utils/langchain_enhancements.py

# utils/langchain_enhancements.py

from langchain.memory import ConversationSummaryMemory
from langchain.agents import Tool, AgentExecutor, OpenAIFunctions
from langchain.tools import DuckDuckGoSearchRun
from langchain.retrievers import TimeWeightedVectorStoreRetriever
from langchain.embeddings import OpenAIEmbeddings
from langchain.vectorstores import Chroma
from langchain.chat_models import ChatAnthropic

class EnhancedLearningSystem:
    def __init__(self, anthropic_api_key):
        self.llm = ChatAnthropic(anthropic_api_key=anthropic_api_key)
        self.memory = self._setup_memory()
        self.tools = self._setup_tools()
        self.agent_executor = self._setup_agent()

    def _setup_memory(self):
        """Setup enhanced memory system"""
        # Create vectorstore for storing chat history
        embeddings = OpenAIEmbeddings()
        vectorstore = Chroma(
            collection_name="chat_history",
            embedding_function=embeddings
        )

        # Create time-weighted retriever
        retriever = TimeWeightedVectorStoreRetriever(
            vectorstore=vectorstore,
            decay_rate=0.01,
            k=5
        )

        # Create summary memory
        memory = ConversationSummaryMemory(
            llm=self.llm,
            memory_key="chat_history",
            return_messages=True
        )

        return {
            'summary': memory,
            'retriever': retriever
        }

    def _setup_tools(self):
        """Setup tools for the learning system"""
        search = DuckDuckGoSearchRun()

        tools = [
            Tool(
                name="Search",
                func=search.run,
                description="Useful for finding current trading information and examples"
            ),
            Tool(
                name="Historical Context",
                func=self._get_historical_context,
                description="Get relevant historical chat context"
            ),
            Tool(
                name="Learning Progress",
                func=self._check_learning_progress,
                description="Check user's learning progress and suggest next topics"
            )
        ]

        return tools

    def _setup_agent(self):
        """Setup the learning agent"""
        agent = OpenAIFunctions.from_llm_and_tools(
            llm=self.llm,
            tools=self.tools,
            memory=self.memory['summary']
        )

        return AgentExecutor.from_agent_and_tools(
            agent=agent,
            tools=self.tools,
            memory=self.memory['summary'],
            verbose=True
        )

    async def _get_historical_context(self, topic):
        """Retrieve relevant historical context"""
        relevant_docs = await self.memory['retriever'].aget_relevant_documents(topic)
        return [doc.page_content for doc in relevant_docs]

    def _check_learning_progress(self, topic):
        """Check user's progress in a topic"""
        # Implement progress tracking logic
        return {
            'mastered_concepts': [],
            'in_progress': [],
            'suggested_next': []
        }

    async def process_question(self, question):
        """Process a learning question with enhanced features"""
        # Get historical context
        history = await self._get_historical_context(question)
        
        # Execute agent
        response = await self.agent_executor.arun(
            input={
                'question': question,
                'history': history,
                'progress': self._check_learning_progress(question)
            }
        )

        # Update memory
        self.memory['summary'].save_context(
            {'input': question},
            {'output': response}
        )

        return {
            'response': response,
            'context': history,
            'progress': self._check_learning_progress(question)
        }