flatpickles / random_wiki Goto Github PK

random_wiki
Matt Nichols

random_wiki is a Python script which tweets the first sentence of a random Wikipedia article and a shortened link to said article several times a day. Using several rules, the script attempts to generate an aesthetic tweet (avoiding common garbage from in-process pages, uncommon formattings, and the like), and determine a reasonable amount of time to wait before tweeting again (randomly, between two and four hours).

Known issues:

- Sentence endings are determined using periods, but excepting circumstances to which any of several rules apply ("Dr.", "U.S.A.", or "Ph.D." are examples where this is necessary). However, there's no way I've catalogued all possible abbreviations, so there will inevitably be some errors here.
- The body of the article is found by the first <p> segment of the Wikipedia webpage, but occasionally a page will have an infoboxe which precedes the actual body of text in the markup, and have content which can be mistaken for the article's real beginning. The "Synonyms" section of some pages is an example. As a workaround, random_wiki will not post from pages including these infoboxes, but eventually I'll incorporate a HTML parser for even more intelligent first sentence identification.
- This readme could be more detailed.