:wrench: Fix audio

app.py

@@ -1,23 +1,21 @@
 import os
 import io
-import json
 import base64
 import gradio as gr
 import requests
-import inspect
 import pandas as pd
 from dotenv import load_dotenv
 from smolagents import CodeAgent, OpenAIServerModel, DuckDuckGoSearchTool
-import openai
-from smolagents import tool
-from PIL import Image
-import soundfile as sf
+from openai import OpenAI
+
+from PIL import Image, UnidentifiedImageError
 
 load_dotenv()
 
 # (Keep Constants as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
+client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
 
 
 # --- Basic Agent Definition ---
@@ -32,8 +30,6 @@ class BasicAgent:
 
         OPENAI_MODEL_ID = os.getenv("OPENAI_MODEL_ID", "gpt-4o")
 
-        openai.api_key = OPENAI_API_KEY
-
         self.agent = CodeAgent(
             tools=[DuckDuckGoSearchTool()],
             model=OpenAIServerModel(model_id=OPENAI_MODEL_ID, api_key=OPENAI_API_KEY),
@@ -83,7 +79,7 @@ def load_file_from_response(response):
 
     try:
         if "application/json" in content_type:
-            if "No file path" in response.json()["detail"]:
+            if "No file path" in response.json().detail:
                 return None
             return {"type": "json", "data": response.json()}
 
@@ -98,7 +94,7 @@ def load_file_from_response(response):
 
         elif "audio/" in content_type:
             # Transcribe audio using OpenAI Whisper
-            transcript = openai.Audio.transcribe(
+            transcript = client.audio.transcriptions.create(
                 model="whisper-1", file=io.BytesIO(content_bytes)
             )
             return {"type": "text", "data": transcript.get("text", "")}
@@ -121,7 +117,7 @@ def load_file_from_response(response):
             # Try audio
             try:
                 # Transcribe audio from raw bytes
-                transcript = openai.Audio.transcribe(
+                transcript = client.audio.transcriptions.create(
                     model="whisper-1", file=io.BytesIO(content_bytes)
                 )
                 return {"type": "text", "data": transcript.get("text", "")}
@@ -170,8 +166,8 @@ def describe_image(image_path: str) -> str:
             ],
         }
     ]
-    response = openai.ChatCompletion.create(model="gpt-4o", messages=messages)
-    return response.choices[0].message["content"]
+    response = client.chat.completions.create(model="gpt-4o", messages=messages)
+    return response.choices[0].message.content
 
 
 def run_and_submit_all(profile: gr.OAuthProfile | None):