Open
Conversation
Implements LangChainRunner to support LangChain agents with the same orchestration capabilities as OpenAIRunner. Changes: - Add LangChainRunner class with support for AgentExecutor - Implement async invoke and streaming via astream_events - Add max iterations recovery similar to OpenAI's max turns - Include report_status tool injection for activity tracking - Add optional langchain dependencies to pyproject.toml - Create comprehensive LangChain example in examples/langchain-agents-fastapi/ - Update main README with LangChain documentation and examples The LangChainRunner provides the same features as OpenAIRunner: - Automatic activity tracking - Agent self-reporting via report_activity tool - Max iterations recovery with wrap-up prompts - Streaming support - Compatible with ReAct, tool-calling, and LangGraph agents
…t-runner-01GheBmLofGM74vWbokKfPvK
Refactored LangChain integration to align with: 1. Latest agentexec architecture (v0.1.0+ changes) 2. LangChain's native early_stopping_method feature 3. Latest LangChain conventions and best practices Changes to LangChain Runner: - Leverage LangChain's early_stopping_method='generate' instead of manual exception handling - Simplified run() and run_streamed() methods - no try/except needed - AgentExecutor handles max iterations gracefully without throwing exceptions - Better documentation explaining the difference from OpenAI's approach - Added notes about LangGraph's GraphRecursionError for future support Changes to LangChain Example: - Updated to match new agentexec architecture with typed contexts - Context now uses Pydantic BaseModel (ResearchCompanyContext) - Task handlers signature: async def handler(agent_id, context) -> Result - Added typed return values (ResearchCompanyResult) - Created context.py and db.py following OpenAI example pattern - Updated views.py to use typed context instead of generic payload dict - Updated main.py to use new imports and patterns - Changed pool.start()/shutdown() to pool.run() - Updated README with corrected examples and curl commands Key Improvements: - Defers to LangChain's built-in max iterations recovery - Cleaner, simpler code without manual error handling - Better alignment with LangChain conventions - Matches latest agentexec patterns from main branch Sources: - https://python.langchain.com/docs/modules/agents/how_to/max_iterations/ - https://python.langchain.com/api_reference/langchain/agents/langchain.agents.agent.AgentExecutor.html
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Implements LangChainRunner to support LangChain agents with the same orchestration capabilities as OpenAIRunner.
Changes:
The LangChainRunner provides the same features as OpenAIRunner: