scrape-heal: self‑healing web scraper that auto‑repairs broken selectors
scrape-heal monitors each scraper run and validates the extracted data against a baseline of a previous successful run. When a selector no longer returns the expected shape, it locates the known values on the live page and generates new selectors, then re‑runs the scraper to verify the fix before applying it. This hands‑off approach saves developers from manual debugging of broken scrapers and ensures continuous data quality. It stands out by coupling detection, automatic healing, and verification in a single CLI tool.
View on GitHub →Swastikbhat-lab/scrape-heal