Update app.py

app.py

@@ -6,7 +6,6 @@ import openpyxl
 from pdfminer.high_level import extract_text
 import csv
 from pptx import Presentation
-from io import StringIO  # For handling string-based CSV
 
 # Configure Gemini API
 GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
@@ -34,11 +33,11 @@ def load_and_extract_text(file_path):
         elif file_extension == 'pdf':
             text = extract_text(file_path)
         elif file_extension == 'csv':
-            with open(file_path, 'r', encoding='utf-8') as csvfile: # Explicit encoding
+            with open(file_path, 'r', encoding='utf-8') as csvfile:
                 reader = csv.reader(csvfile)
-                text = '\n'.join([' '.join(row) for row in reader]) # Join rows with spaces
+                text = '\n'.join([' '.join(row) for row in reader])
         elif file_extension == 'txt':
-            with open(file_path, 'r', encoding='utf-8') as txtfile: #Explicit encoding
+            with open(file_path, 'r', encoding='utf-8') as txtfile:
                 text = txtfile.read()
         elif file_extension == 'pptx':
             prs = Presentation(file_path)
@@ -51,11 +50,8 @@ def load_and_extract_text(file_path):
         return f"Error processing file: {e}"
 
 
-def process_document_and_answer(extracted_text, question):
-    if "Error processing file" in extracted_text or "Unsupported file format" in extracted_text:
-        return extracted_text
-
-    prompt = f"Context: {extracted_text}\n\nQuestion: {question}\n\nAnswer:"
+def process_document_and_answer(context, question, history=""):
+    prompt = f"Context: {context}\n\nConversation History: {history}\n\nQuestion: {question}\n\nAnswer:"
 
     try:
         response = model.generate_content(prompt)
@@ -67,19 +63,52 @@ def process_document_and_answer(extracted_text, question):
 # Streamlit UI
 st.title("Gemini Document Q&A")
 
-uploaded_file = st.file_uploader("Upload a document", type=["docx", "xlsx", "pdf", "csv", "txt", "pptx"])
+# Initialize session state for conversation history
+if 'conversation_history' not in st.session_state:
+    st.session_state.conversation_history = ""
+
+if 'document_content' not in st.session_state:
+    st.session_state.document_content = None
 
-question = st.text_input("Ask a question about the document:")
+# File Upload
+uploaded_file = st.file_uploader("Upload a document", type=["docx", "xlsx", "pdf", "csv", "txt", "pptx"])
 
-if uploaded_file is not None and question:
-    # Save the uploaded file to a temporary location
+if uploaded_file is not None:
+    # Save the uploaded file to a temporary location (important for Streamlit)
     temp_file_path = "temp." + uploaded_file.name.split('.')[-1].lower()
     with open(temp_file_path, "wb") as temp_file:
         temp_file.write(uploaded_file.read())
 
-    extracted_text = load_and_extract_text(temp_file_path)
+    st.session_state.document_content = load_and_extract_text(temp_file_path)
+    os.remove(temp_file_path)  # Clean up
+
+    if "Error processing file" in st.session_state.document_content:
+        st.error(st.session_state.document_content)
+        st.session_state.document_content = None # Clear document content to prevent further processing
+
+if st.session_state.document_content:
+    st.subheader("Document Content Preview:")
+    st.text(st.session_state.document_content[:500] + "..." if len(st.session_state.document_content) > 500 else st.session_state.document_content)  # Display a snippet
+
+
+# Conversation Loop
+while True:
+    question = st.text_input("Ask a question about the document (type 'exit' to end):", key=f"question_{len(st.session_state.conversation_history)}") # Unique key for each input
+
+    if question.lower() == "exit":
+        st.write("Ending conversation.")
+        break
+
+    if question and st.session_state.document_content:
+        answer = process_document_and_answer(st.session_state.document_content, question, st.session_state.conversation_history)
+
+        st.write("Answer:", answer)
+
+        # Update conversation history
+        st.session_state.conversation_history += f"\nQuestion: {question}\nAnswer: {answer}"
 
-    os.remove(temp_file_path)  # Clean up the temporary file
 
-    answer = process_document_and_answer(extracted_text, question)
-    st.write("Answer:", answer)
+# Display conversation history
+if st.session_state.conversation_history:
+    st.subheader("Conversation History:")
+    st.text(st.session_state.conversation_history)