Imageye's picture
Update app.py
bc87def verified
raw
history blame
8.06 kB
import streamlit as st
from openai import OpenAI
from youtube_transcript_api import YouTubeTranscriptApi
import re
import tempfile
import os
from transformers import pipeline
import soundfile as sf
# Initialize the pipeline with the model
pipe = pipeline("automatic-speech-recognition", model="openai/whisper-small")
# Function to transcribe audio using Hugging Face Whisper
def transcribe_audio(file_path):
# Load audio file into NumPy array
audio_input, _ = sf.read(file_path)
transcription = pipe(audio_input)["text"]
return transcription
# Function to get YouTube transcript
def get_transcript(url):
try:
video_id_match = re.search(r"(?:v=|\/)([0-9A-Za-z_-]{11}).*", url)
if video_id_match:
video_id = video_id_match.group(1)
else:
return "Error: Invalid YouTube URL"
transcript = YouTubeTranscriptApi.get_transcript(video_id)
transcript_text = ' '.join([entry['text'] for entry in transcript])
return transcript_text
except Exception as e:
return str(e)
# Function to summarize text using OpenAI API
def summarize_text(client, text):
response = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": f"Summarize the following text:\n\n{text}"}
]
)
summary = response.choices[0].message.content.strip()
return summary
# Function to generate quiz questions using OpenAI API
def generate_quiz_questions(client, text):
response = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": f"Generate ten quiz questions and four multiple choice answers for each question from the following text. Mark the correct answer with an asterisk (*) at the beginning:\n\n{text}"}
]
)
quiz_questions = response.choices[0].message.content.strip()
return quiz_questions
# Function to parse quiz questions
def parse_quiz_questions(quiz_text):
questions = []
question_blocks = quiz_text.split("\n\n")
for block in question_blocks:
lines = block.strip().split("\n")
if len(lines) >= 5:
question = lines[0]
choices = [line.replace('*', '').strip() for line in lines[1:5]]
correct_answer_lines = [line for line in lines[1:5] if '*' in line]
if correct_answer_lines:
correct_answer = correct_answer_lines[0].replace('*', '').strip()
else:
correct_answer = "No correct answer provided"
questions.append({"question": question, "choices": choices, "correct_answer": correct_answer})
return questions
# Function to generate explanation using OpenAI API
def generate_explanation(client, question, correct_answer, user_answer):
prompt = f"Explain why the correct answer to the following question is '{correct_answer}' and not '{user_answer}':\n\n{question}"
response = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": prompt}
]
)
explanation = response.choices[0].message.content.strip()
return explanation
# Function to check answers and provide feedback
def check_answers(client, questions, user_answers):
feedback = []
correct_count = 0
for i, question in enumerate(questions):
correct_answer = question['correct_answer']
user_answer = user_answers.get(f"question_{i+1}", "")
if user_answer == correct_answer:
feedback.append({
"question": question['question'],
"user_answer": user_answer,
"correct_answer": correct_answer,
"status": "Correct"
})
correct_count += 1
else:
explanation = generate_explanation(client, question['question'], correct_answer, user_answer)
feedback.append({
"question": question['question'],
"user_answer": user_answer,
"correct_answer": correct_answer,
"status": "Incorrect",
"explanation": explanation
})
return feedback
# Function to handle uploaded file
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
# Streamlit UI
st.title("YouTube Transcript Quiz Generator")
st.markdown("**Instructions:** Enter your OpenAI API key and paste a YouTube link or upload a media file to generate a quiz.")
api_key = st.text_input("Enter your OpenAI API Key", type="password")
if api_key:
client = OpenAI(api_key=api_key)
option = st.selectbox("Choose input type", ("YouTube URL", "Upload audio/video file"))
if "generated_quiz" not in st.session_state:
st.session_state.generated_quiz = False
if option == "YouTube URL":
url = st.text_input("YouTube URL", value="")
if api_key and url:
if st.button("Generate Quiz"):
transcript_text = get_transcript(url)
if "Error" not in transcript_text:
summary = summarize_text(client, transcript_text)
quiz_text = generate_quiz_questions(client, transcript_text)
questions = parse_quiz_questions(quiz_text)
st.session_state.summary = summary
st.session_state.questions = questions
st.session_state.user_answers = {}
st.session_state.generated_quiz = True
if option == "Upload audio/video file":
uploaded_file = st.file_uploader("Choose an audio or video file", type=["mp3", "wav", "mp4", "mov"])
if uploaded_file and api_key:
if st.button("Generate Quiz"):
tmp_file_path = handle_uploaded_file(uploaded_file)
with st.spinner('Transcribing audio...'):
transcript_text = transcribe_audio(tmp_file_path)
os.remove(tmp_file_path)
if "Error" not in transcript_text:
summary = summarize_text(client, transcript_text)
quiz_text = generate_quiz_questions(client, transcript_text)
questions = parse_quiz_questions(quiz_text)
st.session_state.summary = summary
st.session_state.questions = questions
st.session_state.user_answers = {}
st.session_state.generated_quiz = True
if st.session_state.generated_quiz:
st.write("## Summary")
st.write(st.session_state.summary)
st.write("## Quiz Questions")
for i, question in enumerate(st.session_state.questions):
st.write(f"### Question {i+1}")
st.write(question['question'])
st.session_state.user_answers[f"question_{i+1}"] = st.radio(
label="",
options=question['choices'],
key=f"question_{i+1}"
)
if st.button("Submit Answers"):
if "questions" in st.session_state and st.session_state.questions:
with st.spinner('Processing your answers...'):
feedback = check_answers(client, st.session_state.questions, st.session_state.user_answers)
st.write("## Feedback")
for i, item in enumerate(feedback):
with st.expander(f"Question {i+1} Feedback"):
st.write(f"### {item['question']}")
st.write(f"**Your answer:** {item['user_answer']}")
st.write(f"**Correct answer:** {item['correct_answer']}")
if item['status'] == "Incorrect":
st.write(f"**Explanation:** {item['explanation']}")
else:
st.write("Please generate the quiz first.")