Update chat_with_doc.py

chat with doc use chunked data

chat_with_doc.py

@@ -1,7 +1,7 @@
 import os
 import openai
 import streamlit as st
-import chunk  # Import the chunking function from chunk.py
+import json
 
 # Load environment variables
 from dotenv import load_dotenv
@@ -25,15 +25,15 @@ def get_openai_response(prompt, model="gpt-4", max_tokens=150):
 # Streamlit UI
 st.title("Chat with Your Document")
 
-# Get the chunked data directly from chunk.py (assuming `chunk_text()` is already executed)
-# For example, you can load the chunked data from a file or directly invoke the chunking logic
-document_text = """Your long document text goes here..."""
-chunked_data = chunk.chunk_text(document_text)
-
-# Display chunked data
-st.write("Document has been chunked into the following parts:")
-for i, chunk_part in enumerate(chunked_data, 1):
-    st.write(f"**Chunk {i}:**\n{chunk_part}\n")
+# Load chunked data from the file (chunked_data.json)
+try:
+    with open("chunked_data.json", "r") as f:
+        chunked_data = json.load(f)
+        st.write("Document has been chunked into the following parts:")
+        for i, chunk_part in enumerate(chunked_data, 1):
+            st.write(f"**Chunk {i}:**\n{chunk_part}\n")
+except FileNotFoundError:
+    st.error("Chunked data not found. Please run the search and chunk process first.")
 
 # Input field for user to ask questions about the chunked document
 st.subheader("Ask a question about the document:")
@@ -48,3 +48,4 @@ if st.button("Get Answer"):
         st.write(f"**Answer**: {response}")
     else:
         st.error("Please provide a question.")
+