[06:38:33] I see some gmane.org email address subscribed to a mailing list switched to gmane-mx.org. [06:39:50] (Which currently doesn't have an MX record, AFAICT.) [07:46:09] Nemo_bis: that's expected [07:46:32] https://lars.ingebrigtsen.no/2020/01/06/whatever-happened-to-news-gmane-org/ [07:51:35] Ah ok, I was left to http://home.gmane.org/ [07:53:00] right [15:20:35] zhuyifei1999_: around ? [15:20:35] Can you please have a look at https://phabricator.wikimedia.org/T234783 and give your feedback/comments on `Implementaion of commons notifications/emails`? [15:24:53] * zhuyifei1999_ looking [15:25:50] This might not be as user-friendly as the users might not know how much progress did the video encoding processed. <= no percentages? [15:35:03] zhuyifei1999_: So our plan is to notify users after video computations are done? [15:35:36] I would make it optional toggleable [15:36:06] Oh good idea [15:36:46] the way v2c does it is with giving precentages. people are more patient if they can estimate the ETA ;) [15:37:46] Something like this https://phabricator.wikimedia.org/T232457 ? [15:39:36] Do you have any resources to know the video computation status of ffmpeg? [15:43:04] Gopa: old and ugly code: https://github.com/toolforge/video2commons/blob/master/video2commons/backend/encode/transcodejob.py#L467 [15:43:31] * Gopa looking [15:46:56] Also, Currently we are authenticating users with Wikimedia Commons, If the user is logged in VideoCutTool How to implement notifing users after completion of video encoding ? [15:50:00] zhuyifei1999_: Your dirrect comments in both the tasks T234783 & T232457 would be much appreciated :) [15:50:01] T232457: Implement progress bar for video encodings in tool. - https://phabricator.wikimedia.org/T232457 [15:50:01] T234783: Notify users after computing & uploading videos to commons. - https://phabricator.wikimedia.org/T234783 [15:54:23] I can co-mentor if that is a good idea, will only check if the implementation works, not whether a functionality should or shouldn't be implemented or the design. I think that's your choices ;) [15:58:11] zhuyifei1999_: woow thats so cool, thanks for mentoring along with me :) [16:00:13] k [16:00:49] zhuyifei1999_: If you would like to add as a mentor in Google Code-in, Please add your name in mentors list over https://www.mediawiki.org/wiki/Google_Code-in/2019 [16:01:06] I am already in :P [16:01:33] wonderful :) [16:02:25] Ahh sorry, just found you :D [16:03:52] how do I add myself to co-mentor list? [16:07:31] zhuyifei1999_: I will add you as a mentor to both of these tasks in GCI website soon :) [16:07:50] ok thanks [16:08:44] zhuyifei1999_: Done :) [16:09:49] zhuyifei1999_: Can I publish these two tasks or desc needs improvement ? [16:10:46] (sorry, later) [16:12:51] zhuyifei1999_: no worries, thanks :) [16:22:57] Gopa: Currently, users have to wait until this process completes, <= can't the user just leave? I mean, they don't have to literally wait for it right? [16:23:28] also, how do you send the email? [16:24:37] currently users cant close the tab and just leave the session. [16:25:18] The back-end might still executes but the users is not notified in that case. [16:25:51] how does the backend notify the user if the tab is kept open? [16:26:28] if oauth with sendemail to self, it will probably require the sendemail grant. or you might send from a 'bot account' to the user [16:26:48] >how do you send the email? [16:26:48] planning to send notification as they are logged in video-cut-tool? [16:27:34] 00:25:51 how does the backend notify the user if the tab is kept open? <= or, what happens if the user re-opens the tab? [16:27:55] After video encodings complete in back-end, We can see in the video-cut-tool front-end. [16:28:25] Giving a try migt give your clear idea https://tools.wmflabs.org/video-cut-tool/ [16:28:47] try with the videos having shorter length/size :) [16:28:51] what i meant was, with what process are you going to either obtain the email address yourself and send a mail or somehow ask someone else (mediawiki) with the email addresses to send the mail [16:31:52] the the method in which the server notifies the client is via returning an XHR? [16:31:58] sending notification email throgh mediawiki mail would be good go, but no idea how to request such permissions. [16:32:09] (how does the connection not time out?) [16:33:14] so yeah that was