BTW, because SuggestBot uses number of backlinks to penalize highly linked articles in its link traversal code, I maintain backlink counts for all main namespace pages that are not redirects for en, pt, ru, sv, no, fa, and hu wikis (I think this would cover #1 in Navino's list posted earlier). The tables are on the replicated servers (because we need to join them with the page table) and are updated once a day. If anyone would like access to any of them, get in touch with me off-list.
Regards, Morten On 9 September 2014 10:49, Tim Landscheidt <[email protected]> wrote: > Navino Evans <[email protected]> wrote: > > > That's great to know, thank you. > > > We'll make sure we only use the API within that limit - basically just > for > > individual calls when a user adds a new event to our database. > > > For the bulk processing, we would need to update the backlinks > information > > as a monthly maintenance task, so I wouldn't want to trouble you with > this > > each time. > > > Would you rather we stick with data dump processing for the large scale > > stuff? > > > [...] > > Do note that you don't have to channel all your requests > through John :-). Apparently, you have developers who could > probably set up this themselves easily, so they could just > apply for access. > > I'm not sure if there have been precedents, but I assume if > you: > > 1. don't try to have WMF subsidize your business model, > i. e. don't move heavy lifting stuff to Labs just because > you want to save some CPU time, > > 2. ensure that all tools you create and run on Labs are re- > leased as open source, and > > 3. ideally make the query results usable by others > > there is nothing wrong with running those queries yourself. > > Tim > > > _______________________________________________ > Labs-l mailing list > [email protected] > https://lists.wikimedia.org/mailman/listinfo/labs-l >
_______________________________________________ Labs-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/labs-l
