Update main.py

main.py

@@ -10,9 +10,9 @@ from langchain_huggingface import HuggingFaceEmbeddings
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_community.vectorstores import FAISS
 from langchain.prompts import PromptTemplate
+from langchain_together import Together
 from langchain.memory import ConversationBufferMemory
 from langchain.chains import ConversationalRetrievalChain
-from langchain.schema import Document
 
 # ==========================
 # Logging Configuration
@@ -30,12 +30,17 @@ warnings.filterwarnings("ignore", message="You are using `torch.load` with `weig
 # Load Environment Variables
 # ==========================
 load_dotenv()
+TOGETHER_AI_API = os.getenv("TOGETHER_AI")
 HF_HOME = os.getenv("HF_HOME", "./cache")
 os.environ["HF_HOME"] = HF_HOME
 
 # Ensure the HF_HOME directory exists
 os.makedirs(HF_HOME, exist_ok=True)
 
+# Validate required environment variables
+if not TOGETHER_AI_API:
+    raise ValueError("The TOGETHER_AI_API environment variable is missing. Please set it in your .env file.")
+
 # ==========================
 # Initialize Embeddings
 # ==========================
@@ -74,18 +79,32 @@ ANSWER:
 prompt = PromptTemplate(template=prompt_template, input_variables=["context", "question", "chat_history"])
 
 # ==========================
-# Conversational Retrieval Chain
+# Initialize Together API
 # ==========================
-memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
+try:
+    llm = Together(
+        model="mistralai/Mistral-7B-Instruct-v0.2",
+        temperature=0.5,
+        max_tokens=1024,
+        together_api_key=TOGETHER_AI_API,
+    )
+    logger.info("Together API successfully initialized.")
+except Exception as e:
+    logger.error(f"Error initializing Together API: {e}")
+    raise RuntimeError("Something went wrong with the Together API setup. Please verify your API key and configuration.")
 
+# ==========================
+# Conversational Retrieval Chain (RAG Implementation)
+# ==========================
+memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
 qa = ConversationalRetrievalChain.from_llm(
-    llm="nomic-ai/nomic-embed-text-v1",
+    llm=llm,
     memory=memory,
     retriever=db_retriever,
     combine_docs_chain_kwargs={"prompt": prompt},
+    return_source_documents=True  # This enables logging of retrieved content
 )
-
-logger.info("Conversational Retrieval Chain initialized.")
+logger.info("Conversational Retrieval Chain initialized with RAG capabilities.")
 
 # ==========================
 # FastAPI Backend
@@ -106,25 +125,16 @@ async def root():
 async def chat(request: ChatRequest):
     try:
         logger.debug(f"Received user question: {request.question}")
-        
-        # Retrieve relevant documents
-        docs = db_retriever.get_relevant_documents(request.question)
-        if not docs:
-            logger.warning("No relevant documents found.")
-            return ChatResponse(answer="I'm sorry, I couldn't find relevant information for your query.")
-
-        # Log retrieved documents for debugging
-        for i, doc in enumerate(docs, start=1):
-            logger.debug(f"Retrieved Document {i}: {doc.page_content[:500]}...")
-
-        # Invoke the conversational retrieval chain
         result = qa.invoke(input=request.question)
-        answer = result.get("answer", "I'm sorry, I couldn't find relevant information for your query.")
         
-        # Log the final answer
-        logger.debug(f"Final Answer: {answer}")
+        # Log the retrieved source documents for debugging purposes
+        source_docs = result.get("source_documents")
+        logger.debug(f"Retrieved source documents: {source_docs}")
+        
+        answer = result.get("answer")
+        if not answer or "The information is not available in the provided context" in answer:
+            answer = "I'm sorry, I couldn't find relevant information for your query. Please try rephrasing or providing more details."
         return ChatResponse(answer=answer)
-
     except Exception as e:
         logger.error(f"Error during chat invocation: {e}")
         raise HTTPException(status_code=500, detail="Oops! Something went wrong on our end. Please try again later.")