Update app.py

app.py

@@ -465,11 +465,14 @@ def get_response_from_pdf(query, model, selected_docs, num_calls=3, temperature=
 
     for doc_name in selected_docs:
         logging.info(f"Retrieving documents for: {doc_name}")
-        doc_filter = lambda doc: doc.metadata["source"] == doc_name
-        doc_retriever = database.as_retriever(search_kwargs={"k": k_per_doc, "filter": doc_filter})
-        relevant_docs = doc_retriever.get_relevant_documents(query)
-        all_filtered_docs.extend(relevant_docs)
-        logging.info(f"Retrieved {len(relevant_docs)} documents for {doc_name}")
+        try:
+            doc_filter = lambda doc: isinstance(doc, dict) and doc.get('metadata', {}).get('source') == doc_name
+            doc_retriever = database.as_retriever(search_kwargs={"k": k_per_doc, "filter": doc_filter})
+            relevant_docs = doc_retriever.get_relevant_documents(query)
+            all_filtered_docs.extend(relevant_docs)
+            logging.info(f"Retrieved {len(relevant_docs)} documents for {doc_name}")
+        except Exception as e:
+            logging.error(f"Error retrieving documents for {doc_name}: {str(e)}")
 
     logging.info(f"Total number of filtered documents: {len(all_filtered_docs)}")
     
@@ -479,10 +482,15 @@ def get_response_from_pdf(query, model, selected_docs, num_calls=3, temperature=
         return
 
     for doc in all_filtered_docs:
-        logging.info(f"Document source: {doc.metadata['source']}")
-        logging.info(f"Document content preview: {doc.page_content[:100]}...")  # Log first 100 characters of each document
+        try:
+            source = doc.metadata['source'] if isinstance(doc, Document) else doc.get('metadata', {}).get('source', 'Unknown')
+            content = doc.page_content if isinstance(doc, Document) else doc.get('page_content', '')
+            logging.info(f"Document source: {source}")
+            logging.info(f"Document content preview: {content[:100]}...")  # Log first 100 characters of each document
+        except Exception as e:
+            logging.error(f"Error processing document: {str(e)}")
 
-    context_str = "\n".join([doc.page_content for doc in all_filtered_docs])
+    context_str = "\n".join([doc.page_content if isinstance(doc, Document) else doc.get('page_content', '') for doc in all_filtered_docs])
     logging.info(f"Total context length: {len(context_str)}")
 
     if model == "@cf/meta/llama-3.1-8b-instruct":