youtube2text_web.py

import streamlit as st
from youtube_transcript_api import YouTubeTranscriptApi
from youtube_transcript_api.formatters import TextFormatter
from yt_dlp import YoutubeDL
import re
import g4f


# Include the Vazirmatn font via CDN
st.markdown(
    """
    <link href="https://cdn.jsdelivr.net/gh/rastikerdar/vazirmatn@v33.003/Vazirmatn-font-face.css" rel="stylesheet" type="text/css" />
    """,
    unsafe_allow_html=True
)

st.title('🎥 YouTube Transcript Extractor')

transcript_extracted = False

def extract_video_id(video_url: str) -> str:
    pattern = r'(https?://)?(www\.)?(youtube|youtu|youtube-nocookie)\.(com|be)/(watch\?v=|embed/|v/|.+\?v=)?([^&=%\?]{11})'
    match = re.search(pattern, video_url)
    if match is None:
        st.error("Invalid YouTube URL. Please enter a valid YouTube video URL.")
        return None
    return match.group(6)

def fetch_transcript(video_id: str) -> str:
    try:
        transcript = YouTubeTranscriptApi.get_transcript(video_id)
        formatter = TextFormatter()
        formatted_transcript = formatter.format_transcript(transcript)
        return formatted_transcript
    except Exception as e:
        if "No transcript found" in str(e):
            st.error("No transcript found for this video.")
        else:
            st.error(f"Failed to fetch transcript: {e}")
        return None

def translate_to_persian(video_id: str) -> str:
    try:
        transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
        transcript = transcript_list.find_transcript(['en'])
        translated_transcript = transcript.translate('fa').fetch()
        formatter = TextFormatter()
        formatted_translated_transcript = formatter.format_transcript(translated_transcript)
        return formatted_translated_transcript
    except Exception as e:
        st.error(f"Failed to translate transcript: {e}")
        return None

def get_video_info(video_url: str) -> tuple:
    opts = dict()
    with YoutubeDL(opts) as yt:
        info = yt.extract_info(video_url, download=False)
        title = info.get("title", "")
        description = info.get("description", "")
        thumbnails = info.get("thumbnails", [])
        thumbnail_url = thumbnails[-1]["url"] if thumbnails else None
        return title, description, thumbnail_url

def structure_with_ai(transcript_text: str) -> str:

    request = f''' 
    We are seeking a professional transformation of a YouTube video transcript into a structured and teaching-oriented format, utilizing Markdown for its readability and structure. The transcript content to be transformed is as follows:

    ```
    {transcript_text}
    ```

    Please ensure the transformation adheres to a teaching tone, enhancing clarity and engagement for the audience. The final output should be neatly formatted in Markdown, preserving the original content's meaning while enhancing its presentation for educational purposes.

    Your assistance in this matter is greatly appreciated.
              '''
    response = g4f.ChatCompletion.create(
        model= "",
        provider=g4f.Provider.Koala,
        messages=[
            {"role": "user", "content": request}
        ]
    )
    messages = "".join(response)
    
    return messages

video_url_input = st.text_input("Enter YouTube Video URL", "")
structure_with_ai_checkbox = st.checkbox("✨ Structure to Markdown format with AI")
translate_to_persian_checkbox = st.checkbox("🌏 Extract Persian Transcript")
submit_button = st.button("Extract Transcript")

if submit_button and video_url_input:
    video_id = extract_video_id(video_url_input)

    if video_id:
        try:
            video_title, video_description, video_thumbnail = get_video_info(video_url_input)
            video_url = f"https://www.youtube.com/watch?v={video_id}"

            with st.spinner('Fetching transcript...'):
                transcript_text = fetch_transcript(video_id)
                if transcript_text:
                    transcript_extracted = True

            if translate_to_persian_checkbox and transcript_text:
                with st.spinner('Translating to Persian...'):
                    translated_text = translate_to_persian(video_id)
                    st.markdown(f'''
                    <div dir="rtl" style="text-align: right; font-family: 'Vazirmatn', sans-serif;">
                        <p style="margin-bottom: 0.5rem;"> </p>
                        <article class="media">
                            <p>{translated_text}</p>
                        </article>
                    </div>
                    ''', unsafe_allow_html=True)

            if structure_with_ai_checkbox and transcript_text:
                with st.spinner('Structuring Using Gemini...'):
                    structured_transcript = structure_with_ai(transcript_text)
                    st.markdown(structured_transcript)
            else:
                if video_title: st.markdown(f"## {video_title}")
                st.markdown(f" {transcript_text} ")
            
            with st.sidebar:
                st.markdown(f"## Video Information: ")
                if video_title: st.markdown(f"### [{video_title}]({video_url})")
                if video_thumbnail: st.image(video_thumbnail, use_column_width=True)
                if video_description: st.markdown(f" {video_description}")

        except Exception as e:
            st.error(f"An error occurred: {e}")
    else:
        st.error("Please enter a valid YouTube video URL.")