To solve this problem, we'll create a Python script that fetches the transcript of a YouTube video and uses OpenAI's API to generate a concise summary. The solution involves three main steps: extracting the video ID from the URL, retrieving the transcript, and generating the summary via OpenAI.

Step-by-Step Explanation

1. Extract Video ID: Parse the YouTube URL to get the unique video ID.
2. Fetch Transcript: Use the youtube-transcript-api library to retrieve the video's transcript.
3. Generate Summary: Send the transcript to OpenAI's GPT-3.5-turbo model to get a concise summary.

Prerequisites

• Install required libraries:
  pip install youtube-transcript-api openai
• Set your OpenAI API key as an environment variable:
  Linux/macOS: export OPENAI_API_KEY="your_api_key"
  Windows: set OPENAI_API_KEY="your_api_key"

Python Script

import re
import os
import openai
from youtube_transcript_api import YouTubeTranscriptApi

def get_video_id(url: str) -> str:
    """Extract video ID from YouTube URL."""
    patterns = [
        r'(?:https?://)?(?:www\.)?youtube\.com/watch\?v=([^&]+)',  # Standard URL
        r'(?:https?://)?youtu\.be/([^?]+)',  # Short URL
        r'(?:https?://)?(?:www\.)?youtube\.com/shorts/([^?]+)'  # Shorts URL
    ]
    for pattern in patterns:
        match = re.search(pattern, url)
        if match:
            return match.group(1)
    raise ValueError("Invalid YouTube URL: Could not extract video ID.")

def get_transcript(video_id: str, languages: list = ['en']) -> str:
    """Retrieve transcript for the given video ID."""
    try:
        transcript = YouTubeTranscriptApi.get_transcript(video_id, languages=languages)
        return ' '.join([entry['text'] for entry in transcript])
    except Exception as e:
        raise Exception(f"Transcript error: {str(e)}")

def generate_summary(transcript: str, api_key: str) -> str:
    """Generate summary using OpenAI's GPT-3.5-turbo."""
    openai.api_key = api_key
    try:
        response = openai.ChatCompletion.create(
            model="gpt-3.5-turbo",
            messages=[
                {"role": "system", "content": "You are a concise summarizer. Summarize the YouTube transcript, covering key points and takeaways."},
                {"role": "user", "content": f"Transcript:\n{transcript}"}
            ],
            temperature=0.3,  # Low temperature for factual accuracy
            max_tokens=500  # Adjust based on desired summary length
        )
        return response.choices[0].message['content'].strip()
    except Exception as e:
        raise Exception(f"OpenAI API error: {str(e)}")

def main():
    api_key = os.getenv('OPENAI_API_KEY')
    if not api_key:
        print("Error: OPENAI_API_KEY environment variable not set.")
        return

    url = input("Enter YouTube video URL: ")
    try:
        video_id = get_video_id(url)
        print(f"Extracted Video ID: {video_id}")

        transcript = get_transcript(video_id)
        print("Transcript fetched successfully.")

        summary = generate_summary(transcript, api_key)
        print("\nSummary:\n")
        print(summary)
    except Exception as e:
        print(f"Error: {str(e)}")

if __name__ == "__main__":
    main()

Key Features

• URL Handling: Supports standard, short, and Shorts YouTube URLs.
• Multilingual Transcripts: Defaults to English but can be extended to other languages (e.g., languages=['en', 'es']).
• Error Handling: Catches invalid URLs, missing transcripts, and API errors.
• Concise Summaries: Uses GPT-3.5-turbo with low temperature for accurate, focused summaries.

Limitations & Improvements

• Long Transcripts: For transcripts exceeding the model's context window (4096 tokens), split the transcript into chunks and summarize each, then combine the results.
• Language Support: Add user input for preferred languages to fetch transcripts in non-English languages.
• Summary Length: Allow users to adjust the max_tokens parameter for longer/shorter summaries.

This script provides a robust foundation for summarizing YouTube videos using AI, with room for customization based on specific needs.

作者声明:本文包含人工智能生成内容。