Update document_chat.py

document_chat.py

@@ -6,6 +6,9 @@ from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.chains import ConversationalRetrievalChain
 from langchain.memory import ConversationBufferMemory
 from langchain.llms import HuggingFaceHub
+from langchain.prompts import PromptTemplate
+from langchain.chains import LLMChain
+from langchain.chains.combine_documents import DocumentCompressor
 
 # Constants
 CHROMA_DB_PATH = "chroma_db"
@@ -43,11 +46,20 @@ def process_query_with_memory(query, chat_history=[]):
     # Load the LLM model from Hugging Face
     llm = HuggingFaceHub(repo_id=LLM_Model, model_kwargs={"max_new_tokens": 500})
 
+    # Create a PromptTemplate for the question generator
+    question_generator_template = "Generate a question based on the user's request: {query}"
+    question_generator = LLMChain(llm=llm, prompt=PromptTemplate(template=question_generator_template, input_variables=["query"]))
+
+    # Document combiner (example: just concatenate documents)
+    combine_docs_chain = DocumentCompressor(combine_fn=lambda docs: " ".join([doc.page_content for doc in docs]))
+
     # Create a ConversationalRetrievalChain with the loaded model and retriever
     qa_chain = ConversationalRetrievalChain(
         llm=llm,
         retriever=retriever,
-        memory=memory
+        memory=memory,
+        question_generator=question_generator,
+        combine_docs_chain=combine_docs_chain
     )
 
     # Run the query with the current chat history and return the response