feat: Implement LangChain adapter for agent conversion

[Robbe1991]

Description

Implement adapter to convert trained models to LangChain agents.

Acceptance Criteria

• LangChainAdapter in src/agentgym/integrations/langchain.py
• Inherits from FrameworkAdapter
• Converts trained model to LangChain AgentExecutor
• Tool extraction from LangChain agents
• Type hints and docstrings
• Integration test with real LangChain agent

Technical Details

from langchain.agents import AgentExecutor

class LangChainAdapter(FrameworkAdapter):
    """Adapter for LangChain agents."""

    def wrap_agent(self, trained_model):
        """Convert trained model to LangChain agent."""
        # Create LangChain-compatible agent
        agent = self.create_langchain_agent(trained_model)
        return AgentExecutor(agent=agent, tools=self.tools)

    def extract_tools(self, agent):
        """Extract tools from LangChain agent."""
        return agent.tools

    def create_environment(self, agent):
        """Wrap LangChain agent for RL training."""
        return LangChainEnvironment(
            agent=agent,
            tools=self.extract_tools(agent),
        )

References

• LangChain docs: https://python.langchain.com/docs/modules/agents/
• Community analysis: LangChain pain points

Estimated Time

4-6 hours