Update app.py

app.py

@@ -1,12 +1,12 @@
 import streamlit as st
 import os
 import logging
+import faiss
+import numpy as np
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
 from langchain_community.embeddings import HuggingFaceEmbeddings
-from langchain_community.vectorstores import Chroma
-from langchain_community.llms import HuggingFacePipeline
 from langchain.chains import RetrievalQA
-from ingest import create_chroma_db
+from langchain.vectorstores import FAISS
 
 # Set up logging
 logging.basicConfig(level=logging.INFO)
@@ -29,29 +29,24 @@ def load_llm():
     )
     return HuggingFacePipeline(pipeline=pipe)
 
-def load_chroma_db():
-    chroma_dir = "chroma_db"
-    if not os.path.exists(chroma_dir):
-        st.warning("Chroma database not found. Creating a new one...")
-        create_chroma_db()
-
-    if not os.path.exists(chroma_dir):
-        st.error("Failed to create the Chroma database. Please check the 'docs' directory and try again.")
-        raise RuntimeError("Chroma database creation failed.")
-
+def load_faiss_index():
+    index_path = "faiss_index.index"
+    if not os.path.exists(index_path):
+        st.warning("FAISS index not found. Please create the index first.")
+        raise RuntimeError("FAISS index not found.")
+    
     try:
-        embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
-        db = Chroma.load_local(chroma_dir, embeddings)
-        logger.info(f"Chroma database loaded successfully from {chroma_dir}")
-        return db.as_retriever()
+        faiss_index = faiss.read_index(index_path)
+        logger.info(f"FAISS index loaded successfully from {index_path}")
+        return FAISS(faiss_index)
     except Exception as e:
-        st.error(f"Failed to load Chroma database: {e}")
-        logger.exception("Exception in load_chroma_db")
+        st.error(f"Failed to load FAISS index: {e}")
+        logger.exception("Exception in load_faiss_index")
         raise
 
 def process_answer(instruction):
     try:
-        retriever = load_chroma_db()
+        retriever = load_faiss_index()
         llm = load_llm()
         qa = RetrievalQA.from_chain_type(
             llm=llm,