Update app.py

app.py

@@ -75,9 +75,29 @@ def create_retriever_from_chroma(vectorstore_path="./docs/chroma/", search_type=
     
     retriever=vectorstore.as_retriever(search_type = search_type, search_kwargs={"k": k})
 
+    llm = llamacpp.LlamaCpp(
+        model_path = "JCHAVEROT_Qwen2-0.5B-Chat_SFT_DPO.Q8_0.gguf",
+        seed = 41,
+        n_gpu_layers=0,
+        temperature=0.0,
+        n_ctx=15000,
+        n_batch=2000,
+        max_tokens=1500,
+        repeat_penalty=1.8,
+        last_n_tokens_size = 200,
+        callback_manager=callback_manager,
+        verbose=False,
+    )
+
+    compressor = LLMChainExtractor.from_llm(llm)
+    
+    compression_retriever = ContextualCompressionRetriever(
+    base_compressor=compressor,
+    base_retriever=retriever
+)
     
 
-    return retriever
+    return compression_retriever
 
 
 
@@ -139,7 +159,7 @@ def main():
 def handle_userinput(user_question,retriever,rag_chain):
     st.session_state.messages.append({"role": "user", "content": user_question})
     st.chat_message("user").write(user_question)
-    docs = retriever.invoke(user_question)
+    docs = retriever.get_relevant_documents(user_question)
 
     with st.sidebar:
         st.subheader("Your documents")