ONDLS Pharmacy Data Scraper

I want a robust Python scraper that harvests every publicly listed pharmacy record from India’s Online National Drugs Licensing System (https://statedrugs.gov.in/SFDA/Homepage). The public search pages load results through background JSON calls, so the code should hit those endpoints directly and fall back to Playwright or Selenium whenever tokens, cookies or dynamic content make that easier—whichever blend of both techniques keeps the run stable. The script must: • Iterate through every state/UT and its districts, following pagination until the very last record. • Capture at minimum the Firm Name, full Address and Phone Number for each licence entry. • Handle CSRF tokens, session cookies, time-outs and retries gracefully, resuming from the precise point of failure if the run is interrupted. • De-duplicate and gently clean obvious inconsistencies before saving. Deliverables 1. Well-commented Python code (and requirements.txt) ready to run from the command line. 2. A CSV and an .xlsx file containing the complete cleaned dataset. 3. A brief README describing setup, execution steps and how to restart a partial run. If anything in the portal changes while you are working, please surface the new request pattern in the code rather than hard-coding values so the scraper remains maintainable.