community[patch]: Fix more issues (#26116)

This PR resolves more type checking issues and fixes some bugs.

libs/community/langchain_community/chat_models/oci_generative_ai.py

@@ -548,6 +548,9 @@ class ChatOCIGenAI(BaseChatModel, OCIGenAIBase):
 
         chat_params = {**_model_kwargs, **kwargs, **oci_params}
 
+        if not self.model_id:
+            raise ValueError("Model ID is required to chat")
+
         if self.model_id.startswith(CUSTOM_ENDPOINT_PREFIX):
             serving_mode = models.DedicatedServingMode(endpoint_id=self.model_id)
         else:

libs/community/langchain_community/document_compressors/llmlingua_filter.py

@@ -8,7 +8,7 @@ from langchain_core.documents import Document
 from langchain_core.documents.compressor import (
     BaseDocumentCompressor,
 )
-from pydantic import ConfigDict, model_validator
+from pydantic import ConfigDict, Field, model_validator
 
 DEFAULT_LLM_LINGUA_INSTRUCTION = (
     "Given this documents, please answer the final question"
@@ -35,9 +35,9 @@ class LLMLinguaCompressor(BaseDocumentCompressor):
     """The target number of compressed tokens"""
     rank_method: str = "longllmlingua"
     """The ranking method to use"""
-    model_config: dict = {}
+    model_configuration: dict = Field(default_factory=dict, alias="model_config")
     """Custom configuration for the model"""
-    open_api_config: dict = {}
+    open_api_config: dict = Field(default_factory=dict)
     """open_api configuration"""
     instruction: str = DEFAULT_LLM_LINGUA_INSTRUCTION
     """The instruction for the LLM"""
@@ -75,6 +75,7 @@ class LLMLinguaCompressor(BaseDocumentCompressor):
     model_config = ConfigDict(
         arbitrary_types_allowed=True,
         extra="forbid",
+        populate_by_name=True,
     )
 
     @staticmethod

libs/community/langchain_community/document_loaders/onenote.py

@@ -1,7 +1,7 @@
 """Loads data from OneNote Notebooks"""
 
 from pathlib import Path
-from typing import Dict, Iterator, List, Optional
+from typing import Any, Dict, Iterator, List, Optional
 
 import requests
 from langchain_core.documents import Document
@@ -10,6 +10,7 @@ from pydantic import (
     Field,
     FilePath,
     SecretStr,
+    model_validator,
 )
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
@@ -17,13 +18,14 @@ from langchain_community.document_loaders.base import BaseLoader
 
 
 class _OneNoteGraphSettings(BaseSettings):
-    client_id: str = Field(..., alias="MS_GRAPH_CLIENT_ID")
+    client_id: str = Field(...)
-    client_secret: SecretStr = Field(..., alias="MS_GRAPH_CLIENT_SECRET")
+    client_secret: SecretStr = Field(...)
 
     model_config = SettingsConfigDict(
         case_sensitive=False,
+        populate_by_name=True,
         env_file=".env",
-        env_prefix="",
+        env_prefix="MS_GRAPH_",
     )
 
 
@@ -51,6 +53,14 @@ class OneNoteLoader(BaseLoader, BaseModel):
     object_ids: Optional[List[str]] = None
     """ The IDs of the objects to load data from."""
 
+    @model_validator(mode="before")
+    @classmethod
+    def init(cls, values: Dict) -> Any:
+        """Initialize the class."""
+        if "settings" in values and isinstance(values["settings"], dict):
+            values["settings"] = _OneNoteGraphSettings(**values["settings"])
+        return values
+
     def lazy_load(self) -> Iterator[Document]:
         """
         Get pages from OneNote notebooks.

libs/community/langchain_community/embeddings/oci_generative_ai.py

@@ -181,6 +181,9 @@ class OCIGenAIEmbeddings(BaseModel, Embeddings):
         """
         from oci.generative_ai_inference import models
 
+        if not self.model_id:
+            raise ValueError("Model ID is required to embed documents")
+
         if self.model_id.startswith(CUSTOM_ENDPOINT_PREFIX):
             serving_mode = models.DedicatedServingMode(endpoint_id=self.model_id)
         else:

libs/community/langchain_community/llms/exllamav2.py

@@ -1,4 +1,4 @@
-from typing import Any, Dict, Iterator, List, Optional
+from typing import Any, Callable, Dict, Iterator, List, Optional
 
 from langchain_core.callbacks import CallbackManagerForLLMRun
 from langchain_core.language_models import LLM
@@ -41,7 +41,7 @@ class ExLlamaV2(LLM):
     settings: Any = None
 
     # Langchain parameters
-    logfunc: callable = print
+    logfunc: Callable = print
 
     stop_sequences: List[str] = Field("")
     """Sequences that immediately will stop the generator."""

