j-555/wedinos-scraper — reverse-engineered prompt

Build me a small Python tool that collects all public drug sample results from the WEDINOS sample results website and saves them as clean JSON files for harm reduction work.

It should open the site in a real browser because the pages load with JavaScript, go through every results page, pull out each sample reference, date received, purchase intent, lab findings, postcode area, sample type, colour, use method, reported effects, and any reason a sample was not analysed. It should avoid duplicates if I run it more than once.

Please also add simple flags for important warnings. If the lab result does not match what the person thought they bought, mark it as an alert. Also mark alerts when especially risky substances show up, like nitazenes, fentanyl type drugs, novel benzos, veterinary sedatives, or synthetic cannabinoids.

Save the full dataset, a separate alerts file, a summary file, logs, and checkpoints so a long scrape can resume later. Include command line options for testing a few pages, resuming, changing output folder, slowing the scrape, and showing the browser for debugging.