
+        # Black's cache uses file (mtime, size) to check whether a lookup is a cache hit.
+        # Without this command, files in the repo would have the current time as the modified time,
+        # since the previous action step just created them.
+        # This command resets the mtime to the last time the files were modified in git instead,
+        # which is a high-quality and stable representation of the last modification date.
+        run: |
+          # Important considerations:
+          # - These commands run at base of the repo, since we never `cd` to the `WORKDIR`.
+          # - We only want to alter mtimes for Python files, since that's all black checks.
+          # - We don't need to alter mtimes for directories, since black doesn't look at those.
+          # - We also only alter mtimes inside the `WORKDIR` since that's all we'll lint.
+          # - This should run before `poetry install`, because poetry's venv also contains
+          #   Python files, and we don't want to alter their mtimes since they aren't linted.
+
+          # Ensure we fail on non-zero exits and on undefined variables.
+          # Also print executed commands, for easier debugging.
+          set -eux
+
+          # Restore the mtimes of Python files in the workdir based on git history.
+          .github/tools/git-restore-mtime --no-directories "$WORKDIR/**/*.py"
+
+          # Since CI only does a partial fetch (to `FETCH_DEPTH`) for efficiency,
+          # the local git repo doesn't have full history. There are probably files
+          # that were last modified in a commit *older than* the oldest fetched commit.
+          # After `git-restore-mtime`, such files have a mtime set to the oldest fetched commit.
+          #
+          # As new commits get added, that timestamp will keep moving forward.
+          # If left unchanged, this will make `black` think that the files were edited
+          # more recently than its cache suggests. Instead, we can set their mtime
+          # to a fixed date in the far past that won't change and won't cause cache misses in black.
+          #
+          # For all workdir Python files modified in or before the oldest few fetched commits,
+          # make their mtime be 2000-01-01 00:00:00.
+          OLDEST_COMMIT="$(git log --reverse '--pretty=format:%H' | head -1)"
+          OLDEST_COMMIT_TIME="$(git show -s '--format=%ai' "$OLDEST_COMMIT")"
+          find "$WORKDIR" -name '*.py' -type f -not -newermt "$OLDEST_COMMIT_TIME" -exec touch -c -m -t '200001010000' '{}' '+'
+
+          echo "oldest-commit=$OLDEST_COMMIT" >> "$GITHUB_OUTPUT"
+      - uses: actions/cache@v3
+        id: cache-pip
+        name: Cache langchain editable pip install - ${{ matrix.python-version }}
+        env:
+          SEGMENT_DOWNLOAD_TIMEOUT_MIN: "4"
+        with:
+          path: |
+            ~/.cache/pip
+          key: pip-editable-langchain-deps-${{ runner.os }}-${{ runner.arch }}-py-${{ matrix.python-version }}
       - name: Install poetry
         run: |
-          pipx install poetry==$POETRY_VERSION
+          pipx install "poetry==$POETRY_VERSION"
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v4
+        env:
+          SEGMENT_DOWNLOAD_TIMEOUT_MIN: "4"
         with:
           python-version: ${{ matrix.python-version }}
           cache: poetry
+          cache-dependency-path: |
+            ${{ env.WORKDIR }}/**/poetry.lock
       - name: Install dependencies
+        working-directory: ${{ inputs.working-directory }}
         run: |
           poetry install
       - name: Install langchain editable
-        if: ${{ inputs.working-directory != 'langchain' }}
+        working-directory: ${{ inputs.working-directory }}
+        if: ${{ inputs.working-directory != 'libs/langchain' }}
         run: |
           pip install -e ../langchain
+
+      - name: Restore black cache
+        uses: actions/cache@v3
+        env:
+          CACHE_BASE: black-${{ runner.os }}-${{ runner.arch }}-py${{ matrix.python-version }}-${{ inputs.working-directory }}-${{ hashFiles(format('{0}/poetry.lock', env.WORKDIR)) }}
+          SEGMENT_DOWNLOAD_TIMEOUT_MIN: "1"
+        with:
+          path: |
+            ${{ env.WORKDIR }}/.black_cache
+          key: ${{ env.CACHE_BASE }}-${{ steps.restore-mtimes.outputs.oldest-commit }}
+          restore-keys:
+            # If we can't find an exact match for our cache key, accept any with this prefix.
+            ${{ env.CACHE_BASE }}-
+
+      - name: Get .mypy_cache to speed up mypy
+        uses: actions/cache@v3
+        env:
+          SEGMENT_DOWNLOAD_TIMEOUT_MIN: "2"
+        with:
+          path: |
+            ${{ env.WORKDIR }}/.mypy_cache
+          key: mypy-${{ runner.os }}-${{ runner.arch }}-py${{ matrix.python-version }}-${{ inputs.working-directory }}-${{ hashFiles(format('{0}/poetry.lock', env.WORKDIR)) }}
+
       - name: Analysing the code with our lint
+        working-directory: ${{ inputs.working-directory }}
+        env:
+          BLACK_CACHE_DIR: .black_cache
         run: |
           make lint

.github/workflows/_pydantic_compatibility.yml

@@ -0,0 +1,76 @@
+name: pydantic v1/v2 compatibility
+
+on:
+  workflow_call:
+    inputs:
+      working-directory:
+        required: true
+        type: string
+        description: "From which folder this pipeline executes"
+
+env:
+  POETRY_VERSION: "1.5.1"
+
+jobs:
+  build:
+    defaults:
+      run:
+        working-directory: ${{ inputs.working-directory }}
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version:
+          - "3.8"
+          - "3.9"
+          - "3.10"
+          - "3.11"
+    name: Pydantic v1/v2 compatibility - 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 }}
+          working-directory: ${{ inputs.working-directory }}
+          poetry-version: ${{ env.POETRY_VERSION }}
+          cache-key: pydantic-cross-compat
+          install-command: poetry install
+      - name: Install the opposite major version of pydantic
+        # If normal tests use pydantic v1, here we'll use v2, and vice versa.
+        shell: bash
+        run: |
+          # Determine the major part of pydantic version
+          REGULAR_VERSION=$(poetry run python -c "import pydantic; print(pydantic.__version__)" | cut -d. -f1)
+
+          if [[ "$REGULAR_VERSION" == "1" ]]; then
+            PYDANTIC_DEP=">=2.1,<3"
+            TEST_WITH_VERSION="2"
+          elif [[ "$REGULAR_VERSION" == "2" ]]; then
+            PYDANTIC_DEP="<2"
+            TEST_WITH_VERSION="1"
+          else
+            echo "Unexpected pydantic major version '$REGULAR_VERSION', cannot determine which version to use for cross-compatibility test."
+            exit 1
+          fi
+
+          # Install via `pip` instead of `poetry add` to avoid changing lockfile,
+          # which would prevent caching from working: the cache would get saved
+          # to a different key than where it gets loaded from.
+          poetry run pip install "pydantic${PYDANTIC_DEP}"
+
+          # Ensure that the correct pydantic is installed now.
+          echo "Checking pydantic version... Expecting ${TEST_WITH_VERSION}"
+
+          # Determine the major part of pydantic version
+          CURRENT_VERSION=$(poetry run python -c "import pydantic; print(pydantic.__version__)" | cut -d. -f1)
+
+          # Check that the major part of pydantic version is as expected, if not
+          # raise an error
+          if [[ "$CURRENT_VERSION" != "$TEST_WITH_VERSION" ]]; then
+            echo "Error: expected pydantic version ${CURRENT_VERSION} to have been installed, but found: ${TEST_WITH_VERSION}"
+            exit 1
+          fi
+          echo "Found pydantic version ${CURRENT_VERSION}, as expected"
+      - name: Run pydantic compatibility tests
+        shell: bash
+        run: make test

.github/workflows/_release.yml

@@ -9,21 +9,30 @@ on:
         description: "From which folder this pipeline executes"
 
 env:
-  POETRY_VERSION: "1.4.2"
+  POETRY_VERSION: "1.5.1"
 
 jobs:
   if_release:
-    if: |
-        ${{ github.event.pull_request.merged == true }}
-        && ${{ contains(github.event.pull_request.labels.*.name, 'release') }}
+    # Disallow publishing from branches that aren't `master`.
+    if: github.ref == 'refs/heads/master'
     runs-on: ubuntu-latest
+    permissions:
+      # This permission is used for trusted publishing:
+      # https://blog.pypi.org/posts/2023-04-20-introducing-trusted-publishers/
+      #
+      # Trusted publishing has to also be configured on PyPI for each package:
+      # https://docs.pypi.org/trusted-publishers/adding-a-publisher/
+      id-token: write
+
+      # This permission is needed by `ncipollo/release-action` to create the GitHub release.
+      contents: write
     defaults:
       run:
         working-directory: ${{ inputs.working-directory }}
     steps:
       - uses: actions/checkout@v3
       - name: Install poetry
-        run: pipx install poetry==$POETRY_VERSION
+        run: pipx install "poetry==$POETRY_VERSION"
       - name: Set up Python 3.10
         uses: actions/setup-python@v4
         with:
@@ -37,6 +46,7 @@ jobs:
           echo version=$(poetry version --short) >> $GITHUB_OUTPUT
       - name: Create Release
         uses: ncipollo/release-action@v1
+        if: ${{ inputs.working-directory == 'libs/langchain' }}
         with:
           artifacts: "dist/*"
           token: ${{ secrets.GITHUB_TOKEN }}
@@ -44,8 +54,9 @@ jobs:
           generateReleaseNotes: true
           tag: v${{ steps.check-version.outputs.version }}
           commit: master
-      - name: Publish to PyPI
-        env:
-          POETRY_PYPI_TOKEN_PYPI: ${{ secrets.PYPI_API_TOKEN }}
-        run: |
-          poetry publish
+      - name: Publish package distributions to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: ${{ inputs.working-directory }}/dist/
+          verbose: true
+          print-hash: true

.github/workflows/_test.yml

@@ -7,13 +7,9 @@ on:
         required: true
         type: string
         description: "From which folder this pipeline executes"
-      test_type:
-        type: string
-        description: "Test types to run"
-        default: '["core", "extended"]'
 
 env:
-  POETRY_VERSION: "1.4.2"
+  POETRY_VERSION: "1.5.1"
 
 jobs:
   build:
@@ -28,8 +24,7 @@ jobs:
           - "3.9"
           - "3.10"
           - "3.11"
-        test_type: ${{ fromJSON(inputs.test_type) }}
-    name: Python ${{ matrix.python-version }} ${{ matrix.test_type }}
+    name: Python ${{ matrix.python-version }}
     steps:
       - uses: actions/checkout@v3
       - name: Set up Python ${{ matrix.python-version }}
@@ -37,25 +32,9 @@ jobs:
         with:
           python-version: ${{ matrix.python-version }}
           working-directory: ${{ inputs.working-directory }}
-          poetry-version: "1.4.2"
-          cache-key: ${{ matrix.test_type }}
-          install-command: |
-              if [ "${{ matrix.test_type }}" == "core" ]; then
-                echo "Running core tests, installing dependencies with poetry..."
-                poetry install
-              else
-                echo "Running extended tests, installing dependencies with poetry..."
-                poetry install -E extended_testing
-              fi
-      - name: Install langchain editable
-        if: ${{ inputs.working-directory != 'langchain' }}
-        run: |
-          pip install -e ../langchain
-      - name: Run ${{matrix.test_type}} tests
-        run: |
-          if [ "${{ matrix.test_type }}" == "core" ]; then
-            make test
-          else
-            make extended_tests
-          fi
+          poetry-version: ${{ env.POETRY_VERSION }}
+          cache-key: core
+          install-command: poetry install
+      - name: Run core tests
         shell: bash
+        run: make test

.github/workflows/codespell.yml

@@ -20,3 +20,5 @@ jobs:
         uses: actions/checkout@v3
       - name: Codespell
         uses: codespell-project/actions-codespell@v2
+        with:
+          skip: guide_imports.json

.github/workflows/langchain_ci.yml

@@ -8,10 +8,15 @@ on:
     paths:
       - '.github/workflows/_lint.yml'
       - '.github/workflows/_test.yml'
+      - '.github/workflows/_pydantic_compatibility.yml'
       - '.github/workflows/langchain_ci.yml'
       - 'libs/langchain/**'
   workflow_dispatch:  # Allows to trigger the workflow manually in GitHub UI
 
+env:
+  POETRY_VERSION: "1.5.1"
+  WORKDIR: "libs/langchain"
+
 jobs:
   lint:
     uses:
@@ -19,9 +24,45 @@ jobs:
     with:
       working-directory: libs/langchain
     secrets: inherit
+
   test:
     uses:
       ./.github/workflows/_test.yml
     with:
       working-directory: libs/langchain
-    secrets: inherit
+    secrets: inherit
+
+  pydantic-compatibility:
+    uses:
+      ./.github/workflows/_pydantic_compatibility.yml
+    with:
+      working-directory: libs/langchain
+    secrets: inherit
+
+  extended-tests:
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: ${{ env.WORKDIR }}
+    strategy:
+      matrix:
+        python-version:
+          - "3.8"
+          - "3.9"
+          - "3.10"
+          - "3.11"
+    name: Python ${{ matrix.python-version }} extended tests
+    steps:
+      - uses: actions/checkout@v3
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: "./.github/actions/poetry_setup"
+        with:
+          python-version: ${{ matrix.python-version }}
+          working-directory: ${{ env.WORKDIR }}
+          poetry-version: ${{ env.POETRY_VERSION }}
+          cache-key: extended
+          install-command: |
+            echo "Running extended tests, installing dependencies with poetry..."
+            poetry install -E extended_testing
+      - name: Run extended tests
+        run: make extended_tests

.github/workflows/langchain_experimental_ci.yml

@@ -1,5 +1,5 @@
 ---
-name: libs/langchain-experimental CI
+name: libs/experimental CI
 
 on:
   push:
@@ -25,5 +25,4 @@ jobs:
       ./.github/workflows/_test.yml
     with:
       working-directory: libs/experimental
-      test_type: '["core"]'
-    secrets: inherit
+    secrets: inherit

.github/workflows/langchain_experimental_release.yml

@@ -1,14 +1,7 @@
 ---
-name: libs/langchain-experimental Release
+name: libs/experimental Release
 
 on:
-  pull_request:
-    types:
-      - closed
-    branches:
-      - master
-    paths:
-      - 'libs/experimental/pyproject.toml'
   workflow_dispatch:  # Allows to trigger the workflow manually in GitHub UI
 
 jobs:
@@ -17,4 +10,4 @@ jobs:
       ./.github/workflows/_release.yml
     with:
       working-directory: libs/experimental
-    secrets: inherit
+    secrets: inherit

.github/workflows/langchain_release.yml

@@ -2,13 +2,6 @@
 name: libs/langchain Release
 
 on:
-  pull_request:
-    types:
-      - closed
-    branches:
-      - master
-    paths:
-      - 'libs/langchain/pyproject.toml'
   workflow_dispatch:  # Allows to trigger the workflow manually in GitHub UI
 
 jobs:
@@ -17,4 +10,4 @@ jobs:
       ./.github/workflows/_release.yml
     with:
       working-directory: libs/langchain
-    secrets: inherit
+    secrets: inherit

.github/workflows/scheduled_test.yml

@@ -0,0 +1,42 @@
+name: Scheduled tests
+
+on:
+  workflow_dispatch:  # Allows to trigger the workflow manually in GitHub UI
+  schedule:
+    - cron:  '0 13 * * *'
+
+env:
+  POETRY_VERSION: "1.5.1"
+
+jobs:
+  build:
+    defaults:
+      run:
+        working-directory: libs/langchain
+    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: ${{ env.POETRY_VERSION }}
+          working-directory: libs/langchain
+          install-command: |
+            echo "Running scheduled tests, installing dependencies with poetry..."
+            poetry install --with=test_integration
+      - name: Run tests
+        env:
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+        run: |
+          make scheduled_tests
+        shell: bash

.gitignore

@@ -162,6 +162,7 @@ docs/.docusaurus/
 docs/.cache-loader/
 docs/_dist
 docs/api_reference/api_reference.rst
+docs/api_reference/experimental_api_reference.rst
 docs/api_reference/_build
 docs/api_reference/*/
 !docs/api_reference/_static/

MIGRATE.md

@@ -1,6 +1,6 @@
-# Migrating to `langchain.experimental`
+# Migrating to `langchain_experimental`
 
-We are moving any experimental components of langchain, or components with vulnerability issues, into `langchain.experimental`.
+We are moving any experimental components of LangChain, or components with vulnerability issues, into `langchain_experimental`.
 This guide covers how to migrate.
 
 ## Installation
@@ -9,9 +9,19 @@ Previously:
 
 `pip install -U langchain`
 
+Now (only if you want to access things in experimental):
+
+`pip install -U langchain langchain_experimental`
+
+## Things in `langchain.experimental`
+
+Previously:
+
+`from langchain.experimental import ...`
+
 Now:
 
-`pip install -U langchain langchain.experimental`
+`from langchain_experimental import ...`
 
 ## PALChain
 
@@ -21,7 +31,7 @@ Previously:
 
 Now:
 
-`from langchain.experimental.pal_chain import PALChain`
+`from langchain_experimental.pal_chain import PALChain`
 
 ## SQLDatabaseChain
 
@@ -31,7 +41,11 @@ Previously:
 
 Now:
 
-`from langchain.experimental.sql import SQLDatabaseChain`
+`from langchain_experimental.sql import SQLDatabaseChain`
+
+Alternatively, if you are just interested in using the query generation part of the SQL chain, you can check out [`create_sql_query_chain`](https://github.com/langchain-ai/langchain/blob/master/docs/extras/use_cases/tabular/sql_query.ipynb)
+
+`from langchain.chains import create_sql_query_chain`
 
 ## `load_prompt` for Python files
 
@@ -44,4 +58,4 @@ Previously:
 
 Now:
 
-`from langchain.experimental.prompts import load_prompt`
+`from langchain_experimental.prompts import load_prompt`

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'

README.md

@@ -2,24 +2,32 @@
 
 ⚡ Building applications with LLMs through composability ⚡
 
-[![Release Notes](https://img.shields.io/github/release/hwchase17/langchain)](https://github.com/hwchase17/langchain/releases)
-[![CI](https://github.com/hwchase17/langchain/actions/workflows/langchain_ci.yml/badge.svg)](https://github.com/hwchase17/langchain/actions/workflows/langchain_ci.yml)
-[![Experimental CI](https://github.com/hwchase17/langchain/actions/workflows/langchain_experimental_ci.yml/badge.svg)](https://github.com/hwchase17/langchain/actions/workflows/langchain_experimental_ci.yml)
+[![Release Notes](https://img.shields.io/github/release/langchain-ai/langchain)](https://github.com/langchain-ai/langchain/releases)
+[![CI](https://github.com/langchain-ai/langchain/actions/workflows/langchain_ci.yml/badge.svg)](https://github.com/langchain-ai/langchain/actions/workflows/langchain_ci.yml)
+[![Experimental CI](https://github.com/langchain-ai/langchain/actions/workflows/langchain_experimental_ci.yml/badge.svg)](https://github.com/langchain-ai/langchain/actions/workflows/langchain_experimental_ci.yml)
 [![Downloads](https://static.pepy.tech/badge/langchain/month)](https://pepy.tech/project/langchain)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![Twitter](https://img.shields.io/twitter/url/https/twitter.com/langchainai.svg?style=social&label=Follow%20%40LangChainAI)](https://twitter.com/langchainai)
 [![](https://dcbadge.vercel.app/api/server/6adMQxSpJS?compact=true&style=flat)](https://discord.gg/6adMQxSpJS)
-[![Open in Dev Containers](https://img.shields.io/static/v1?label=Dev%20Containers&message=Open&color=blue&logo=visualstudiocode)](https://vscode.dev/redirect?url=vscode://ms-vscode-remote.remote-containers/cloneInVolume?url=https://github.com/hwchase17/langchain)
-[![Open in GitHub Codespaces](https://github.com/codespaces/badge.svg)](https://codespaces.new/hwchase17/langchain)
-[![GitHub star chart](https://img.shields.io/github/stars/hwchase17/langchain?style=social)](https://star-history.com/#hwchase17/langchain)
-[![Dependency Status](https://img.shields.io/librariesio/github/hwchase17/langchain)](https://libraries.io/github/hwchase17/langchain)
-[![Open Issues](https://img.shields.io/github/issues-raw/hwchase17/langchain)](https://github.com/hwchase17/langchain/issues)
+[![Open in Dev Containers](https://img.shields.io/static/v1?label=Dev%20Containers&message=Open&color=blue&logo=visualstudiocode)](https://vscode.dev/redirect?url=vscode://ms-vscode-remote.remote-containers/cloneInVolume?url=https://github.com/langchain-ai/langchain)
+[![Open in GitHub Codespaces](https://github.com/codespaces/badge.svg)](https://codespaces.new/langchain-ai/langchain)
+[![GitHub star chart](https://img.shields.io/github/stars/langchain-ai/langchain?style=social)](https://star-history.com/#langchain-ai/langchain)
+[![Dependency Status](https://img.shields.io/librariesio/github/langchain-ai/langchain)](https://libraries.io/github/langchain-ai/langchain)
+[![Open Issues](https://img.shields.io/github/issues-raw/langchain-ai/langchain)](https://github.com/langchain-ai/langchain/issues)
 
 
 Looking for the JS/TS version? Check out [LangChain.js](https://github.com/hwchase17/langchainjs).
 
