Lab — Context-Based Extraction Tool#

Objective#

Build a unified tool that an agent can call to extract text from any source URL (PDF, Website, or YouTube Video).

Instructions#

Create a Python module exposing a single extract_context(url: str) function.

1. If URL is a PDF: Use PyMuPDF to download and extract the text.
2. If URL is a Website: Use Playwright to render JavaScript and extract the main article text via BeautifulSoup.
3. If URL is a YouTube Video: Use yt-dlp to download the auto-generated VTT subtitles and parse them into raw text.

Testing#

Provide a script that calls your function with three different URLs and prints the first 500 characters of each.