library · inferred from evidence
LangChain
A library mentioned as a source of unhashable types (e.g., Document lists).
- create_tool_calling_agent raises ValueError: This function requires a .bind_tools method be implemented on...
Wrapping a chat model with .with_fallbacks([fallback_model]) and passing the result to create_tool_calling_agent raised ValueError: This function requires a .bind_tools method be implemented on the LLM. The same error also hit AzureChatOpen
- Qwen3 tool calls not parsed into AIMessage.tool_calls on OCI
Using Qwen3 models through ChatOCIGenAI with LangChain tool calling, the tool calls are not parsed into AIMessage.tool_calls. The model returns valid tool calls wrapped in [tool_call] tags, but they stay as raw text in content and tool_call
- LangSmith Python SDK memory grows in long-running apps (RunTree kept in memory)
Long-running Python app using LangGraph/LangChain (asyncio.gather with create_task, one graph per task). Memory grows steadily over time, but only when LangSmith tracing is enabled. tracemalloc and objgraph show RunTree objects accumulating
- LangChain agent: {tool_name} is not a valid tool, try another one (long tool names)
An agent intermittently failed with "{tool_name} is not a valid tool, try another one" — e.g. "retrieve_from_somewhere_else is not a valid tool, try [retrieve_from_vectorstore]" — even though both tools were registered. Wrapping an agent as
- ScrapeGraphAI 1.14.0+ SmartScraperGraph fails with pydantic ValidationError: 1 validation error for...
Since ScrapeGraphAI 1.14.0 (tested on 1.15.0), SmartScraperGraph on OpenAI died in the GenerateAnswer node with: pydantic.v1.error_wrappers.ValidationError: 1 validation error for Generation, text str type expected (type=type_error.str). Up
- OpenAIEmbeddings fails against Azure OpenAI: Too many inputs for model
Using langchain's OpenAIEmbeddings against Azure OpenAI failed with "InvalidRequestError: Too many inputs for model None. The max number of inputs is 1." Azure's embedding endpoint only accepts one input per request, so the default batched
- AI21 LLM is not Runnable when using via LangChain
A user tried to use AI21's models through LangChain agents by passing an AI21Client instance directly as the llm to load_tools and initialize_agent, and the agent would not run. The problem: AI21Client is the raw API client, not a LangChain
- ValueError: create_tool_calling_agent requires .bind_tools on the LLM when using with_fallbacks
Wrapping a chat model with .with_fallbacks([fallback_model]) and passing the result to create_tool_calling_agent raised ValueError: This function requires a .bind_tools method be implemented on the LLM. The same error also hit AzureChatOpen
- DeepSeek multi-turn agent 400 error: The `reasoning_content` in the thinking mode must be passed back to the API
Running a multi-turn agent with tool calls on a DeepSeek thinking model (deepseek-v4-flash via langchain-deepseek's ChatDeepSeek, built with create_agent). The first turn succeeded, but the second turn raised openai.BadRequestError: Error c
- streaming not working when using bind_tools with ChatOllama - llm.stream() returns no tokens
Streaming chat responses from an Ollama model (e.g. llama3.1:70b via ChatOllama, or ChatOpenAI pointed at Ollama's /v1 endpoint) with tools bound via llm.bind_tools([...]).stream(). With tools bound, streaming produced no tokens at all and
- Error in RootListenersTracer.on_chain_end callback: KeyError('answer') while streaming a RAG chain
Building a conversational RAG chatbot with create_retrieval_chain wrapped in RunnableWithMessageHistory (input_messages_key="input", history_messages_key="chat_history", output_messages_key="answer"). Calling astream() on the chain printed
- Chroma.from_documents() raises sqlite3.OperationalError: attempt to write a readonly database
Using Chroma as a persistent vector store via Chroma.from_documents(docs, embedding, persist_directory=...) on Ubuntu with langchain 0.0.351. The first creation worked, but deleting the persist directory with shutil.rmtree and recreating th