-**Production Support:** As you move your LangChains into production, we'd love to offer more comprehensive support.
-Please fill out [this form](https://forms.gle/57d8AmXBYp8PP8tZA) and we'll set up a dedicated support Slack channel.
+**Production Support:** As you move your LangChains into production, we'd love to offer more hands-on support.
+Fill out [this form](https://airtable.com/appwQzlErAS2qiP0L/shrGtGaVBVAz7NcV2) to share more about what you're building, and our team will get in touch.
+
+## 🚨Breaking Changes for select chains (SQLDatabase) on 7/28/23
+
+In an effort to make `langchain` leaner and safer, we are moving select chains to `langchain_experimental`.
+This migration has already started, but we are remaining backwards compatible until 7/28.
+On that date, we will remove functionality from `langchain`.
+Read more about the motivation and the progress [here](https://github.com/hwchase17/langchain/discussions/8043).
+Read how to migrate your code [here](MIGRATE.md).
 
 ## Quick Install
 

SECURITY.md

@@ -0,0 +1,6 @@
+# Security Policy
+
+## Reporting a Vulnerability
+
+Please report security vulnerabilities by email to `security@langchain.dev`.
+This email is an alias to a subset of our maintainers, and will ensure the issue is promptly triaged and acted upon as needed.

docs/.local_build.sh

@@ -13,5 +13,6 @@ cp -r {docs_skeleton,snippets} _dist
 cp -r extras/* _dist/docs_skeleton/docs
 cd _dist/docs_skeleton
 poetry run nbdoc_build
+poetry run python generate_api_reference_links.py
 yarn install
 yarn start

docs/api_reference/conf.py

@@ -7,20 +7,67 @@
 
 # -- Path setup --------------------------------------------------------------
 
+import json
+import os
+import sys
+from pathlib import Path
+
+import toml
+from docutils import nodes
+from sphinx.util.docutils import SphinxDirective
+
 # If extensions (or modules to document with autodoc) are in another directory,
 # add these directories to sys.path here. If the directory is relative to the
 # documentation root, use os.path.abspath to make it absolute, like shown here.
-#
-import os
-import sys
-
-import toml
 
+_DIR = Path(__file__).parent.absolute()
 sys.path.insert(0, os.path.abspath("."))
 sys.path.insert(0, os.path.abspath("../../libs/langchain"))
+sys.path.insert(0, os.path.abspath("../../libs/experimental"))
 
-with open("../../libs/langchain/pyproject.toml") as f:
+with (_DIR.parents[1] / "libs" / "langchain" / "pyproject.toml").open("r") as f:
     data = toml.load(f)
+with (_DIR / "guide_imports.json").open("r") as f:
+    imported_classes = json.load(f)
+
+
+class ExampleLinksDirective(SphinxDirective):
+    """Directive to generate a list of links to examples.
+
+    We have a script that extracts links to API reference docs
+    from our notebook examples. This directive uses that information
+    to backlink to the examples from the API reference docs."""
+
+    has_content = False
+    required_arguments = 1
+
+    def run(self):
+        """Run the directive.
+
+        Called any time :example_links:`ClassName` is used
+        in the template *.rst files."""
+        class_or_func_name = self.arguments[0]
+        links = imported_classes.get(class_or_func_name, {})
+        list_node = nodes.bullet_list()
+        for doc_name, link in links.items():
+            item_node = nodes.list_item()
+            para_node = nodes.paragraph()
+            link_node = nodes.reference()
+            link_node["refuri"] = link
+            link_node.append(nodes.Text(doc_name))
+            para_node.append(link_node)
+            item_node.append(para_node)
+            list_node.append(item_node)
+        if list_node.children:
+            title_node = nodes.title()
+            title_node.append(nodes.Text(f"Examples using {class_or_func_name}"))
+            return [title_node, list_node]
+        return [list_node]
+
+
+def setup(app):
+    app.add_directive("example_links", ExampleLinksDirective)
+
 
 # -- Project information -----------------------------------------------------
 
@@ -53,6 +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
@@ -65,13 +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

@@ -1,83 +1,264 @@
-"""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"
-WRITE_FILE = Path(__file__).parent / "api_reference.rst"
+EXP_DIR = ROOT_DIR / "libs" / "experimental" / "langchain_experimental"
+WRITE_FILE = HERE / "api_reference.rst"
+EXP_WRITE_FILE = HERE / "experimental_api_reference.rst"
 
 
-def load_members() -> dict:
-    members: dict = {}
-    for py in glob.glob(str(PKG_DIR) + "/**/*.py", recursive=True):
-        module = py[len(str(PKG_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(members: dict) -> str:
-    full_doc = """\
-.. _api_reference:
+class ClassInfo(TypedDict):
+    """Information about a class."""
 
-=============
-API Reference
-=============
+    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 file_path.name.startswith("_"):
+            continue
+
+        relative_module_name = file_path.relative_to(package_path)
+
+        # Skip if any module part starts with an underscore
+        if any(part.startswith("_") for part in relative_module_name.parts):
+            continue
+
+        # 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
-
-        module_title = module.replace("_", " ").title()
-        if module_title == "Llms":
-            module_title = "LLMs"
-        section = f":mod:`langchain.{module}`: {module_title}"
+        section = f":mod:`{pkg}.{module}`"
+        underline = "=" * (len(section) + 1)
         full_doc += f"""\
 {section}
-{'=' * (len(section) + 1)}
+{underline}
 
-.. automodule:: langchain.{module}
+.. automodule:: {pkg}.{module}
     :no-members:
     :no-inherited-members:
 
 """
 
         if classes:
-            cstring = "\n    ".join(sorted(classes))
             full_doc += f"""\
 Classes
 --------------
-.. currentmodule:: langchain
+.. currentmodule:: {pkg}
 
 .. 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
 --------------
-.. currentmodule:: langchain
+.. currentmodule:: {pkg}
 
 .. autosummary::
     :toctree: {module}
+    :template: function.rst
 
     {fstring}
 
@@ -86,10 +267,17 @@ Functions
 
 
 def main() -> None:
-    members = load_members()
-    full_doc = construct_doc(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(full_doc)
+        f.write(lc_doc)
+    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)
 
 
 if __name__ == "__main__":

docs/api_reference/modules/evaluation.rst

@@ -1,9 +0,0 @@
-Evaluation
-=======================
-
-LangChain has a number of convenient evaluation chains you can use off the shelf to grade your models' oupputs.
-
-.. automodule:: langchain.evaluation
-   :members:
-   :undoc-members:
-   :inherited-members:

docs/api_reference/requirements.txt

@@ -1,4 +1,6 @@
 -e libs/langchain
+-e libs/experimental
+pydantic<2
 autodoc_pydantic==1.8.0
 myst_parser
 nbsphinx==0.8.9
@@ -10,4 +12,4 @@ 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') }}
@@ -26,3 +15,22 @@
    {%- endfor %}
    {% 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/function.rst

@@ -0,0 +1,8 @@
+:mod:`{{module}}`.{{objname}}
+{{ underline }}==============
+
+.. currentmodule:: {{ module }}
+
+.. autofunction:: {{ objname }}
+
+.. 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 %}
@@ -45,6 +34,9 @@
         <li class="nav-item">
           <a class="sk-nav-link nav-link" href="{{ pathto('api_reference') }}">API</a>
         </li>
+        <li class="nav-item">
+          <a class="sk-nav-link nav-link" href="{{ pathto('experimental_api_reference') }}">Experimental</a>
+        </li>
         <li class="nav-item">
           <a class="sk-nav-link nav-link" target="_blank" rel="noopener noreferrer" href="https://python.langchain.com/">Python Docs</a>
         </li>

docs/api_reference/themes/scikit-learn-modern/static/css/theme.css

@@ -745,6 +745,11 @@ span.descname {
   background-color: transparent;
   padding: 0;
   font-family: monospace;
+  font-size: 1.2rem;
+}
+
+em.property {
+  font-weight: normal;
 }
 
 span.descclassname {

docs/docs_skeleton/docs/community.md

@@ -0,0 +1,54 @@
+# Community navigator
+
+Hi! Thanks for being here. We’re lucky to have a community of so many passionate developers building with LangChain–we have so much to teach and learn from each other. Community members contribute code, host meetups, write blog posts, amplify each other’s work, become each other's customers and collaborators, and so much more.
+
+Whether you’re new to LangChain, looking to go deeper, or just want to get more exposure to the world of building with LLMs, this page can point you in the right direction. 
+
+- **🦜 Contribute to LangChain**
+
+- **🌍 Meetups, Events, and Hackathons**
+
+- **📣 Help Us Amplify Your Work**
+
+- **💬 Stay in the loop**
+
+
+# 🦜 Contribute to LangChain
+
+LangChain is the product of over 5,000+ contributions by 1,500+ contributors, and there is ******still****** so much to do together. Here are some ways to get involved:
+
+- **[Open a pull request](https://github.com/langchain-ai/langchain/issues):** we’d appreciate all forms of contributions–new features, infrastructure improvements, better documentation, bug fixes, etc. If you have an improvement or an idea, we’d love to work on it with you.
+- **[Read our contributor guidelines:](https://github.com/langchain-ai/langchain/blob/bbd22b9b761389a5e40fc45b0570e1830aabb707/.github/CONTRIBUTING.md)** We ask contributors to follow a ["fork and pull request"](https://docs.github.com/en/get-started/quickstart/contributing-to-projects) workflow, run a few local checks for formatting, linting, and testing before submitting, and follow certain documentation and testing conventions.
+    - **First time contributor?** [Try one of these PRs with the “good first issue” tag](https://github.com/langchain-ai/langchain/contribute).
+- **Become an expert:** our experts help the community by answering product questions in Discord. If that’s a role you’d like to play, we’d be so grateful! (And we have some special experts-only goodies/perks we can tell you more about). Send us an email to introduce yourself at hello@langchain.dev and we’ll take it from there!
+- **Integrate with LangChain:** if your product integrates with LangChain–or aspires to–we want to help make sure the experience is as smooth as possible for you and end users. Send us an email at hello@langchain.dev and tell us what you’re working on.
+    - **Become an Integration Maintainer:** Partner with our team to ensure your integration stays up-to-date and talk directly with users (and answer their inquiries) in our Discord. Introduce yourself at hello@langchain.dev if you’d like to explore this role.
+
+
+# 🌍 Meetups, Events, and Hackathons
+
+One of our favorite things about working in AI is how much enthusiasm there is for building together. We want to help make that as easy and impactful for you as possible! 
+- **Find a meetup, hackathon, or webinar:** you can find the one for you on our [global events calendar](https://mirror-feeling-d80.notion.site/0bc81da76a184297b86ca8fc782ee9a3?v=0d80342540df465396546976a50cfb3f).  
+    - **Submit an event to our calendar:** email us at events@langchain.dev with a link to your event page! We can also help you spread the word with our local communities.
+- **Host a meetup:** If you want to bring a group of builders together, we want to help! We can publicize your event on our event calendar/Twitter, share with our local communities in Discord, send swag, or potentially hook you up with a sponsor. Email us at events@langchain.dev to tell us about your event!
+- **Become a meetup sponsor:** we often hear from groups of builders that want to get together, but are blocked or limited on some dimension (space to host, budget for snacks, prizes to distribute, etc.). If you’d like to help, send us an email to events@langchain.dev we can share more about how it works!
+- **Speak at an event:** meetup hosts are always looking for great speakers, presenters, and panelists. If you’d like to do that at an event, send us an email to hello@langchain.dev with more information about yourself, what you want to talk about, and what city you’re based in and we’ll try to match you with an upcoming event!
+- **Tell us about your LLM community:** If you host or participate in a community that would welcome support from LangChain and/or our team, send us an email at hello@langchain.dev and let us know how we can help.
+
+# 📣 Help Us Amplify Your Work
+
+If you’re working on something you’re proud of, and think the LangChain community would benefit from knowing about it, we want to help you show it off.
+
+- **Post about your work and mention us:** we love hanging out on Twitter to see what people in the space are talking about and working on. If you tag [@langchainai](https://twitter.com/LangChainAI), we’ll almost certainly see it and can show you some love.
+- **Publish something on our blog:** if you’re writing about your experience building with LangChain, we’d love to post (or crosspost) it on our blog! E-mail hello@langchain.dev with a draft of your post! Or even an idea for something you want to write about.
+- **Get your product onto our [integrations hub](https://integrations.langchain.com/):** Many developers take advantage of our seamless integrations with other products, and come to our integrations hub to find out who those are. If you want to get your product up there, tell us about it (and how it works with LangChain) at hello@langchain.dev.
+
+# ☀️ Stay in the loop
+
+Here’s where our team hangs out, talks shop, spotlights cool work, and shares what we’re up to. We’d love to see you there too.
+
+- **[Twitter](https://twitter.com/LangChainAI):** we post about what we’re working on and what cool things we’re seeing in the space. If you tag @langchainai in your post, we’ll almost certainly see it, and can snow you some love!
+- **[Discord](https://discord.gg/6adMQxSpJS):** connect with with >30k developers who are building with LangChain
+- **[GitHub](https://github.com/langchain-ai/langchain):** open pull requests, contribute to a discussion, and/or contribute
+- **[Subscribe to our bi-weekly Release Notes](https://6w1pwbss0py.typeform.com/to/KjZB1auB):** a twice/month email roundup of the coolest things going on in our orbit
+- **Slack:** if you’re building an application in production at your company, we’d love to get into a Slack channel together. Fill out [this form](https://airtable.com/appwQzlErAS2qiP0L/shrGtGaVBVAz7NcV2) and we’ll get in touch about setting one up.

