Update app.py

app.py

@@ -1,26 +1,25 @@
 #using codes from mistralai official cookbook
 import gradio as gr
-from mistralai.async_client import MistralAsyncClient
+from mistralai.client import MistralClient
 from mistralai.models.chat_completion import ChatMessage
 import numpy as np
 import PyPDF2
 import faiss
 import os
-import httpx
-import asyncio
+
 
 mistral_api_key = os.environ.get("API_KEY")
 
-cli = MistralAsyncClient(api_key = mistral_api_key)
+cli = MistralClient(api_key = mistral_api_key)
 
-async def get_text_embedding(input: str):
+def get_text_embedding(input: str):
     embeddings_batch_response = cli.embeddings(
           model = "mistral-embed",
           input = input
       )
     return embeddings_batch_response.data[0].embedding
 
-async def rag_pdf(pdfs: list, question: str) -> str:
+def rag_pdf(pdfs: list, question: str) -> str:
     chunk_size = 4096
     chunks = []
     for pdf in pdfs:
@@ -34,11 +33,10 @@ async def rag_pdf(pdfs: list, question: str) -> str:
     question_embeddings = np.array([get_text_embedding(question)])
     D, I = index.search(question_embeddings, k = 4)
     retrieved_chunk = [chunks[i] for i in I.tolist()[0]]
-    print(retrieved_chunk)
     text_retrieved = "\n\n".join(retrieved_chunk)
     return text_retrieved
 
-async def ask_mistral(message: str, history: list):
+def ask_mistral(message: str, history: list):
     messages = []
     pdfs = message["files"]
     for couple in history:
@@ -58,23 +56,17 @@ async def ask_mistral(message: str, history: list):
             pdfs_extracted.append(txt)
 
         retrieved_text = rag_pdf(pdfs_extracted, message["text"])
-        print(retrieved_text)
         messages.append(ChatMessage(role = "user", content = retrieved_text + "\n\n" + message["text"]))
     else:
         messages.append(ChatMessage(role = "user", content = message["text"]))
 
     full_response = ""
-
-    async_response = cli.chat_stream(
-        model = "open-mistral-7b", 
-        messages = messages, 
-        max_tokens = 1024 
-    )
-
-    async for chunk in async_response:
+    for chunk in cli.chat_stream(model = "open-mistral-7b", messages = messages, max_tokens = 1024):
         full_response += chunk.choices[0].delta.content
         yield full_response
 
+
+
 chatbot = gr.Chatbot()
 
 with gr.Blocks(theme="soft") as demo: