Screenshots and PDFs

When text isn't enough

Every lesson so far has extracted text: a title, a price, a link. But sometimes you need a visual artifact instead: a screenshot to prove what an agent actually saw at a given moment (useful for debugging a scraper that suddenly breaks), or a PDF as a shareable, print-style record of a page. Playwright can produce both directly, no separate screenshot tool or PDF library needed.

The code, piece by piece

OUTPUT_DIR = Path(__file__).parent / "output"

Path(__file__) is this very script's own location on disk; .parent is the folder containing it. Building paths this way means the lesson always writes its output next to itself, no matter what folder you happened to run the command from.

page.screenshot(path=viewport_path)

Captures exactly what's visible in the browser's viewport (the visible window area) at that moment, nothing that would require scrolling to see. This is the fast, default option.

page.screenshot(path=full_page_path, full_page=True)

full_page=True instead scrolls through the entire page and stitches everything into one tall image, capturing content below the fold too. It costs a bit more time, since Playwright has to render and capture the whole page rather than just the current viewport.

page.pdf(path=pdf_path, format="A4")

Generates a PDF the same way a browser's own "Print to PDF" feature would. This only works in headless Chromium (no visible window, which is the default when you call p.chromium.launch() without headless=False), and it's Chromium-specific, not every browser engine implements PDF export. format="A4" sets the paper size, the same setting you'd see in a real print dialog.