docs/docs_skeleton/docs/get_started/introduction.mdx

@@ -28,7 +28,7 @@ LangChain provides standard, extendable interfaces and external integrations for
 
 #### [Model I/O](/docs/modules/model_io/)
 Interface with language models
-#### [Data connection](/docs/modules/data_connection/)
+#### [Retrieval](/docs/modules/data_connection/)
 Interface with application-specific data
 #### [Chains](/docs/modules/chains/)
 Construct sequences of calls
@@ -51,7 +51,7 @@ Walkthroughs and best-practices for common end-to-end use cases, like:
 Learn best practices for developing with LangChain.
 
 ### [Ecosystem](/docs/ecosystem/)
-LangChain is part of a rich ecosystem of tools that integrate with our framework and build on top of it. Check out our growing list of [integrations](/docs/ecosystem/integrations/) and [dependent repos](/docs/ecosystem/dependents.html).
+LangChain is part of a rich ecosystem of tools that integrate with our framework and build on top of it. Check out our growing list of [integrations](/docs/integrations/) and [dependent repos](/docs/ecosystem/dependents).
 
 ### [Additional resources](/docs/additional_resources/)
 Our community is full of prolific developers, creative builders, and fantastic teachers. Check out [YouTube tutorials](/docs/additional_resources/youtube.html) for great tutorials from folks in the community, and [Gallery](https://github.com/kyrolabs/awesome-langchain) for a list of awesome LangChain projects, compiled by the folks at [KyroLabs](https://kyrolabs.com).

docs/docs_skeleton/docs/get_started/quickstart.mdx

@@ -22,28 +22,74 @@ import OpenAISetup from "@snippets/get_started/quickstart/openai_setup.mdx"
 
 ## Building an application
 
-Now we can start building our language model application. LangChain provides many modules that can be used to build language model applications. Modules can be used as stand-alones in simple applications and they can be combined for more complex use cases.
+Now we can start building our language model application. LangChain provides many modules that can be used to build language model applications.
+Modules can be used as stand-alones in simple applications and they can be combined for more complex use cases.
+
+The core building block of LangChain applications is the LLMChain.
+This combines three things:
+- LLM: The language model is the core reasoning engine here. In order to work with LangChain, you need to understand the different types of language models and how to work with them.
+- Prompt Templates: This provides instructions to the language model. This controls what the language model outputs, so understanding how to construct prompts and different prompting strategies is crucial.
+- Output Parsers: These translate the raw response from the LLM to a more workable format, making it easy to use the output downstream.
+
+In this getting started guide we will cover those three components by themselves, and then cover the LLMChain which combines all of them.
+Understanding these concepts will set you up well for being able to use and customize LangChain applications.
+Most LangChain applications allow you to configure the LLM and/or the prompt used, so knowing how to take advantage of this will be a big enabler.
 
 ## LLMs
-#### Get predictions from a language model
 
-The basic building block of LangChain is the LLM, which takes in text and generates more text.
+There are two types of language models, which in LangChain are called:
 
-As an example, suppose we're building an application that generates a company name based on a company description. In order to do this, we need to initialize an OpenAI model wrapper. In this case, since we want the outputs to be MORE random, we'll initialize our model with a HIGH temperature.
+- LLMs: this is a language model which takes a string as input and returns a string
+- ChatModels: this is a language model which takes a list of messages as input and returns a message
 
-import LLM from "@snippets/get_started/quickstart/llm.mdx"
+The input/output for LLMs is simple and easy to understand - a string.
+But what about ChatModels? The input there is a list of `ChatMessage`s, and the output is a single `ChatMessage`.
+A `ChatMessage` has two required components:
 
-<LLM/>
+- `content`: This is the content of the message.
+- `role`: This is the role of the entity from which the `ChatMessage` is coming from.
 
-## Chat models
+LangChain provides several objects to easily distinguish between different roles:
 
-Chat models are a variation on language models. While chat models use language models under the hood, the interface they expose is a bit different: rather than expose a "text in, text out" API, they expose an interface where "chat messages" are the inputs and outputs.
+- `HumanMessage`: A `ChatMessage` coming from a human/user.
+- `AIMessage`: A `ChatMessage` coming from an AI/assistant.
+- `SystemMessage`: A `ChatMessage` coming from the system.
+- `FunctionMessage`: A `ChatMessage` coming from a function call.
 
-You can get chat completions by passing one or more messages to the chat model. The response will be a message. The types of messages currently supported in LangChain are `AIMessage`, `HumanMessage`, `SystemMessage`, and `ChatMessage` -- `ChatMessage` takes in an arbitrary role parameter. Most of the time, you'll just be dealing with `HumanMessage`, `AIMessage`, and `SystemMessage`.
+If none of those roles sound right, there is also a `ChatMessage` class where you can specify the role manually.
+For more information on how to use these different messages most effectively, see our prompting guide.
 
-import ChatModel from "@snippets/get_started/quickstart/chat_model.mdx"
+LangChain exposes a standard interface for both, but it's useful to understand this difference in order to construct prompts for a given language model.
+The standard interface that LangChain exposes has two methods:
+- `predict`: Takes in a string, returns a string
+- `predict_messages`: Takes in a list of messages, returns a message.
+
+Let's see how to work with these different types of models and these different types of inputs.
+First, let's import an LLM and a ChatModel.
+
+import ImportLLMs from "@snippets/get_started/quickstart/import_llms.mdx"
+
+<ImportLLMs/>
+
+The `OpenAI` and `ChatOpenAI` objects are basically just configuration objects.
+You can initialize them with parameters like `temperature` and others, and pass them around.
+
+Next, let's use the `predict` method to run over a string input.
+
+import InputString from "@snippets/get_started/quickstart/input_string.mdx"
+
+<InputString/>
+
+Finally, let's use the `predict_messages` method to run over a list of messages.
+
+import InputMessages from "@snippets/get_started/quickstart/input_messages.mdx"
+
+<InputMessages/>
+
+For both these methods, you can also pass in parameters as key word arguments.
+For example, you could pass in `temperature=0` to adjust the temperature that is used from what the object was configured with.
+Whatever values are passed in during run time will always override what the object was configured with.
 
-<ChatModel/>
 
 ## Prompt templates
 
@@ -51,108 +97,66 @@ Most LLM applications do not pass user input directly into an LLM. Usually they
 
 In the previous example, the text we passed to the model contained instructions to generate a company name. For our application, it'd be great if the user only had to provide the description of a company/product, without having to worry about giving the model instructions.
 
+PromptTemplates help with exactly this!
+They bundle up all the logic for going from user input into a fully formatted prompt.
+This can start off very simple - for example, a prompt to produce the above string would just be:
+
 import PromptTemplateLLM from "@snippets/get_started/quickstart/prompt_templates_llms.mdx"
 import PromptTemplateChatModel from "@snippets/get_started/quickstart/prompt_templates_chat_models.mdx"
 
-<Tabs>
-    <TabItem value="llms" label="LLMs" default>
-
-With PromptTemplates this is easy! In this case our template would be very simple:
-
 <PromptTemplateLLM/>
-</TabItem>
-<TabItem value="chat_models" label="Chat models">
 
-Similar to LLMs, you can make use of templating by using a `MessagePromptTemplate`. You can build a `ChatPromptTemplate` from one or more `MessagePromptTemplate`s. You can use `ChatPromptTemplate`'s `format_messages` method to generate the formatted messages.
+However, the advantages of using these over raw string formatting are several.
+You can "partial" out variables - eg you can format only some of the variables at a time.
+You can compose them together, easily combining different templates into a single prompt.
+For explanations of these functionalities, see the [section on prompts](/docs/modules/model_io/prompts) for more detail.
 
-Because this is generating a list of messages, it is slightly more complex than the normal prompt template which is generating only a string. Please see the detailed guides on prompts to understand more options available to you here.
+PromptTemplates can also be used to produce a list of messages.
+In this case, the prompt not only contains information about the content, but also each message (its role, its position in the list, etc)
+Here, what happens most often is a ChatPromptTemplate is a list of ChatMessageTemplates.
+Each ChatMessageTemplate contains instructions for how to format that ChatMessage - its role, and then also its content.
+Let's take a look at this below:
 
 <PromptTemplateChatModel/>
-    </TabItem>
-</Tabs>
 
-## Chains
+ChatPromptTemplates can also include other things besides ChatMessageTemplates - see the [section on prompts](/docs/modules/model_io/prompts) for more detail.
 
-Now that we've got a model and a prompt template, we'll want to combine the two. Chains give us a way to link (or chain) together multiple primitives, like models, prompts, and other chains.
+## Output Parsers
 
-import ChainLLM from "@snippets/get_started/quickstart/chains_llms.mdx"
-import ChainChatModel from "@snippets/get_started/quickstart/chains_chat_models.mdx"
+OutputParsers convert the raw output of an LLM into a format that can be used downstream.
+There are few main type of OutputParsers, including:
 
-<Tabs>
-<TabItem value="llms" label="LLMs" default>
+- Convert text from LLM -> structured information (eg JSON)
+- Convert a ChatMessage into just a string
+- Convert the extra information returned from a call besides the message (like OpenAI function invocation) into a string.
 
-The simplest and most common type of chain is an LLMChain, which passes an input first to a PromptTemplate and then to an LLM. We can construct an LLM chain from our existing model and prompt template.
+For full information on this, see the [section on output parsers](/docs/modules/model_io/output_parsers)
 
-<ChainLLM/>
+In this getting started guide, we will write our own output parser - one that converts a comma separated list into a list.
 
-There we go, our first chain! Understanding how this simple chain works will set you up well for working with more complex chains.
+import OutputParser from "@snippets/get_started/quickstart/output_parser.mdx"
 
-</TabItem>
-<TabItem value="chat_models" label="Chat models">
+<OutputParser/>
 
-The `LLMChain` can be used with chat models as well:
+## LLMChain
 
-<ChainChatModel/>
-</TabItem>
-</Tabs>
+We can now combine all these into one chain.
+This chain will take input variables, pass those to a prompt template to create a prompt, pass the prompt to an LLM, and then pass the output through an (optional) output parser.
+This is a convenient way to bundle up a modular piece of logic.
+Let's see it in action!
 
-## Agents
+import LLMChain from "@snippets/get_started/quickstart/llm_chain.mdx"
 
-import AgentLLM from "@snippets/get_started/quickstart/agents_llms.mdx"
-import AgentChatModel from "@snippets/get_started/quickstart/agents_chat_models.mdx"
+<LLMChain/>
 
-Our first chain ran a pre-determined sequence of steps. To handle complex workflows, we need to be able to dynamically choose actions based on inputs.
+## Next Steps
 
-Agents do just this: they use a language model to determine which actions to take and in what order. Agents are given access to tools, and they repeatedly choose a tool, run the tool, and observe the output until they come up with a final answer.
+This is it!
+We've now gone over how to create the core building block of LangChain applications - the LLMChains.
+There is a lot more nuance in all these components (LLMs, prompts, output parsers) and a lot more different components to learn about as well.
+To continue on your journey:
 
-To load an agent, you need to choose a(n):
-- LLM/Chat model: The language model powering the agent.
-- Tool(s): A function that performs a specific duty. This can be things like: Google Search, Database lookup, Python REPL, other chains. For a list of predefined tools and their specifications, see the [Tools documentation](/docs/modules/agents/tools/).
-- Agent name: A string that references a supported agent class. An agent class is largely parameterized by the prompt the language model uses to determine which action to take. Because this notebook focuses on the simplest, highest level API, this only covers using the standard supported agents. If you want to implement a custom agent, see [here](/docs/modules/agents/how_to/custom_agent.html). For a list of supported agents and their specifications, see [here](/docs/modules/agents/agent_types/).
-
-For this example, we'll be using SerpAPI to query a search engine.
-
-You'll need to install the SerpAPI Python package:
-
-```bash
-pip install google-search-results
-```
-
-And set the `SERPAPI_API_KEY` environment variable.
-
-<Tabs>
-<TabItem value="llms" label="LLMs" default>
-<AgentLLM/>
-</TabItem>
-<TabItem value="chat_models" label="Chat models">
-
-Agents can also be used with chat models, you can initialize one using `AgentType.CHAT_ZERO_SHOT_REACT_DESCRIPTION` as the agent type.
-
-<AgentChatModel/>
-</TabItem>
-</Tabs>
-
-## Memory
-
-The chains and agents we've looked at so far have been stateless, but for many applications it's necessary to reference past interactions. This is clearly the case with a chatbot for example, where you want it to understand new messages in the context of past messages.
-
-The Memory module gives you a way to maintain application state. The base Memory interface is simple: it lets you update state given the latest run inputs and outputs and it lets you modify (or contextualize) the next input using the stored state.
-
-There are a number of built-in memory systems. The simplest of these is a buffer memory which just prepends the last few inputs/outputs to the current input - we will use this in the example below.
-
-import MemoryLLM from "@snippets/get_started/quickstart/memory_llms.mdx"
-import MemoryChatModel from "@snippets/get_started/quickstart/memory_chat_models.mdx"
-
-<Tabs>
-<TabItem value="llms" label="LLMs" default>
-
-<MemoryLLM/>
-</TabItem>
-<TabItem value="chat_models" label="Chat models">
-
-You can use Memory with chains and agents initialized with chat models. The main difference between this and Memory for LLMs is that rather than trying to condense all previous messages into a string, we can keep them as their own unique memory object.
-
-<MemoryChatModel/>
-
-</TabItem>
-</Tabs>
+- [Dive deeper](/docs/modules/model_io) into LLMs, prompts, and output parsers
+- Learn the other [key components](/docs/modules)
+- Check out our [helpful guides](/docs/guides) for detailed walkthroughs on particular topics
+- Explore [end-to-end use cases](/docs/use_cases)

docs/docs_skeleton/docs/guides/evaluation/comparison/index.mdx

@@ -0,0 +1,24 @@
+---
+sidebar_position: 3 
+---
+# Comparison Evaluators
+
+Comparison evaluators in LangChain help measure two different chain or LLM outputs. These evaluators are helpful for comparative analyses, such as A/B testing between two language models, or comparing different versions of the same model. They can also be useful for things like generating preference scores for ai-assisted reinforcement learning.
+
+These evaluators inherit from the `PairwiseStringEvaluator` class, providing a comparison interface for two strings - typically, the outputs from two different prompts or models, or two versions of the same model. In essence, a comparison evaluator performs an evaluation on a pair of strings and returns a dictionary containing the evaluation score and other relevant details.
+
+To create a custom comparison evaluator, inherit from the `PairwiseStringEvaluator` class and overwrite the `_evaluate_string_pairs` method. If you require asynchronous evaluation, also overwrite the `_aevaluate_string_pairs` method.
+
+Here's a summary of the key methods and properties of a comparison evaluator:
+
+- `evaluate_string_pairs`: Evaluate the output string pairs. This function should be overwritten when creating custom evaluators.
+- `aevaluate_string_pairs`: Asynchronously evaluate the output string pairs. This function should be overwritten for asynchronous evaluation.
+- `requires_input`: This property indicates whether this evaluator requires an input string.
+- `requires_reference`: This property specifies whether this evaluator requires a reference label.
+
+Detailed information about creating custom evaluators and the available built-in comparison evaluators are provided in the following sections.
+
+import DocCardList from "@theme/DocCardList";
+
+<DocCardList />
+

docs/docs_skeleton/docs/guides/evaluation/index.mdx

@@ -0,0 +1,31 @@
+---
+sidebar_position: 6
+---
+
+import DocCardList from "@theme/DocCardList";
+
+# Evaluation
+
+Building applications with language models involves many moving parts. One of the most critical components is ensuring that the outcomes produced by your models are reliable and useful across a broad array of inputs, and that they work well with your application's other software components. Ensuring reliability usually boils down to some combination of application design, testing & evaluation, and runtime checks. 
+
+The guides in this section review the APIs and functionality LangChain provides to help you better evaluate your applications. Evaluation and testing are both critical when thinking about deploying LLM applications, since production environments require repeatable and useful outcomes.
+
+LangChain offers various types of evaluators to help you measure performance and integrity on diverse data, and we hope to encourage the community to create and share other useful evaluators so everyone can improve. These docs will introduce the evaluator types, how to use them, and provide some examples of their use in real-world scenarios.
+
+Each evaluator type in LangChain comes with ready-to-use implementations and an extensible API that allows for customization according to your unique requirements. Here are some of the types of evaluators we offer:
+
+- [String Evaluators](/docs/guides/evaluation/string/): These evaluators assess the predicted string for a given input, usually comparing it against a reference string.
+- [Trajectory Evaluators](/docs/guides/evaluation/trajectory/): These are used to evaluate the entire trajectory of agent actions.
+- [Comparison Evaluators](/docs/guides/evaluation/comparison/): These evaluators are designed to compare predictions from two runs on a common input.
+
+These evaluators can be used across various scenarios and can be applied to different chain and LLM implementations in the LangChain library.
+
+We also are working to share guides and cookbooks that demonstrate how to use these evaluators in real-world scenarios, such as:
+
+- [Chain Comparisons](/docs/guides/evaluation/examples/comparisons): This example uses a comparison evaluator to predict the preferred output. It reviews ways to measure confidence intervals to select statistically significant differences in aggregate preference scores across different models or prompts.
+
+## Reference Docs
+
+For detailed information on the available evaluators, including how to instantiate, configure, and customize them, check out the [reference documentation](https://api.python.langchain.com/en/latest/api_reference.html#module-langchain.evaluation) directly.
+
+<DocCardList />

docs/docs_skeleton/docs/guides/evaluation/string/index.mdx

@@ -0,0 +1,27 @@
+---
+sidebar_position: 2 
+---
+# String Evaluators
+
+A string evaluator is a component within LangChain designed to assess the performance of a language model by comparing its generated outputs (predictions) to a reference string or an input. This comparison is a crucial step in the evaluation of language models, providing a measure of the accuracy or quality of the generated text.
+
+In practice, string evaluators are typically used to evaluate a predicted string against a given input, such as a question or a prompt. Often, a reference label or context string is provided to define what a correct or ideal response would look like. These evaluators can be customized to tailor the evaluation process to fit your application's specific requirements.
+
+To create a custom string evaluator, inherit from the `StringEvaluator` class and implement the `_evaluate_strings` method. If you require asynchronous support, also implement the `_aevaluate_strings` method.
+
+Here's a summary of the key attributes and methods associated with a string evaluator:
+
+- `evaluation_name`: Specifies the name of the evaluation.
+- `requires_input`: Boolean attribute that indicates whether the evaluator requires an input string. If True, the evaluator will raise an error when the input isn't provided. If False, a warning will be logged if an input _is_ provided, indicating that it will not be considered in the evaluation.
+- `requires_reference`: Boolean attribute specifying whether the evaluator requires a reference label. If True, the evaluator will raise an error when the reference isn't provided. If False, a warning will be logged if a reference _is_ provided, indicating that it will not be considered in the evaluation.
+
+String evaluators also implement the following methods:
+
+- `aevaluate_strings`: Asynchronously evaluates the output of the Chain or Language Model, with support for optional input and label.
+- `evaluate_strings`: Synchronously evaluates the output of the Chain or Language Model, with support for optional input and label.
+
+The following sections provide detailed information on available string evaluator implementations as well as how to create a custom string evaluator.
+
+import DocCardList from "@theme/DocCardList";
+
+<DocCardList />

