community: Add ruff rule PGH003 (#30812)

See https://docs.astral.sh/ruff/rules/blanket-type-ignore/

---------

Co-authored-by: Chester Curme <[email protected]>

Author: cbornet

libs/community/langchain_community/agent_toolkits/sql/base.py

@@ -196,7 +196,7 @@ def create_sql_agent(
             ]
             prompt = ChatPromptTemplate.from_messages(messages)
         agent = RunnableAgent(
-            runnable=create_openai_functions_agent(llm, tools, prompt),  # type: ignore
+            runnable=create_openai_functions_agent(llm, tools, prompt),  # type: ignore[arg-type]
             input_keys_arg=["input"],
             return_keys_arg=["output"],
             **kwargs,
@@ -211,9 +211,9 @@ def create_sql_agent(
             ]
             prompt = ChatPromptTemplate.from_messages(messages)
         if agent_type == "openai-tools":
-            runnable = create_openai_tools_agent(llm, tools, prompt)  # type: ignore
+            runnable = create_openai_tools_agent(llm, tools, prompt)  # type: ignore[arg-type]
         else:
-            runnable = create_tool_calling_agent(llm, tools, prompt)  # type: ignore
+            runnable = create_tool_calling_agent(llm, tools, prompt)  # type: ignore[arg-type]
         agent = RunnableMultiActionAgent(  # type: ignore[assignment]
             runnable=runnable,
             input_keys_arg=["input"],

libs/community/langchain_community/agents/openai_assistant/base.py

@@ -135,7 +135,7 @@ def _get_assistants_tool(
         Dict[str, Any]: A dictionary of tools that are converted into OpenAI tools.
     """
     if _is_assistants_builtin_tool(tool):
-        return tool  # type: ignore
+        return tool  # type: ignore[return-value]
     else:
         return convert_to_openai_tool(tool)
 
@@ -288,7 +288,7 @@ def create_assistant(
         assistant = client.beta.assistants.create(
             name=name,
             instructions=instructions,
-            tools=[_get_assistants_tool(tool) for tool in tools],  # type: ignore
+            tools=[_get_assistants_tool(tool) for tool in tools],
             tool_resources=tool_resources,  # type: ignore[arg-type]
             model=model,
             extra_body=extra_body,
@@ -430,7 +430,7 @@ async def acreate_assistant(
         assistant = await async_client.beta.assistants.create(
             name=name,
             instructions=instructions,
-            tools=openai_tools,  # type: ignore
+            tools=openai_tools,
             tool_resources=tool_resources,  # type: ignore[arg-type]
             model=model,
         )

libs/community/langchain_community/cache.py

@@ -238,7 +238,7 @@ async def aclear(self, **kwargs: Any) -> None:
 Base = declarative_base()
 
 
-class FullLLMCache(Base):  # type: ignore
+class FullLLMCache(Base):  # type: ignore[misc,valid-type]
     """SQLite table for full LLM Cache (all generations)."""
 
     __tablename__ = "full_llm_cache"
@@ -261,7 +261,7 @@ def lookup(self, prompt: str, llm_string: str) -> Optional[RETURN_VAL_TYPE]:
         """Look up based on prompt and llm_string."""
         stmt = (
             select(self.cache_schema.response)
-            .where(self.cache_schema.prompt == prompt)  # type: ignore
+            .where(self.cache_schema.prompt == prompt)
             .where(self.cache_schema.llm == llm_string)
             .order_by(self.cache_schema.idx)
         )
@@ -1531,7 +1531,7 @@ async def aclear(self, **kwargs: Any) -> None:
         await self.table.aclear()
 
 
-class FullMd5LLMCache(Base):  # type: ignore
+class FullMd5LLMCache(Base):  # type: ignore[misc,valid-type]
     """SQLite table for full LLM Cache (all generations)."""
 
     __tablename__ = "full_md5_llm_cache"
@@ -1583,7 +1583,7 @@ def update(self, prompt: str, llm_string: str, return_val: RETURN_VAL_TYPE) -> N
     def _delete_previous(self, session: Session, prompt: str, llm_string: str) -> None:
         stmt = (
             delete(self.cache_schema)
-            .where(self.cache_schema.prompt_md5 == self.get_md5(prompt))  # type: ignore
+            .where(self.cache_schema.prompt_md5 == self.get_md5(prompt))
             .where(self.cache_schema.llm == llm_string)
             .where(self.cache_schema.prompt == prompt)
         )
@@ -1593,7 +1593,7 @@ def _search_rows(self, prompt: str, llm_string: str) -> Sequence[Row]:
         prompt_pd5 = self.get_md5(prompt)
         stmt = (
             select(self.cache_schema.response)
-            .where(self.cache_schema.prompt_md5 == prompt_pd5)  # type: ignore
+            .where(self.cache_schema.prompt_md5 == prompt_pd5)
             .where(self.cache_schema.llm == llm_string)
             .where(self.cache_schema.prompt == prompt)
             .order_by(self.cache_schema.idx)
@@ -1796,7 +1796,7 @@ def __init__(self, awaitable: Awaitable[Any]):
     def __await__(self) -> Generator:
         if self.result is _unset:
             self.result = yield from self.awaitable.__await__()
-        return self.result  # type: ignore
+        return self.result  # type: ignore[return-value]
 
 
 def _reawaitable(func: Callable) -> Callable:

libs/community/langchain_community/callbacks/comet_ml_callback.py

@@ -584,7 +584,7 @@ def _reset(
         )
         _custom_metrics = custom_metrics if custom_metrics else self.custom_metrics
 
