Update app.py

app.py

@@ -3,34 +3,30 @@ from youtube_transcript_api import YouTubeTranscriptApi
 import re
 import tempfile
 import os
-import warnings
-import torch
-from transformers import pipeline
-from groq.client import Client
+import openai
 
-# Set up device for torch
-device = "cuda:0" if torch.cuda.is_available() else "cpu"
+# Set up OpenAI API key
+openai.api_key = os.getenv("OPENAI_API_KEY")
 
-# 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,
-)
+# Supported file types for OpenAI API
+SUPPORTED_FILE_TYPES = ["mp3", "mp4", "mpeg", "mpga", "m4a", "wav", "webm"]
 
-# Suppress specific warning
-warnings.filterwarnings("ignore", message="FP16 is not supported on CPU; using FP32 instead")
-
-# Set up Groq client
-client = Client(api_key=os.environ.get("GROQ_API_KEY"))
-
-# Function to transcribe audio using ASR pipeline
+# Function to transcribe audio using OpenAI Whisper API
 def transcribe_audio(file_path):
-    with open(file_path, "rb") as f:
-        audio_bytes = f.read()
-    prediction = asr_pipeline(audio_bytes)["text"]
-    return prediction
+    file_extension = os.path.splitext(file_path)[1][1:]
+    if file_extension not in SUPPORTED_FILE_TYPES:
+        return f"Error: Unsupported file type '{file_extension}'. Please upload a valid file."
+
+    try:
+        with open(file_path, "rb") as file:
+            response = openai.Audio.transcribe(
+                model="whisper-1",
+                file=file,
+                response_format="json"
+            )
+        return response['text']
+    except Exception as e:
+        return f"Error during transcription: {e}"
 
 # Function to get transcript from YouTube
 def get_transcript(url):
@@ -45,36 +41,34 @@ def get_transcript(url):
     except Exception as e:
         return str(e)
 
-# Function to summarize text using Groq API
+# Function to summarize text using OpenAI API
 def summarize_text(text):
     try:
-        response = client.chat_completions.create(
+        response = openai.ChatCompletion.create(
+            model="gpt-3.5-turbo",
             messages=[
-                {
-                    "role": "user",
-                    "content": f"Summarize the following text:\n\n{text}"
-                }
+                {"role": "system", "content": "You are a helpful assistant."},
+                {"role": "user", "content": f"Summarize the following text:\n\n{text}"}
             ],
-            model="llama3-8b-8192",
+            max_tokens=100,
         )
-        summary = response['choices'][0]['message']['content'].strip()
+        summary = response.choices[0].message['content'].strip()
         return summary
     except Exception as e:
         return f"Error summarizing text: {e}"
 
-# Function to generate quiz questions using Groq API
+# Function to generate quiz questions using OpenAI API
 def generate_quiz_questions(text):
     try:
-        response = client.chat_completions.create(
+        response = openai.ChatCompletion.create(
+            model="gpt-3.5-turbo",
             messages=[
-                {
-                    "role": "user",
-                    "content": f"Generate quiz questions for the following text:\n\n{text}"
-                }
+                {"role": "system", "content": "You are a helpful assistant."},
+                {"role": "user", "content": f"Generate 10 quiz questions from the following text. Make sure the questions cover all key points and are varied in type (e.g., multiple choice, short answer). Number each question and provide the correct answer after the question number:\n\n{text}"}
             ],
-            model="llama3-8b-8192",
+            max_tokens=1000,
         )
-        quiz_questions = response['choices'][0]['message']['content'].strip()
+        quiz_questions = response.choices[0].message['content'].strip()
         return quiz_questions
     except Exception as e:
         return f"Error generating quiz questions: {e}"
@@ -82,50 +76,50 @@ def generate_quiz_questions(text):
 # Function to parse quiz questions from generated text
 def parse_quiz_questions(quiz_text):
     questions = []
-    question_blocks = quiz_text.split("\n\n")
     current_question = None
     current_choices = []
     correct_answer = None
-    
-    for block in question_blocks:
-        lines = block.strip().split("\n")
-        if lines:
-            if re.match(r'^\d+\.', lines[0]):  # This line is a question number
-                if current_question and current_choices and correct_answer:
-                    questions.append({
-                        "question": current_question,
-                        "choices": current_choices,
-                        "correct_answer": correct_answer
-                    })
-                current_question = lines[0]
-                current_choices = lines[1:5]
-                correct_answer = lines[-1].split(": ")[-1].strip() if len(lines) > 5 else None
-            else:  # This line is an answer
-                correct_answer = lines[-1].split(": ")[-1].strip()
+
+    lines = quiz_text.split("\n")
+    for line in lines:
+        line = line.strip()
+        if re.match(r'^\d+\.', line):  # This line is a question number
+            if current_question:
+                questions.append({
+                    "question": current_question,
+                    "choices": current_choices,
+                    "correct_answer": correct_answer
+                })
+            current_question = line
+            current_choices = []
+            correct_answer = None
+        elif line.startswith("Answer:"):
+            correct_answer = line.split(":", 1)[1].strip()
+        elif line:
+            current_choices.append(line)
 