docs/docs_skeleton/docs/guides/evaluation/trajectory/index.mdx

@@ -0,0 +1,28 @@
+---
+sidebar_position: 4
+---
+# Trajectory Evaluators
+
+Trajectory Evaluators in LangChain provide a more holistic approach to evaluating an agent. These evaluators assess the full sequence of actions taken by an agent and their corresponding responses, which we refer to as the "trajectory". This allows you to better measure an agent's effectiveness and capabilities.
+
+A Trajectory Evaluator implements the `AgentTrajectoryEvaluator` interface, which requires two main methods:
+
+- `evaluate_agent_trajectory`: This method synchronously evaluates an agent's trajectory.
+- `aevaluate_agent_trajectory`: This asynchronous counterpart allows evaluations to be run in parallel for efficiency.
+
+Both methods accept three main parameters:
+
+- `input`: The initial input given to the agent.
+- `prediction`: The final predicted response from the agent.
+- `agent_trajectory`: The intermediate steps taken by the agent, given as a list of tuples.
+
+These methods return a dictionary. It is recommended that custom implementations return a `score` (a float indicating the effectiveness of the agent) and `reasoning` (a string explaining the reasoning behind the score).
+
+You can capture an agent's trajectory by initializing the agent with the `return_intermediate_steps=True` parameter. This lets you collect all intermediate steps without relying on special callbacks.
+
+For a deeper dive into the implementation and use of Trajectory Evaluators, refer to the sections below.
+
+import DocCardList from "@theme/DocCardList";
+
+<DocCardList />
+

docs/docs_skeleton/docs/guides/expression_language/index.mdx

@@ -0,0 +1,9 @@
+# LangChain Expression Language
+
+import DocCardList from "@theme/DocCardList";
+
+LangChain Expression Language is a declarative way to easily compose chains together.
+Any chain constructed this way will automatically have full sync, async, and streaming support.
+See guides below for how to interact with chains constructed this way as well as cookbook examples.
+
+<DocCardList />

docs/docs_skeleton/docs/guides/langsmith/index.md

@@ -5,8 +5,8 @@ import DocCardList from "@theme/DocCardList";
 LangSmith helps you trace and evaluate your language model applications and intelligent agents to help you
 move from prototype to production.
 
-Check out the [interactive walkthrough](walkthrough) below to get started.
+Check out the [interactive walkthrough](/docs/guides/langsmith/walkthrough) below to get started.
 
 For more information, please refer to the [LangSmith documentation](https://docs.smith.langchain.com/)
 
-<DocCardList />
+<DocCardList />

docs/docs_skeleton/docs/guides/safety/index.mdx

@@ -0,0 +1,6 @@
+# Preventing harmful outputs
+
+One of the key concerns with using LLMs is that they may generate harmful or unethical text. This is an area of active research in the field. Here we present some built-in chains inspired by this research, which are intended to make the outputs of LLMs safer.
+
+- [Moderation chain](/docs/use_cases/safety/moderation): Explicitly check if any output text is harmful and flag it.
+- [Constitutional chain](/docs/use_cases/safety/constitutional_chain): Prompt the model with a set of principles which should guide it's behavior.

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.
 
@@ -28,7 +28,7 @@ navigating around a browser.
 ### [OpenAI Functions](/docs/modules/agents/agent_types/openai_functions_agent.html)
 
 Certain OpenAI models (like gpt-3.5-turbo-0613 and gpt-4-0613) have been explicitly fine-tuned to detect when a
-function should to be called and respond with the inputs that should be passed to the function.
+function should be called and respond with the inputs that should be passed to the function.
 The OpenAI Functions Agent is designed to work with these models.
 
 ### [Conversational](/docs/modules/agents/agent_types/chat_conversation_agent.html)

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

@@ -1,6 +1,6 @@
 # OpenAI functions
 
-Certain OpenAI models (like gpt-3.5-turbo-0613 and gpt-4-0613) have been fine-tuned to detect when a function should to be called and respond with the inputs that should be passed to the function.
+Certain OpenAI models (like gpt-3.5-turbo-0613 and gpt-4-0613) have been fine-tuned to detect when a function should be called and respond with the inputs that should be passed to the function.
 In an API call, you can describe functions and have the model intelligently choose to output a JSON object containing arguments to call those functions.
 The goal of the OpenAI Function APIs is to more reliably return valid and useful function calls than a generic text completion or chat API.
 

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

@@ -3,46 +3,80 @@ sidebar_position: 4
 ---
 # Agents
 
-Some applications require a flexible chain of calls to LLMs and other tools based on user input. The **Agent** interface provides the flexibility for such applications. An agent has access to a suite of tools, and determines which ones to use depending on the user input. Agents can use multiple tools, and use the output of one tool as the input to the next.
+The core idea of agents is to use an LLM to choose a sequence of actions to take.
+In chains, a sequence of actions is hardcoded (in code).
+In agents, a language model is used as a reasoning engine to determine which actions to take and in which order.
 
-There are two main types of agents:
+There are several key components here:
 
-- **Action agents**: at each timestep, decide on the next action using the outputs of all previous actions
-- **Plan-and-execute agents**: decide on the full sequence of actions up front, then execute them all without updating the plan
+## Agent
 
-Action agents are suitable for small tasks, while plan-and-execute agents are better for complex or long-running tasks that require maintaining long-term objectives and focus. Often the best approach is to combine the dynamism of an action agent with the planning abilities of a plan-and-execute agent by letting the plan-and-execute agent use action agents to execute plans.
+This is the class responsible for deciding what step to take next.
+This is powered by a language model and a prompt.
+This prompt can include things like:
 
-For a full list of agent types see [agent types](/docs/modules/agents/agent_types/). Additional abstractions involved in agents are:
-- [**Tools**](/docs/modules/agents/tools/): the actions an agent can take. What tools you give an agent highly depend on what you want the agent to do
-- [**Toolkits**](/docs/modules/agents/toolkits/): wrappers around collections of tools that can be used together a specific use case. For example, in order for an agent to
-  interact with a SQL database it will likely need one tool to execute queries and another to inspect tables
+1. The personality of the agent (useful for having it respond in a certain way)
+2. Background context for the agent (useful for giving it more context on the types of tasks it's being asked to do)
+3. Prompting strategies to invoke better reasoning (the most famous/widely used being [ReAct](https://arxiv.org/abs/2210.03629))
 
-## Action agents
+LangChain provides a few different types of agents to get started.
+Even then, you will likely want to customize those agents with parts (1) and (2).
+For a full list of agent types see [agent types](/docs/modules/agents/agent_types/)
 
-At a high-level an action agent:
-1. Receives user input
-2. Decides which tool, if any, to use and the tool input
-3. Calls the tool and records the output (also known as an "observation")
-4. Decides the next step using the history of tools, tool inputs, and observations
-5. Repeats 3-4 until it determines it can respond directly to the user
+## Tools
 
-Action agents are wrapped in **agent executors**, which are responsible for calling the agent, getting back an action and action input, calling the tool that the action references with the generated input, getting the output of the tool, and then passing all that information back into the agent to get the next action it should take.
+Tools are functions that an agent calls.
+There are two important considerations here:
 
-Although an agent can be constructed in many ways, it typically involves these components:
+1. Giving the agent access to the right tools
+2. Describing the tools in a way that is most helpful to the agent
 
-- **Prompt template**: Responsible for taking the user input and previous steps and constructing a prompt
-  to send to the language model
-- **Language model**: Takes the prompt with use input and action history and decides what to do next
-- **Output parser**: Takes the output of the language model and parses it into the next action or a final answer
+Without both, the agent you are trying to build will not work.
+If you don't give the agent access to a correct set of tools, it will never be able to accomplish the objective.
+If you don't describe the tools properly, the agent won't know how to properly use them.
 
-## Plan-and-execute agents
+LangChain provides a wide set of tools to get started, but also makes it easy to define your own (including custom descriptions).
+For a full list of tools, see [here](/docs/modules/agents/tools/)
 
-At a high-level a plan-and-execute agent:
-1. Receives user input
-2. Plans the full sequence of steps to take
-3. Executes the steps in order, passing the outputs of past steps as inputs to future steps
+## Toolkits
 
-The most typical implementation is to have the planner be a language model, and the executor be an action agent. Read more [here](/docs/modules/agents/agent_types/plan_and_execute.html).
+Often the set of tools an agent has access to is more important than a single tool.
+For this LangChain provides the concept of toolkits - groups of tools needed to accomplish specific objectives.
+There are generally around 3-5 tools in a toolkit.
+
+LangChain provides a wide set of toolkits to get started.
+For a full list of toolkits, see [here](/docs/modules/agents/toolkits/)
+
+## AgentExecutor
+
+The agent executor is the runtime for an agent.
+This is what actually calls the agent and executes the actions it chooses.
+Pseudocode for this runtime is below:
+
+```python
+next_action = agent.get_action(...)
+while next_action != AgentFinish:
+    observation = run(next_action)
+    next_action = agent.get_action(..., next_action, observation)
+return next_action
+```
+
+While this may seem simple, there are several complexities this runtime handles for you, including:
+
+1. Handling cases where the agent selects a non-existent tool
+2. Handling cases where the tool errors
+3. Handling cases where the agent produces output that cannot be parsed into a tool invocation
+4. Logging and observability at all levels (agent decisions, tool calls) either to stdout or [LangSmith](https://smith.langchain.com).
+
+## Other types of agent runtimes
+
+The `AgentExecutor` class is the main agent runtime supported by LangChain.
+However, there are other, more experimental runtimes we also support.
+These include:
+
+- [Plan-and-execute Agent](/docs/modules/agents/agent_types/plan_and_execute.html)
+- [Baby AGI](/docs/use_cases/autonomous_agents/baby_agi.html)
+- [Auto GPT](/docs/use_cases/autonomous_agents/autogpt.html)
 
 ## Get started
 

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

@@ -3,8 +3,8 @@ sidebar_position: 3
 ---
 # Toolkits
 
+:::info
+Head to [Integrations](/docs/integrations/toolkits/) for documentation on built-in toolkit integrations.
+:::
+
 Toolkits are collections of tools that are designed to be used together for specific tasks and have convenience loading methods.
-
-import DocCardList from "@theme/DocCardList";
-
-<DocCardList />

docs/docs_skeleton/docs/modules/agents/tools/how_to/_category_.yml

@@ -1,2 +0,0 @@
-label: 'How-to'
-position: 0

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

@@ -3,6 +3,10 @@ sidebar_position: 2
 ---
 # Tools
 
+:::info
+Head to [Integrations](/docs/integrations/tools/) for documentation on built-in tool integrations.
+:::
+
 Tools are interfaces that an agent can use to interact with the world.
 
