Update app.py

app.py

@@ -4,24 +4,33 @@ import re
 import tempfile
 import os
 import warnings
-import requests
+from groq import Groq
+import torch
 from transformers import pipeline
-from whisper_jax import FlaxWhisperPipeline
+
+# Set up device for torch
+device = "cuda:0" if torch.cuda.is_available() else "cpu"
+
+# Set up ASR pipeline with a smaller model for reduced memory usage
+asr_pipeline = pipeline(
+    "automatic-speech-recognition",
+    model="openai/whisper-medium",
+    chunk_length_s=30,
+    device=device,
+)
 
 # Suppress specific warning
 warnings.filterwarnings("ignore", message="FP16 is not supported on CPU; using FP32 instead")
 
-# Set up Groq API key
-GROQ_API_KEY = os.environ.get("GROQ_API_KEY")
+# Set up Groq client
+client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
 
-# Instantiate FlaxWhisperPipeline
-pipeline = FlaxWhisperPipeline("openai/whisper-large-v2")
-
-# Function to transcribe audio using FlaxWhisperPipeline
+# Function to transcribe audio using ASR pipeline
 def transcribe_audio(file_path):
-    # JIT compile the forward call - slow, but we only do once
-    text = pipeline(file_path)
-    return text
+    with open(file_path, "rb") as f:
+        audio_bytes = f.read()
+    prediction = asr_pipeline(audio_bytes)["text"]
+    return prediction
 
 # Function to get transcript from YouTube
 def get_transcript(url):
@@ -36,30 +45,19 @@ def get_transcript(url):
     except Exception as e:
         return str(e)
 
-# Function to make a request to the Groq API
-def groq_request(payload):
-    headers = {
-        "Authorization": f"Bearer {GROQ_API_KEY}",
-        "Content-Type": "application/json"
-    }
-    response = requests.post("https://api.groq.com/v1/endpoint", json=payload, headers=headers)
-    response.raise_for_status()
-    return response.json()
-
 # Function to summarize text using Groq API
 def summarize_text(text):
     try:
-        payload = {
-            "model": "llama3-8b-8192",
-            "messages": [
+        response = client.chat.completions.create(
+            messages=[
                 {
                     "role": "user",
                     "content": f"Summarize the following text:\n\n{text}"
                 }
-            ]
-        }
-        response = groq_request(payload)
-        summary = response['choices'][0]['message']['content'].strip()
+            ],
+            model="llama3-8b-8192",
+        )
+        summary = response.choices[0].message.content.strip()
         return summary
     except Exception as e:
         return f"Error summarizing text: {e}"
@@ -67,17 +65,16 @@ def summarize_text(text):
 # Function to generate quiz questions using Groq API
 def generate_quiz_questions(text):
     try:
-        payload = {
-            "model": "llama3-8b-8192",
-            "messages": [
+        response = client.chat.completions.create(
+            messages=[
                 {
                     "role": "user",
                     "content": f"Generate quiz questions for the following text:\n\n{text}"
                 }
-            ]
-        }
-        response = groq_request(payload)
-        quiz_questions = response['choices'][0]['message']['content'].strip()
+            ],
+            model="llama3-8b-8192",
+        )
+        quiz_questions = response.choices[0].message.content.strip()
         return quiz_questions
     except Exception as e:
         return f"Error generating quiz questions: {e}"
@@ -119,17 +116,16 @@ def parse_quiz_questions(quiz_text):
 # Function to generate explanation for quiz answers using Groq API
 def generate_explanation(question, correct_answer, user_answer):
     try:
-        payload = {
-            "model": "llama3-8b-8192",
-            "messages": [
+        response = client.chat.completions.create(
+            messages=[
                 {
                     "role": "user",
                     "content": f"Explain why the correct answer to the following question is '{correct_answer}' and not '{user_answer}':\n\n{question}"
                 }
-            ]
-        }
-        response = groq_request(payload)
-        explanation = response['choices'][0]['message']['content'].strip()
+            ],
+            model="llama3-8b-8192",
+        )
+        explanation = response.choices[0].message.content.strip()
         return explanation
     except Exception as e:
         return f"Error generating explanation: {e}"