Best way for a beginner to learn screen scraping by Python -

this might 1 of questions difficult answer, here goes:

i don't consider self programmer - :-) i've learned r, because sick , tired of spss, , because friend introduced me language - not complete stranger programming logic.

now learn python - screen scraping , text analysis, writing webapps pylons or django.

so: how should go learning screen scrape python? started going through scrappy docs feel "magic" going on - after - trying learn, not do.

on other hand: there no reason reinvent wheel, , if scrapy screen scraping django webpages, might after worth jumping straight scrapy. think?

oh - btw: kind of screen scraping: want scrape newspaper sites (i.e. complex , big) mentions of politicians etc. - means need scrape daily, incrementally , recursively - , need log results database of sorts - lead me bonus question: talking nonsql db. should learn use e.g. mongodb right away (i don't think need strong consistency), or foolish want do?

thank thoughts - , apologize if general considered programming question.

i agree scrapy docs give off impression. but, believe, found myself, if patient scrapy, , go through tutorials first, , bury rest of documentation, not start understand different parts scrapy better, appreciate why way it. framework writing spiders , screen scrappers in real sense of framework. still have learn xpath, find best learn regardless. after all, intend scrape websites, , understanding of xpath , how works going make things easier you.

once have, example, understood concept of pipelines in scrapy, able appreciate how easy sorts of stuff scrapped items, including storing them database.

beautifulsoup wonderful python library can used scrape websites. but, in contrast scrapy, not framework means. smaller projects don't have invest time in writing proper spider , have deal scrapping amount of data, can beautifulsoup. else, begin appreciate sort of things scrapy provides.