[07:55:32] [[Tech]]; ArchiverBot; Bot: Archiving 1 thread (older than 30 days) to [[Tech/Archives/2017]].; https://meta.wikimedia.org/w/index.php?diff=16600874&oldid=16579076&rcid=9444990 [15:33:33] Does anibody knows about how labs --> google search/api works exactly? I mean, is there any special contract between wikimedia and google about the google api usage? [18:27:53] Steinsplitter, are you looking for https://wikitech.wikimedia.org/wiki/Nova_Resource:Google-api-proxy , or, for details beyond that? [18:28:55] quiddity: this one does not allow to acess he image search engine (as far i now), which is not part of the regular api. i am wondering if we have some limit expection for the /reverse image search. [18:30:47] Steinsplitter, ah, the commons specific version. I'm not sure. (Do you have any links handy, for details of what you've found so far?) [18:32:38] quiddity: for example https://www.google.com/searchbyimage?image_url=https%3A%2F%2Fupload.wikimedia.org%2Fwikipedia%2Fcommons%2Fthumb%2F0%2F0c%2FL%25C3%25B6rrach_-_Burghof_L%25C3%25B6rrach_-_Abendansicht5.jpg%2F300px-L%25C3%25B6rrach_-_Burghof_L%25C3%25B6rrach_-_Abendansicht5.jpg (this function is not included in google api, but parsing is ratelimit as far [18:32:38] i know. not tested on labs) [18:47:15] Steinsplitter, Hmmm, Are you talking about the old (~2011) gadget on Commons? (i.e. https://commons.wikimedia.org/wiki/MediaWiki:Gadget-GoogleImages.js ) I don't think that uses Labs, nor any specific contract, at all? [18:47:58] beyond that, I only know of https://phabricator.wikimedia.org/T145165 (and linked tasks, which you're already subscribed to), and I can't see any recent updates in those. [18:48:57] i have a script to find copyvios, but not sure if allowed to run on labs. Thus i asked if wmf<>google havin some sort of contract for free google api/tools use. [18:49:55] ah! Not sure. I suggest asking kaldari during the week. [18:51:31] Steinsplitter: See https://wikitech.wikimedia.org/wiki/Nova_Resource:Google-api-proxy/Documentation [18:52:18] Steinsplitter: It's not exactly free though. Google gives us credits to cover the cost, but the credits aren't guaranteed forever, so it's kind of a tenuous service. [18:55:09] But that's strictly for API use, which it sounds like might not be what you're looking for. [18:55:15] kaldari: thanks, it is problematic if i send ca. 100 requests per day at https://www.google.com/searchbyimage?image_url= ? As far i can see i can't send requests to the /searchbyimage using the api-proxy (the function is not part of the api itself) [18:56:32] Steinsplitter: That's correct. Google isn't currently offering any reverse image search by API. [18:57:58] If you mean literally ~100/day, I would assume Google would barely notice that? I don't know what levels would start getting problematic though... ~100/minute? [18:58:23] Steinsplitter: As far as your proposed use of the search service directly, I can't really respond to that for legal reasons (Google Term of Use and all that). [18:59:43] (oh, right, disclaimers: I am not a dev, I am not a lawyer, I have no relevant experience with these questions. etc.) [19:00:17] * quiddity denies everything, and goes back to cleaning his apt. [19:01:09] plausible deniability [19:02:07] thanks to both of you, i try to write google a mail ... tos says approval is needed for automated queryes (or do we have a contac person at the WMF?). I spent hours to search a equivalent api/function - does not exist so far :/ [20:21:56] It's rather easy to get "blocked" with www.google.com screenscraping, but there are simple workarounds.