Update app.py

app.py

@@ -15,6 +15,7 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 import os
 import transformers
 import torch
+from langchain_retrieval import BaseRetrieverChain
 # from dotenv import load_dotenv
 
 # load_dotenv()
@@ -61,21 +62,49 @@ def get_context_retriever_chain(vector_store,llm):
     return retriever_chain
     
 
-def get_conversational_rag_chain(retriever_chain,llm): 
+# def get_conversational_rag_chain(retriever_chain,llm): 
     
-    llm=llm
+#     llm=llm
     
+#     template = "Answer the user's questions based on the below context:\n\n{context}"
+#     human_template = "{input}"
+    
+#     prompt = ChatPromptTemplate.from_messages([
+#       ("system", template),
+#       MessagesPlaceholder(variable_name="chat_history"),
+#       ("user", human_template),
+#     ])
+    
+#     stuff_documents_chain = create_stuff_documents_chain(llm,prompt)
+    
+#     return create_retrieval_chain(retriever_chain, stuff_documents_chain)
+def get_conversational_rag_chain(
+    retriever_chain: Optional[langchain_retrieval.BaseRetrieverChain],
+    llm: Callable[[str], str],
+    chat_history: Optional[langchain_core.prompts.chat.ChatPromptValue] = None,
+) -> langchain_retrieval.BaseRetrieverChain:
+
+    if not retriever_chain:
+        raise ValueError("`retriever_chain` cannot be None or an empty object.")
+
     template = "Answer the user's questions based on the below context:\n\n{context}"
     human_template = "{input}"
-    
+
     prompt = ChatPromptTemplate.from_messages([
       ("system", template),
       MessagesPlaceholder(variable_name="chat_history"),
       ("user", human_template),
     ])
-    
-    stuff_documents_chain = create_stuff_documents_chain(llm,prompt)
-    
+
+    def safe_llm(input_str: str) -> str:
+        if isinstance(input_str, langchain_core.prompts.chat.ChatPromptValue):
+            input_str = str(input_str)
+
+        # Call the original llm, which should now work correctly
+        return llm(input_str)
+
+    stuff_documents_chain = create_stuff_documents_chain(safe_llm, prompt)
+
     return create_retrieval_chain(retriever_chain, stuff_documents_chain)
 
 def get_response(user_input):