libs/community/langchain_community/tools/bearly/tool.py

@@ -37,7 +37,7 @@ class BearlyInterpreterToolArguments(BaseModel):
 
     python_code: str = Field(
         ...,
-        example="print('Hello World')",
+        examples=["print('Hello World')"],
         description=(
             "The pure python script to be evaluated. "
             "The contents will be in main.py. "

libs/community/langchain_community/tools/connery/models.py

@@ -1,4 +1,4 @@
-from typing import List, Optional
+from typing import Any, List, Optional
 
 from pydantic import BaseModel
 
@@ -15,7 +15,7 @@ class Parameter(BaseModel):
     key: str
     title: str
     description: Optional[str] = None
-    type: str
+    type: Any
     validation: Optional[Validation] = None
 
 

libs/community/langchain_community/tools/connery/tool.py

@@ -154,8 +154,8 @@ class ConneryAction(BaseTool):
             type = param.type
 
             dynamic_input_fields[param.key] = (
-                str,
+                type,
-                Field(default, title=title, description=description, type=type),
+                Field(default, title=title, description=description),
             )
 
         InputModel = create_model("InputSchema", **dynamic_input_fields)

libs/community/langchain_community/tools/e2b_data_analysis/tool.py

@@ -84,7 +84,7 @@ class E2BDataAnalysisToolArguments(BaseModel):
 
     python_code: str = Field(
         ...,
-        example="print('Hello World')",
+        examples=["print('Hello World')"],
         description=(
             "The python script to be evaluated. "
             "The contents will be in main.py. "

libs/community/tests/unit_tests/chat_models/test_oci_generative_ai.py

@@ -23,7 +23,11 @@ def test_llm_chat(monkeypatch: MonkeyPatch, test_model_id: str) -> None:
     oci_gen_ai_client = MagicMock()
     llm = ChatOCIGenAI(model_id=test_model_id, client=oci_gen_ai_client)
 
-    provider = llm.model_id.split(".")[0].lower()
+    model_id = llm.model_id
+    if model_id is None:
+        raise ValueError("Model ID is required for OCI Generative AI LLM service.")
+
+    provider = model_id.split(".")[0].lower()
 
     def mocked_response(*args):  # type: ignore[no-untyped-def]
         response_text = "Assistant chat reply."

libs/community/tests/unit_tests/document_loaders/test_onenote.py

@@ -15,7 +15,6 @@ def test_initialization() -> None:
     os.environ["MS_GRAPH_CLIENT_SECRET"] = "CLIENT_SECRET"
 
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         notebook_name="test_notebook",
         section_name="test_section",
         page_title="test_title",
@@ -40,7 +39,6 @@ def test_load(mocker: MockerFixture) -> None:
         return_value=mocker.MagicMock(json=lambda: {"value": []}, links=None),
     )
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         notebook_name="test_notebook",
         section_name="test_section",
         page_title="test_title",
@@ -57,7 +55,6 @@ def test_load(mocker: MockerFixture) -> None:
         ),
     )
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         object_ids=["test_id"],
         access_token="access_token",
     )
@@ -85,7 +82,6 @@ def test_msal_import(monkeypatch: MonkeyPatch, mocker: MockerFixture) -> None:
         return_value=FakeConfidentialClientApplication(),
     )
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         notebook_name="test_notebook",
         section_name="test_section",
         page_title="test_title",
@@ -99,7 +95,6 @@ def test_url() -> None:
     os.environ["MS_GRAPH_CLIENT_SECRET"] = "CLIENT_SECRET"
 
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         notebook_name="test_notebook",
         section_name="test_section",
         page_title="test_title",
@@ -115,7 +110,6 @@ def test_url() -> None:
     )
 
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         notebook_name="test_notebook",
         section_name="test_section",
         access_token="access_token",
@@ -129,7 +123,6 @@ def test_url() -> None:
     )
 
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         notebook_name="test_notebook",
         access_token="access_token",
         onenote_api_base_url="https://graph.microsoft.com/v1.0/me/onenote",
@@ -141,7 +134,6 @@ def test_url() -> None:
     )
 
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         section_name="test_section",
         access_token="access_token",
         onenote_api_base_url="https://graph.microsoft.com/v1.0/me/onenote",
@@ -153,7 +145,6 @@ def test_url() -> None:
     )
 
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         section_name="test_section",
         page_title="test_title",
         access_token="access_token",
@@ -167,7 +158,6 @@ def test_url() -> None:
     )
 
     loader = OneNoteLoader(
-        settings={"client_id": "", "client_secret": ""},
         page_title="test_title",
         access_token="access_token",
         onenote_api_base_url="https://graph.microsoft.com/v1.0/me/onenote",

libs/community/tests/unit_tests/llms/test_oci_generative_ai.py

@@ -22,7 +22,12 @@ def test_llm_complete(monkeypatch: MonkeyPatch, test_model_id: str) -> None:
     oci_gen_ai_client = MagicMock()
     llm = OCIGenAI(model_id=test_model_id, client=oci_gen_ai_client)
 
-    provider = llm.model_id.split(".")[0].lower()
+    model_id = llm.model_id
+
+    if model_id is None:
+        raise ValueError("Model ID is required for OCI Generative AI LLM service.")
+
+    provider = model_id.split(".")[0].lower()
 
     def mocked_response(*args):  # type: ignore[no-untyped-def]
         response_text = "This is the completion."