[05:52] Trying to convince ariel to give IA a copy of all their deleted images [05:52] >:| [05:52] haha [05:58] who be this ariel? [06:00] He's (one of? the?) head tech guys for wikimedia foundation [06:53] more or less [08:40] ah [17:47] Nemo_bis: im thinking about that wikistats project [17:47] emijrp, good [17:47] it is all php? [17:47] I have no idea [17:47] did you think of the images tarballs? [17:47] the rsync mirror should make things much faster [17:48] im thinking about adding some pie charts (mediawiki, dokuwiki, ... % comparison), langs, size, active/inactive ones, etc [17:48] emijrp, also, did you see that list of API urls and what do you think? [17:48] I'm getting hundreds of errors of all sorts [17:48] probably wget is not ok for that [17:48] apparently even functioning api.php handles are not actually able to give any output, alwayas DB problems [17:48] it's not wget [17:54] Nemo_bis: i dont know if rsync mirror allows to downnload images by date [17:54] or just all [17:55] I don't know, probably you can set some filter [17:55] or you'll need some post-processing [17:55] bt could avoid some headaches with corrupted images and so on I think [17:56] or, to download all the stuff at once very fast and then worry only about packaging [17:56] the problem is that you cant download all and then filter [17:56] commons is 18TB [17:56] you have to filter and the downlaod [17:56] maybe IA can give a machine with 18 TB for a shot while [17:57] wasn't 18TB including all wikis? [17:57] not that it changes much [17:57] if you find with underscor a fast path to archive all those images on IA, I'm sure he can find the resources [17:58] by the way, the rsync method doesnt download the .xml description [17:59] i think our script is better [18:17] hm [18:17] but WMF won't be happy if you start mass downloading all Commons [18:18] i dont care [18:18] and they dont care, dont worry [21:01] damn, missed emijrp again [21:01] I wish he'd stay longer!