Update app.py

app.py

@@ -7,7 +7,7 @@ from langchain_community.embeddings import HuggingFaceEmbeddings
 from langchain_community.vectorstores import FAISS
 from langchain_community.llms import HuggingFacePipeline
 from langchain.chains import RetrievalQA
-from ingest import create_faiss_index
+from ingest import create_faiss_index  # Make sure this function is updated
 
 # Set up logging
 logging.basicConfig(level=logging.INFO)
@@ -32,9 +32,6 @@ def load_llm():
 
 def validate_index_file(index_path):
     try:
-        if os.path.getsize(index_path) == 0:
-            st.error(f"Index file '{index_path}' is empty.")
-            return False
         with open(index_path, 'rb') as f:
             data = f.read(100)
         logger.info(f"Successfully read {len(data)} bytes from the index file")
@@ -45,13 +42,13 @@ def validate_index_file(index_path):
 
 def load_faiss_index():
     index_path = "faiss_index/index.faiss"
-    
-    if not os.path.exists(index_path) or not validate_index_file(index_path):
-        st.warning("Index file is missing or corrupted. Creating a new one...")
-        if os.path.exists(index_path):
-            os.remove(index_path)
-            st.info("Deleted the corrupted index file.")
-        create_faiss_index()
+    if not os.path.exists(index_path):
+        st.warning("Index file not found. Creating a new one...")
+        # Pass the required arguments to create_faiss_index
+        documents = []  # Load your documents here or from a specific directory
+        texts = []      # Process your documents to get text chunks
+        embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+        create_faiss_index(texts, embeddings)  # Ensure this is updated with correct function
 
     if not os.path.exists(index_path):
         st.error("Failed to create the FAISS index. Please check the 'docs' directory and try again.")