 ## Get started

docs/docs_skeleton/docs/modules/agents/tools/integrations/_category_.yml

@@ -1 +0,0 @@
-label: 'Integrations'

docs/docs_skeleton/docs/modules/callbacks/how_to/_category_.yml

@@ -1,2 +0,0 @@
-label: 'How-to'
-position: 0

docs/docs_skeleton/docs/modules/callbacks/index.mdx

@@ -3,6 +3,10 @@ sidebar_position: 5
 ---
 # Callbacks
 
+:::info
+Head to [Integrations](/docs/integrations/callbacks/) for documentation on built-in callbacks integrations with 3rd-party tools.
+:::
+
 LangChain provides a callbacks system that allows you to hook into the various stages of your LLM application. This is useful for logging, monitoring, streaming, and other tasks.
 
 import GetStarted from "@snippets/modules/callbacks/get_started.mdx"

docs/docs_skeleton/docs/modules/callbacks/integrations/_category_.yml

@@ -1 +0,0 @@
-label: 'Integrations'

docs/docs_skeleton/docs/modules/chains/additional/multi_prompt_router.mdx

@@ -1,7 +0,0 @@
-# Dynamically selecting from multiple prompts
-
-This notebook demonstrates how to use the `RouterChain` paradigm to create a chain that dynamically selects the prompt to use for a given input. Specifically we show how to use the `MultiPromptChain` to create a question-answering chain that selects the prompt which is most relevant for a given question, and then answers the question using that prompt.
-
-import Example from "@snippets/modules/chains/additional/multi_prompt_router.mdx"
-
-<Example/>

docs/docs_skeleton/docs/modules/chains/foundational/sequential_chains.mdx

@@ -1,6 +1,6 @@
 # Sequential
 
-<!-- WARNING: THIS FILE WAS AUTOGENERATED! DO NOT EDIT! Instead, edit the notebook w/the location & name as this file. -->
+
 
 The next step after calling a language model is make a series of calls to a language model. This is particularly useful when you want to take the output from one call and use it as the input to another.
 

docs/docs_skeleton/docs/modules/chains/popular/api.mdx

@@ -1,9 +0,0 @@
----
-sidebar_position: 0
----
-# API chains
-APIChain enables using LLMs to interact with APIs to retrieve relevant information. Construct the chain by providing a question relevant to the provided API documentation.
-
-import Example from "@snippets/modules/chains/popular/api.mdx"
-
-<Example/>

docs/docs_skeleton/docs/modules/chains/popular/summarize.mdx

@@ -1,8 +0,0 @@
-# Summarization
-
-A summarization chain can be used to summarize multiple documents. One way is to input multiple smaller documents, after they have been divided into chunks, and operate over them with a MapReduceDocumentsChain. You can also choose instead for the chain that does summarization to be a StuffDocumentsChain, or a RefineDocumentsChain.
-
-import Example from "@snippets/modules/chains/popular/summarize.mdx"
-
-<Example/>
-

docs/docs_skeleton/docs/modules/data_connection/document_loaders/how_to/_category_.yml

@@ -1,2 +0,0 @@
-label: 'How-to'
-position: 0

docs/docs_skeleton/docs/modules/data_connection/document_loaders/index.mdx

@@ -3,6 +3,10 @@ sidebar_position: 0
 ---
 # Document loaders
 
+:::info
+Head to [Integrations](/docs/integrations/document_loaders/) for documentation on built-in document loader integrations with 3rd-party tools.
+:::
+
 Use document loaders to load data from a source as `Document`'s. A `Document` is a piece of text
 and associated metadata. For example, there are document loaders for loading a simple `.txt` file, for loading the text
 contents of any web page, or even for loading a transcript of a YouTube video.

docs/docs_skeleton/docs/modules/data_connection/document_loaders/integrations/_category_.yml

@@ -1 +0,0 @@
-label: 'Integrations'

docs/docs_skeleton/docs/modules/data_connection/document_transformers/index.mdx

@@ -3,6 +3,10 @@ sidebar_position: 1
 ---
 # Document transformers
 
+:::info
+Head to [Integrations](/docs/integrations/document_transformers/) for documentation on built-in document transformer integrations with 3rd-party tools.
+:::
+
 Once you've loaded documents, you'll often want to transform them to better suit your application. The simplest example
 is you may want to split a long document into smaller chunks that can fit into your model's context window. LangChain
 has a number of built-in document transformers that make it easy to split, combine, filter, and otherwise manipulate documents.

docs/docs_skeleton/docs/modules/data_connection/index.mdx

@@ -2,15 +2,60 @@
 sidebar_position: 1
 ---
 
-# Data connection
+# Retrieval
 
-Many LLM applications require user-specific data that is not part of the model's training set. LangChain gives you the 
-building blocks to load, transform, store and query your data via:
+Many LLM applications require user-specific data that is not part of the model's training set.
+The primary way of accomplishing this is through Retrieval Augmented Generation (RAG).
+In this process, external data is *retrieved* and then passed to the LLM when doing the *generation* step.
 
-- [Document loaders](/docs/modules/data_connection/document_loaders/): Load documents from many different sources
-- [Document transformers](/docs/modules/data_connection/document_transformers/): Split documents, convert documents into Q&A format, drop redundant documents, and more
-- [Text embedding models](/docs/modules/data_connection/text_embedding/): Take unstructured text and turn it into a list of floating point numbers
-- [Vector stores](/docs/modules/data_connection/vectorstores/): Store and search over embedded data
-- [Retrievers](/docs/modules/data_connection/retrievers/): Query your data
+LangChain provides all the building blocks for RAG applications - from simple to complex.
+This section of the documentation covers everything related to the *retrieval* step - e.g. the fetching of the data.
+Although this sounds simple, it can be subtly complex.
+This encompasses several key modules.
 
 ![data_connection_diagram](/img/data_connection.jpg)
+
+**[Document loaders](/docs/modules/data_connection/document_loaders/)**
+
+Load documents from many different sources.
+LangChain provides over a 100 different document loaders as well as integrations with other major providers in the space,
+like AirByte and Unstructured.
+We provide integrations to load all types of documents (html, PDF, code) from all types of locations (private s3 buckets, public websites).
+
+**[Document transformers](/docs/modules/data_connection/document_transformers/)**
+
+A key part of retrieval is fetching only the relevant parts of documents.
+This involves several transformation steps in order to best prepare the documents for retrieval.
+One of the primary ones here is splitting (or chunking) a large document into smaller chunks.
+LangChain provides several different algorithms for doing this, as well as logic optimized for specific document types (code, markdown, etc).
+
+**[Text embedding models](/docs/modules/data_connection/text_embedding/)**
+
+Another key part of retrieval has become creating embeddings for documents.
+Embeddings capture the semantic meaning of text, allowing you to quickly and
+efficiently find other pieces of text that are similar.
+LangChain provides integrations with over 25 different embedding providers and methods,
+from open-source to proprietary API,
+allowing you to choose the one best suited for your needs.
+LangChain exposes a standard interface, allowing you to easily swap between models.
+
+**[Vector stores](/docs/modules/data_connection/vectorstores/)**
+
+With the rise of embeddings, there has emerged a need for databases to support efficient storage and searching of these embeddings.
+LangChain provides integrations with over 50 different vectorstores, from open-source local ones to cloud-hosted proprietary ones,
+allowing you choose the one best suited for your needs.
+LangChain exposes a standard interface, allowing you to easily swap between vector stores.
+
+**[Retrievers](/docs/modules/data_connection/retrievers/)**
+
+Once the data is in the database, you still need to retrieve it.
+LangChain supports many different retrieval algorithms and is one of the places where we add the most value.
+We support basic methods that are easy to get started - namely simple semantic search.
+However, we have also added a collection of algorithms on top of this to increase performance.
+These include:
+
+- [Parent Document Retriever](/docs/modules/data_connection/retrievers/parent_document_retriever): This allows you to create multiple embeddings per parent document, allowing you to look up smaller chunks but return larger context.
+- [Self Query Retriever](/docs/modules/data_connection/retrievers/self_query): User questions often contain reference to something that isn't just semantic, but rather expresses some logic that can best be represented as a metadata filter. Self-query allows you to parse out the *semantic* part of a query from other *metadata filters* present in the query
+- [Ensemble Retriever](/docs/modules/data_connection/retrievers/ensemble): Sometimes you may want to retrieve documents from multiple different sources, or using multiple different algorithms. The ensemble retriever allows you to easily do this.
+- And more!
+

docs/docs_skeleton/docs/modules/data_connection/retrievers/how_to/_category_.yml

@@ -1,2 +0,0 @@
-label: 'How-to'
-position: 0

docs/docs_skeleton/docs/modules/data_connection/retrievers/index.mdx

@@ -3,6 +3,10 @@ sidebar_position: 4
 ---
 # Retrievers
 
+:::info
+Head to [Integrations](/docs/integrations/retrievers/) for documentation on built-in retriever integrations with 3rd-party tools.
+:::
+
 A retriever is an interface that returns documents given an unstructured query. It is more general than a vector store.
 A retriever does not need to be able to store documents, only to return (or retrieve) it. Vector stores can be used
 as the backbone of a retriever, but there are other types of retrievers as well.

docs/docs_skeleton/docs/modules/data_connection/retrievers/integrations/_category_.yml

@@ -1 +0,0 @@
-label: 'Integrations'

docs/docs_skeleton/docs/modules/data_connection/text_embedding/index.mdx

@@ -3,6 +3,10 @@ sidebar_position: 2
 ---
 # Text embedding models
 
+:::info
+Head to [Integrations](/docs/integrations/text_embedding/) for documentation on built-in integrations with text embedding model providers.
+:::
+
 The Embeddings class is a class designed for interfacing with text embedding models. There are lots of embedding model providers (OpenAI, Cohere, Hugging Face, etc) - this class is designed to provide a standard interface for all of them.
 