-        self.__init__(  # type: ignore
+        self.__init__(  # type: ignore[misc]
             task_type=_task_type,
             workspace=_workspace,
             project_name=_project_name,

libs/community/langchain_community/callbacks/wandb_callback.py

@@ -580,7 +580,7 @@ def flush_tracker(
             self.temp_dir.cleanup()
             self.reset_callback_meta()
         if reset:
-            self.__init__(  # type: ignore
+            self.__init__(  # type: ignore[misc]
                 job_type=job_type if job_type else self.job_type,
                 project=project if project else self.project,
                 entity=entity if entity else self.entity,

libs/community/langchain_community/chains/ernie_functions/base.py

@@ -352,7 +352,7 @@ class Dog(BaseModel):
         class _OutputFormatter(BaseModel):
             """Output formatter. Should always be used to format your response to the user."""  # noqa: E501
 
-            output: output_schema  # type: ignore
+            output: output_schema  # type: ignore[valid-type]
 
         function = _OutputFormatter
         output_parser = output_parser or PydanticAttrOutputFunctionsParser(
@@ -537,7 +537,7 @@ class Dog(BaseModel):
         class _OutputFormatter(BaseModel):
             """Output formatter. Should always be used to format your response to the user."""  # noqa: E501
 
-            output: output_schema  # type: ignore
+            output: output_schema  # type: ignore[valid-type]
 
         function = _OutputFormatter
         output_parser = output_parser or PydanticAttrOutputFunctionsParser(

libs/community/langchain_community/chains/graph_qa/cypher.py

@@ -316,7 +316,7 @@ def from_llm(
                         MessagesPlaceholder(variable_name="function_response"),
                     ]
                 )
-                qa_chain = response_prompt | qa_llm | StrOutputParser()  # type: ignore
+                qa_chain = response_prompt | qa_llm | StrOutputParser()  # type: ignore[operator]
             except (NotImplementedError, AttributeError):
                 raise ValueError("Provided LLM does not support native tools/functions")
         else:
@@ -404,15 +404,15 @@ def _call(
             intermediate_steps.append({"context": context})
             if self.use_function_response:
                 function_response = get_function_response(question, context)
-                final_result = self.qa_chain.invoke(  # type: ignore
+                final_result = self.qa_chain.invoke(  # type: ignore[assignment]
                     {"question": question, "function_response": function_response},
                 )
             else:
-                result = self.qa_chain.invoke(  # type: ignore
+                result = self.qa_chain.invoke(
                     {"question": question, "context": context},
                     callbacks=callbacks,
                 )
-                final_result = result[self.qa_chain.output_key]  # type: ignore
+                final_result = result[self.qa_chain.output_key]  # type: ignore[union-attr]
 
         chain_result: Dict[str, Any] = {self.output_key: final_result}
         if self.return_intermediate_steps:

libs/community/langchain_community/chains/graph_qa/memgraph.py

@@ -225,11 +225,11 @@ def from_llm(
                         MessagesPlaceholder(variable_name="function_response"),
                     ]
                 )
-                qa_chain = response_prompt | qa_llm | StrOutputParser()  # type: ignore
+                qa_chain = response_prompt | qa_llm | StrOutputParser()  # type: ignore[operator]
             except (NotImplementedError, AttributeError):
                 raise ValueError("Provided LLM does not support native tools/functions")
         else:
-            qa_chain = use_qa_llm_kwargs["prompt"] | qa_llm | StrOutputParser()  # type: ignore
+            qa_chain = use_qa_llm_kwargs["prompt"] | qa_llm | StrOutputParser()
 
         prompt = use_cypher_llm_kwargs["prompt"]
         llm_to_use = cypher_llm if cypher_llm is not None else llm
@@ -300,11 +300,11 @@ def _call(
             intermediate_steps.append({"context": context})
             if self.use_function_response:
                 function_response = get_function_response(question, context)
-                result = self.qa_chain.invoke(  # type: ignore
+                result = self.qa_chain.invoke(
                     {"question": question, "function_response": function_response},
                 )
             else:
-                result = self.qa_chain.invoke(  # type: ignore
+                result = self.qa_chain.invoke(
                     {"question": question, "context": context},
                     callbacks=callbacks,
                 )

libs/community/langchain_community/chains/graph_qa/neptune_cypher.py

@@ -67,11 +67,11 @@ def extract_cypher(text: str) -> str:
 
 def use_simple_prompt(llm: BaseLanguageModel) -> bool:
     """Decides whether to use the simple prompt"""
-    if llm._llm_type and "anthropic" in llm._llm_type:  # type: ignore
+    if llm._llm_type and "anthropic" in llm._llm_type:  # type: ignore[attr-defined]
         return True
 
     # Bedrock anthropic
-    if hasattr(llm, "model_id") and "anthropic" in llm.model_id:  # type: ignore
+    if hasattr(llm, "model_id") and "anthropic" in llm.model_id:
         return True
 
     return False

libs/community/langchain_community/chains/pebblo_retrieval/base.py

@@ -313,8 +313,12 @@ async def _aget_docs(
         )
 
     @staticmethod
-    def _get_app_details(  # type: ignore
-        app_name: str, owner: str, description: str, llm: BaseLanguageModel, **kwargs
+    def _get_app_details(
+        app_name: str,
+        owner: str,
+        description: str,
+        llm: BaseLanguageModel,
+        **kwargs: Any,
     ) -> App:
         """Fetch app details. Internal method.
         Returns: