mend3/Scrapling — reverse-engineered prompt

Build me a Python library called Scrapling for people who need reliable web scraping without fighting every website manually. I want it to handle simple one page fetches, dynamic pages that need a browser, stealthier requests for sites with antibot checks like Cloudflare Turnstile, and bigger crawls that can run many sessions at once.

The main thing is that users can select data with familiar CSS style selectors, save what they picked, and have the scraper find the same kind of elements again even if the page layout changes later. Also include a spider style crawler where someone can define start URLs and yield scraped items, with live stats, streaming results, pause and resume, and proxy rotation support.

Please make it feel polished as an installable Python package with clear examples, a small CLI, docs, and tests. Keep the user experience simple, just a few lines of Python for common scraping jobs. Look up current docs online if you need to.