Update app.py

app.py

@@ -2,8 +2,8 @@ from sentence_transformers import SentenceTransformer
 from PyPDF2 import PdfReader
 import tiktoken
 import groq
-import asyncio
-from groq import AsyncGroq
+#import asyncio
+#from groq import AsyncGroq
 import faiss
 import numpy as np
 import gradio as gr
@@ -84,20 +84,16 @@ def load_embeddings(embeddings_path="models/embeddings.pkl"):
         index = pickle.load(f)
 
 # == Integrasi LLaMA via Groq API ==
-#client = groq.Client(api_key=GROQ_API_KEY)
-client = AsyncGroq(api_key=GROQ_API_KEY)
+client = groq.Client(api_key=GROQ_API_KEY)
+#client = AsyncGroq(api_key=GROQ_API_KEY)
 
-async def query_llama(prompt):
-    response = await client.chat.completions.create(
+def query_llama(prompt):
+    response = client.chat.completions.create(
         model="llama3-8b-8192",
         messages=[{"role": "user", "content": prompt}],
         max_tokens=512
     )
-    for chunk in response:
-        print(chunk.choices[0].delta.content)
-        yield chunk.choices[0].delta.content
-        print("****************")
-    #return response.choices[0].message.content.strip()
+    return response.choices[0].message.content.strip()
 
 # == Main Workflow ==
 if __name__ == '__main__':
@@ -112,7 +108,6 @@ if __name__ == '__main__':
     context = "\n".join(retrieved_chunks)
     
     prompt = f"Gunakan informasi berikut untuk menjawab:\n{context}\n\nPertanyaan: Apa isi dokumen ini?"
-    print(prompt)
     answer = query_llama(prompt)
     print(answer)
 
@@ -122,7 +117,7 @@ def chatbot_interface(user_query):
     context = "\n".join(retrieved_chunks)
     
     prompt = f"Gunakan informasi berikut untuk menjawab:\n{context}\n\nPertanyaan: {user_query}"
-    answer = asyncio.run(query_llama(prompt))
+    answer = query_llama(prompt)
     
     return answer