Update app.py

app.py

@@ -7,7 +7,7 @@ from transformers import AutoTokenizer, AutoModel
 from weaviate.classes.init import Auth
 import cohere
 
-# Load credentials from environment variables
+# Load credentials from environment variables or hardcode them temporarily
 WEAVIATE_URL = "vgwhgmrlqrqqgnlb1avjaa.c0.us-west3.gcp.weaviate.cloud"
 WEAVIATE_API_KEY = "7VoeYTjkOS4aHINuhllGpH4JPgE2QquFmSMn"
 COHERE_API_KEY = "LEvCVeZkqZMW1aLYjxDqlstCzWi4Cvlt9PiysqT8"
@@ -67,16 +67,19 @@ def generate_response(context, query):
 
 def qa_pipeline(pdf_file, query):
     """Main pipeline for QA: parse PDF, embed chunks, query Weaviate, and generate answer."""
-    document_text = load_pdf(pdf_file)
-    document_chunks = [document_text[i:i+500] for i in range(0, len(document_text), 500)]
+    try:
+        document_text = load_pdf(pdf_file)
+        document_chunks = [document_text[i:i+500] for i in range(0, len(document_text), 500)]
 
-    upload_document_chunks(document_chunks)
-    top_docs = query_answer(query)
+        upload_document_chunks(document_chunks)
+        top_docs = query_answer(query)
 
-    context = ' '.join([doc.properties['content'] for doc in top_docs])
-    answer = generate_response(context, query)
+        context = ' '.join([doc.properties['content'] for doc in top_docs])
+        answer = generate_response(context, query)
 
-    return context, answer
+        return str(context), str(answer)
+    finally:
+        client.close()  # ✅ Properly close client to avoid memory leaks
 
 # Gradio UI
 with gr.Blocks(theme="compact") as demo:
@@ -103,7 +106,7 @@ with gr.Blocks(theme="compact") as demo:
             answer_output = gr.Textbox(label="💬 Answer", lines=3)
 
     submit_button.click(
-        qa_pipeline,
+        fn=qa_pipeline,
         inputs=[pdf_input, query_input],
         outputs=[doc_segments_output, answer_output]
     )
@@ -141,4 +144,4 @@ with gr.Blocks(theme="compact") as demo:
         """
     )
 
-demo.launch()
+demo.launch(share=True)  # ✅ Required for Hugging Face Spaces