 Embeddings create a vector representation of a piece of text. This is useful because it means we can think about text in the vector space, and do things like semantic search where we look for pieces of text that are most similar in the vector space.

docs/docs_skeleton/docs/modules/data_connection/text_embedding/integrations/_category_.yml

@@ -1 +0,0 @@
-label: 'Integrations'

docs/docs_skeleton/docs/modules/data_connection/vectorstores/index.mdx

@@ -3,6 +3,10 @@ sidebar_position: 3
 ---
 # Vector stores
 
+:::info
+Head to [Integrations](/docs/integrations/vectorstores/) for documentation on built-in integrations with 3rd-party vector stores.
+:::
+
 One of the most common ways to store and search over unstructured data is to embed it and store the resulting embedding
 vectors, and then at query time to embed the unstructured query and retrieve the embedding vectors that are
 'most similar' to the embedded query. A vector store takes care of storing embedded data and performing vector search

docs/docs_skeleton/docs/modules/data_connection/vectorstores/integrations/_category_.yml

@@ -1 +0,0 @@
-label: 'Integrations'

docs/docs_skeleton/docs/modules/evaluation/comparison/index.mdx

@@ -1,8 +0,0 @@
----
-sidebar_position: 3 
----
-# Comparison Evaluators
-
-import DocCardList from "@theme/DocCardList";
-
-<DocCardList />

docs/docs_skeleton/docs/modules/evaluation/index.mdx

@@ -1,28 +0,0 @@
----
-sidebar_position: 6
----
-
-import DocCardList from "@theme/DocCardList";
-
-# Evaluation
-
-Language models can be unpredictable. This makes it challenging to ship reliable applications to production, where repeatable, useful outcomes across diverse inputs are a minimum requirement. Tests help demonstrate each component in an LLM application can produce the required or expected functionality. These tests also safeguard against regressions while you improve interconnected pieces of an integrated system. However, measuring the quality of generated text can be challenging. It can be hard to agree on the right set of metrics for your application, and it can be difficult to translate those into better performance. Furthermore, it's common to lack sufficient evaluation data adequately test the range of inputs and expected outputs for each component when you're just getting started. The LangChain community is building open source tools and guides to help address these challenges.
-
-LangChain exposes different types of evaluators for common types of evaluation. Each type has off-the-shelf implementations you can use to get started, as well as an
- extensible API so you can create your own or contribute improvements for everyone to use. The following sections have example notebooks for you to get started.
-
-- [String Evaluators](/docs/modules/evaluation/string/): Evaluate the predicted string for a given input, usually against a reference string
-- [Trajectory Evaluators](/docs/modules/evaluation/trajectory/): Evaluate the whole trajectory of agent actions
-- [Comparison Evaluators](/docs/modules/evaluation/comparison/): Compare predictions from two runs on a common input
-
-
-This section also provides some additional examples of how you could use these evaluators for different scenarios or apply to different chain implementations in the LangChain library. Some examples include:
-
-- [Preference Scoring Chain Outputs](/docs/modules/evaluation/examples/comparisons): An example using a comparison evaluator on different models or prompts to select statistically significant differences in aggregate preference scores
-
-
-## Reference Docs
-
-For detailed information of the available evaluators, including how to instantiate, configure, and customize them. Check out the [reference documentation](https://api.python.langchain.com/en/latest/api_reference.html#module-langchain.evaluation) directly.
-
-<DocCardList />

docs/docs_skeleton/docs/modules/evaluation/trajectory/index.mdx

@@ -1,8 +0,0 @@
----
-sidebar_position: 4
----
-# Trajectory Evaluators
-
-import DocCardList from "@theme/DocCardList";
-
-<DocCardList />

docs/docs_skeleton/docs/modules/index.mdx

@@ -8,7 +8,7 @@ LangChain provides standard, extendable interfaces and external integrations for
 
 #### [Model I/O](/docs/modules/model_io/)
 Interface with language models
-#### [Data connection](/docs/modules/data_connection/)
+#### [Retrieval](/docs/modules/data_connection/)
 Interface with application-specific data
 #### [Chains](/docs/modules/chains/)
 Construct sequences of calls
@@ -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/memory/chat_messages/index.mdx

@@ -0,0 +1,17 @@
+---
+sidebar_position: 1
+---
+# Chat Messages
+
+:::info
+Head to [Integrations](/docs/integrations/memory/) for documentation on built-in memory integrations with 3rd-party databases and tools.
+:::
+
+One of the core utility classes underpinning most (if not all) memory modules is the `ChatMessageHistory` class.
+This is a super lightweight wrapper which exposes convenience methods for saving Human messages, AI messages, and then fetching them all.
+
+You may want to use this class directly if you are managing memory outside of a chain.
+
+import GetStarted from "@snippets/modules/memory/chat_messages/get_started.mdx"
+
+<GetStarted/>

docs/docs_skeleton/docs/modules/memory/how_to/_category_.yml

@@ -1,2 +0,0 @@
-label: 'How-to'
-position: 0

docs/docs_skeleton/docs/modules/memory/index.mdx

@@ -1,30 +1,62 @@
 ---
 sidebar_position: 3
 ---
-
 # Memory
 
-🚧 _Docs under construction_ 🚧
+Most LLM applications have a conversational interface. An essential component of a conversation is being able to refer to information introduced earlier in the conversation.
+At bare minimum, a conversational system should be able to access some window of past messages directly.
+A more complex system will need to have a world model that it is constantly updating, which allows it to do things like maintain information about entities and their relationships.
 
-By default, Chains and Agents are stateless,
-meaning that they treat each incoming query independently (like the underlying LLMs and chat models themselves).
-In some applications, like chatbots, it is essential
-to remember previous interactions, both in the short and long-term.
-The **Memory** class does exactly that.
+We call this ability to store information about past interactions "memory".
+LangChain provides a lot of utilities for adding memory to a system.
+These utilities can be used by themselves or incorporated seamlessly into a chain.
 
-LangChain provides memory components in two forms.
-First, LangChain provides helper utilities for managing and manipulating previous chat messages.
-These are designed to be modular and useful regardless of how they are used.
-Secondly, LangChain provides easy ways to incorporate these utilities into chains.
+A memory system needs to support two basic actions: reading and writing.
+Recall that every chain defines some core execution logic that expects certain inputs.
+Some of these inputs come directly from the user, but some of these inputs can come from memory.
+A chain will interact with its memory system twice in a given run.
+1. AFTER receiving the initial user inputs but BEFORE executing the core logic, a chain will READ from its memory system and augment the user inputs.
+2. AFTER executing the core logic but BEFORE returning the answer, a chain will WRITE the inputs and outputs of the current run to memory, so that they can be referred to in future runs.
+
+![memory-diagram](/img/memory_diagram.png)
+
+
+## Building memory into a system
+The two core design decisions in any memory system are:
+- How state is stored
+- How state is queried
+
+### Storing: List of chat messages
+Underlying any memory is a history of all chat interactions.
+Even if these are not all used directly, they need to be stored in some form.
+One of the key parts of the LangChain memory module is a series of integrations for storing these chat messages,
+from in-memory lists to persistent databases.
+
+- [Chat message storage](/docs/modules/memory/chat_messages/): How to work with Chat Messages, and the various integrations offered
+
+### Querying: Data structures and algorithms on top of chat messages
+Keeping a list of chat messages is fairly straight-forward.
+What is less straight-forward are the data structures and algorithms built on top of chat messages that serve a view of those messages that is most useful.
+
+A very simply memory system might just return the most recent messages each run. A slightly more complex memory system might return a succinct summary of the past K messages.
+An even more sophisticated system might extract entities from stored messages and only return information about entities referenced in the current run.
+
+Each application can have different requirements for how memory is queried. The memory module should make it easy to both get started with simple memory systems and write your own custom systems if needed.
+
+- [Memory types](/docs/modules/memory/types/): The various data structures and algorithms that make up the memory types LangChain supports
 
 ## Get started
 
-Memory involves keeping a concept of state around throughout a user's interactions with an language model. A user's interactions with a language model are captured in the concept of ChatMessages, so this boils down to ingesting, capturing, transforming and extracting knowledge from a sequence of chat messages. There are many different ways to do this, each of which exists as its own memory type.
-
-In general, for each type of memory there are two ways to understanding using memory. These are the standalone functions which extract information from a sequence of messages, and then there is the way you can use this type of memory in a chain.
-
-Memory can return multiple pieces of information (for example, the most recent N messages and a summary of all previous messages). The returned information can either be a string or a list of messages.
+Let's take a look at what Memory actually looks like in LangChain.
+Here we'll cover the basics of interacting with an arbitrary memory class.
 
 import GetStarted from "@snippets/modules/memory/get_started.mdx"
 
 <GetStarted/>
+
+## Next steps
+
+And that's it for getting started!
+Please see the other sections for walkthroughs of more advanced topics,
+like custom memory, multiple memories, and more.
+

docs/docs_skeleton/docs/modules/memory/integrations/_category_.yml

@@ -1 +0,0 @@
-label: 'Integrations'

docs/docs_skeleton/docs/modules/memory/types/buffer.mdx

@@ -4,6 +4,6 @@ This notebook shows how to use `ConversationBufferMemory`. This memory allows fo
 
 We can first extract it as a string.
 
-import Example from "@snippets/modules/memory/how_to/buffer.mdx"
+import Example from "@snippets/modules/memory/types/buffer.mdx"
 
 <Example/>

docs/docs_skeleton/docs/modules/memory/types/buffer_window.mdx

@@ -4,6 +4,6 @@
 
 Let's first explore the basic functionality of this type of memory.
 
-import Example from "@snippets/modules/memory/how_to/buffer_window.mdx"
+import Example from "@snippets/modules/memory/types/buffer_window.mdx"
 
 <Example/>

docs/docs_skeleton/docs/modules/memory/types/entity_summary_memory.mdx

@@ -4,6 +4,6 @@ Entity Memory remembers given facts about specific entities in a conversation. I
 
 Let's first walk through using this functionality.
 
-import Example from "@snippets/modules/memory/how_to/entity_summary_memory.mdx"
+import Example from "@snippets/modules/memory/types/entity_summary_memory.mdx"
 
 <Example/>

docs/docs_skeleton/docs/modules/memory/types/index.mdx

@@ -0,0 +1,8 @@
+---
+sidebar_position: 2
+---
+# Memory Types
+
+There are many different types of memory.
+Each have their own parameters, their own return types, and are useful in different scenarios.
+Please see their individual page for more detail on each one.

docs/docs_skeleton/docs/modules/memory/types/summary.mdx

@@ -4,6 +4,6 @@ Conversation summary memory summarizes the conversation as it happens and stores
 
 Let's first explore the basic functionality of this type of memory.
 
-import Example from "@snippets/modules/memory/how_to/summary.mdx"
+import Example from "@snippets/modules/memory/types/summary.mdx"
 
 <Example/>

docs/docs_skeleton/docs/modules/memory/types/vectorstore_retriever_memory.mdx

@@ -6,6 +6,6 @@ This differs from most of the other Memory classes in that it doesn't explicitly
 
 In this case, the "docs" are previous conversation snippets. This can be useful to refer to relevant pieces of information that the AI was told earlier in the conversation.
 
-import Example from "@snippets/modules/memory/how_to/vectorstore_retriever_memory.mdx"
+import Example from "@snippets/modules/memory/types/vectorstore_retriever_memory.mdx"
 
 <Example/>

docs/docs_skeleton/docs/modules/model_io/models/chat/how_to/_category_.yml

@@ -1,2 +0,0 @@
-label: 'How-to'
-position: 0