Feature Requests item #1993062, was opened at 2008-06-13 16:47 Message generated for change (Comment added) made by multichill You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603141&aid=1993062&group_id=93107
Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: interwiki Group: None Status: Open Priority: 7 Private: No Submitted By: Melancholie (melancholie) Assigned to: Nobody/Anonymous (nobody) Summary: Use API module 'parse' for retrieving interwiki links Initial Comment: Currently pages are retrieved in a batch by using Special:Export. Although being fast (as only one request is done), there is a huge data overhead with this method! Why not use the API with its 'parse' module? Only interwiki links can be fetched with that, reducing traffic (overhead) a lot! See: http://de.wikipedia.org/w/api.php?action=parse&format=xml&page=Test&prop=langlinks Outputs could be downloaded in parallel to virtualize a batch (faster). ---- At least make this method optional (config.py) for being able of reducing data traffic, if wanted. API is just more efficient. ---------------------------------------------------------------------- >Comment By: Multichill (multichill) Date: 2008-11-13 12:46 Message: We are working on a rewrite. The rewrite uses the api as much as possible. ---------------------------------------------------------------------- Comment By: Melancholie (melancholie) Date: 2008-06-15 01:27 Message: Logged In: YES user_id=2089773 Originator: YES See http://meta.wikimedia.org/wiki/Interwiki_bot_access_protocol concerning disambiguations and redirects: http://de.wikipedia.org/w/api.php?action=parse&format=xml&text={{:Main_Page}}{{:Bot}}&prop=langlinks|templates ---------------------------------------------------------------------- Comment By: Melancholie (melancholie) Date: 2008-06-14 16:38 Message: Logged In: YES user_id=2089773 Originator: YES Backwards compatibility? That's no reason for not making software more efficient, where possible ;-) That's also why I wrote something about "optional", too. Because for current MediaWiki wikis there is a much more efficient way of retrieving (only) certain contents (langlinks, categories), there should be a method of using that advantage! Will reduce load (bot owner's and server's)... ---------------------------------------------------------------------- Comment By: Bryan (btongminh) Date: 2008-06-13 20:44 Message: Logged In: YES user_id=1806226 Originator: NO Backwards compatibility with non Wikimedia wikis? ---------------------------------------------------------------------- Comment By: Melancholie (melancholie) Date: 2008-06-13 17:20 Message: Logged In: YES user_id=2089773 Originator: YES For not being misusable of confusing bots, the yet to be set up MediaWiki message could contain [[foreigncode:{{CURRENTTIMESTAMP}}]] (cache issue?) (sorry for spamming with this request ;-) ---------------------------------------------------------------------- Comment By: Melancholie (melancholie) Date: 2008-06-13 17:08 Message: Logged In: YES user_id=2089773 Originator: YES Important note for getting pages' interwikis in a batch: http://de.wikipedia.org/w/api.php?action=parse&text={{:Test}}{{:Bot}}{{:Haus}}&prop=langlinks Either the bot could figure out what interwikis belong together then, or maybe a marker could placed in between: http://de.wikipedia.org/w/api.php?action=parse&text={{:Test}}{{MediaWiki:Iwmarker}}{{:Bot}}{{MediaWiki:Iwmarker}}{{:Haus}}&prop=langlinks [[MediaWiki:Iwmarker]] (or 'Llmarker'?) would have to be set up by the MediaWiki developers with [[en:/de:Abuse-save-mark]] as content (but this is potentially misusable). ---------------------------------------------------------------------- Comment By: Melancholie (melancholie) Date: 2008-06-13 16:51 Message: Logged In: YES user_id=2089773 Originator: YES Note: Maybe combine it with 'generator'. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603141&aid=1993062&group_id=93107 _______________________________________________ Pywikipedia-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/pywikipedia-l
