Update rag.py

rag.py

@@ -1,4 +1,5 @@
 import os
+import logging
 import streamlit as st
 from dotenv import load_dotenv
 import pickle
@@ -15,20 +16,7 @@ from llama_index.core import get_response_synthesizer
 from llama_index.core.agent import ReActAgent
 from chromadb import PersistentClient
 
-# writable cache directories
-CACHE_DIR = "/app/.cache"
-HUGGINGFACE_CACHE = os.path.join(CACHE_DIR, "huggingface")
-VECTORDb_DIR = "/app/vectordb"
-STREAMLIT_CONFIG = "/app/.streamlit"
-
-os.makedirs(HUGGINGFACE_CACHE, exist_ok=True)
-os.makedirs(VECTORDb_DIR, exist_ok=True)
-os.makedirs(STREAMLIT_CONFIG, exist_ok=True)
-
-os.environ["TRANSFORMERS_CACHE"] = CACHE_DIR
-os.environ["HF_HOME"] = CACHE_DIR
-os.environ["XDG_CACHE_HOME"] = CACHE_DIR
-os.environ["STREAMLIT_HOME"] = STREAMLIT_CONFIG
+logger = logging.getLogger(__name__)
 
 @st.cache_resource
 def setup_rag_system(debug=False):
@@ -48,11 +36,7 @@ def setup_rag_system(debug=False):
     )
 
     # Embeddings
-    embedding_model = HuggingFaceEmbedding(
-        model_name="sentence-transformers/all-MiniLM-L6-v2",
-        cache_folder=HUGGINGFACE_CACHE
-    )
-
+    embedding_model = HuggingFaceEmbedding(model_name="sentence-transformers/all-MiniLM-L6-v2")
 
     # Persisted vector DBs
     persist_dirs = [
@@ -92,7 +76,7 @@ def setup_rag_system(debug=False):
         hybrid_retriever = RecursiveRetriever(
             "vector",
             retriever_dict={"vector": vector_retriever, "bm25": bm25_retriever},
-            verbose=False
+            verbose=True
         )
         hybrid_retrievers.append(hybrid_retriever)
 
@@ -178,11 +162,13 @@ def setup_rag_system(debug=False):
     agent = ReActAgent(
         tools=retriever_tools,
         llm=llm,
-        verbose=False,
+        verbose=True,
         max_iterations=20,
         system_prompt=system_prompt
     )
 
+    logger.info("RAG system setup complete.")
+
     if debug:
         return agent, llm, hybrid_retrievers