     # Add the last question if it exists
-    if current_question and current_choices and correct_answer:
+    if current_question:
         questions.append({
             "question": current_question,
             "choices": current_choices,
             "correct_answer": correct_answer
         })
-    
+
     return questions
 
-# Function to generate explanation for quiz answers using Groq API
+# Function to generate explanation for quiz answers using OpenAI API
 def generate_explanation(question, correct_answer, user_answer):
     try:
-        response = client.chat_completions.create(
+        response = openai.ChatCompletion.create(
+            model="gpt-3.5-turbo",
             messages=[
-                {
-                    "role": "user",
-                    "content": f"Explain why the correct answer to the following question is '{correct_answer}' and not '{user_answer}':\n\n{question}"
-                }
+                {"role": "system", "content": "You are a helpful assistant."},
+                {"role": "user", "content": f"Explain why the correct answer to the following question is '{correct_answer}' and not '{user_answer}':\n\n{question}"}
             ],
-            model="llama3-8b-8192",
+            max_tokens=100,
         )
-        explanation = response['choices'][0]['message']['content'].strip()
+        explanation = response.choices[0].message['content'].strip()
         return explanation
     except Exception as e:
         return f"Error generating explanation: {e}"
@@ -135,7 +129,7 @@ def check_answers(questions, user_answers):
     feedback = []
     correct_count = 0
     for i, question in enumerate(questions):
-        correct_answer = question['correct_answer']
+        correct_answer = question.get('correct_answer')
         user_answer = user_answers.get(f"question_{i+1}", "")
         if user_answer == correct_answer:
             feedback.append({
@@ -158,10 +152,10 @@ def check_answers(questions, user_answers):
 
 # Function to handle uploaded files
 def handle_uploaded_file(uploaded_file):
-    with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
-        tmp_file.write(uploaded_file.read())
-        tmp_file_path = tmp_file.name
-    return tmp_file_path
+    file_path = tempfile.mktemp(suffix=os.path.splitext(uploaded_file.name)[1])
+    with open(file_path, "wb") as f:
+        f.write(uploaded_file.read())
+    return file_path
 
 # Streamlit app layout and functionality
 st.title("YouTube Transcript Quiz Generator")
@@ -182,18 +176,18 @@ if option == "YouTube URL":
                 quiz_text = generate_quiz_questions(transcript_text)
                 questions = parse_quiz_questions(quiz_text)
 
-                if not questions:
+                if len(questions) < 10:
                     st.error("No valid quiz questions could be generated.")
                 else:
                     st.session_state.summary = summary
-                    st.session_state.questions = questions
+                    st.session_state.questions = questions[:10]  # Ensure only 10 questions are taken
                     st.session_state.user_answers = {}
                     st.session_state.generated_quiz = True
             else:
                 st.error(transcript_text)
 
 if option == "Upload audio/video file":
-    uploaded_file = st.file_uploader("Choose an audio or video file", type=["mp3", "wav", "mp4", "mov"])
+    uploaded_file = st.file_uploader("Choose an audio or video file", type=SUPPORTED_FILE_TYPES)
     if uploaded_file:
         if st.button("Generate Quiz"):
             tmp_file_path = handle_uploaded_file(uploaded_file)
@@ -204,11 +198,11 @@ if option == "Upload audio/video file":
                 quiz_text = generate_quiz_questions(transcript_text)
                 questions = parse_quiz_questions(quiz_text)
 
-                if not questions:
+                if len(questions) < 10:
                     st.error("No valid quiz questions could be generated.")
                 else:
                     st.session_state.summary = summary
-                    st.session_state.questions = questions
+                    st.session_state.questions = questions[:10]  # Ensure only 10 questions are taken
                     st.session_state.user_answers = {}
                     st.session_state.generated_quiz = True
             else:
@@ -241,4 +235,4 @@ if st.session_state.generated_quiz:
                         if item['status'] == "Incorrect":
                             st.write(f"**Explanation:** {item['explanation']}")
         else:
-            st.write("Please generate the quiz first.")
+            st.write("Please generate the quiz first.")