what I was asking. if you send from the user's account to itself the oauth will need 'sendemail' grant. or you can send from a bot account [16:33:34] Ahh there are many things to care of, I need to plan more for this. (currently I have no ideas on this) [16:34:16] oh if the latter case make sure to authenticate with oauth. I don't want random people filling out my user name and ticking the box to spam my inbox :P [16:35:01] oh and if you don't have authentication already and needs to add authentication... I'm afraid that might be too much for GCI honestly [16:35:13] agreed :) [16:36:07] zhuyifei1999_: Authentication is already done, did you had a look https://tools.wmflabs.org/video-cut-tool/ anb played around? [16:36:29] yeah [16:36:54] that's where I discovered 00:31:53 the the method in which the server notifies the client is via returning an XHR? 00:32:10 (how does the connection not time out?) [16:38:08] Ooh [16:38:31] so umm... larger and longer videos would definitely time out your XHR. say, like some of those hour-long videos [16:38:45] me testing with https://commons.wikimedia.org/wiki/File:1874_Pierre_Jules_C%C3%A9sar_Janssen_-_Passage_de_Venus.webm will be good go [16:39:38] zhuyifei1999_: yes we had a issues with the longer videos, T237995 [16:39:39] T237995: Sometimes VideoCutTool is failing for the larger videos - https://phabricator.wikimedia.org/T237995 [16:39:44] usually encoding speed is slower than playing, after a certain resolution), so hour long video becomes hours of encoding [16:39:54] you can't keep the XHR alive for so long [16:40:12] ok CC'ed [16:40:21] so what might be the good way ro resolve this issues ? [16:40:46] yeah I was testing with https://commons.wikimedia.org/wiki/File:All_Around.webm. it's less than a minute :) [16:41:00] I would suggest socket.io [16:41:37] you are using nodejs right? [16:41:46] yes [16:42:01] then socket.io would fit right in :P [16:42:12] it would also help with progress bar [16:42:24] since progress bar is server -> client push of data [16:42:48] Ahh I never used this before, yeah I will find more about it soon. [16:42:53] v2c has a seperate nodejs component to handle socket.io [16:43:18] https://github.com/toolforge/video2commons/blob/master/video2commons-socketio/index.js [16:43:24] * Gopa looking [16:43:46] it's inefficient since the rest of server code is python [16:44:18] but you are using nodejs anyways so no such overhead for you [16:44:43] yes that migh be simple to implement in this case. [16:46:35] the way v2c handles progress is that the real backend (where ffmpeg runs) stores the status into redis, which generates an event that nodejs subscribes to. nodejs then queries the python api for the status and send the status to client via socket.io [16:48:02] it's more complicated because there are many hosts that run ffmpeg, which shares data on redis. if the webservice starts ffmpeg directly you can skip this redis part [16:48:10] Ooh that so niec pipelined process. [16:48:18] which shares data on redis <= referring to the hosts [16:49:42] well, yeah, too complicated and painful to work with. I have yet to migrate it to python 3. probably after the 16th when I get back to US :P [16:51:09] anyways hear this node script works fine I guess. [16:51:41] uh, don't think I understand [16:52:01] you mean your current node script works fine? [16:53:37] Nope, [16:53:37] >I have yet to migrate it to python 3. probably after the 16th when I get back to US :P [16:53:37] here I ment, It makes you more tough to migrate again. [16:54:31] oh yeah [16:54:40] lol [16:54:48] :) [16:55:28] zhuyifei1999_: Are we moving this out of GCI then ? [16:56:06] too complex? [16:56:10] the notification is too much. authentication might be too much for GCI [16:56:46] the progress bar... a competent GCI student might be able to do it [16:56:51] yes I also think soo :) [16:57:20] but will need like a week probably? and probably lots of mentoring [16:57:38] i wouldn't say notification is too much...? would just need the oauth scope changed, wouldn't it? assuming you mean email notification [17:00:15] zhuyifei1999_: Are you ready to mentor ? [17:00:38] IAmNetx: authtication doesn't seem to be implemented yet [17:00:59] Gopa: sure [17:01:06] on videocuttool? [17:01:08] zhuyifei1999_: which auth do you mean ? [17:01:41] oh wait [17:02:04] when I was presented with the form, I thought the tool was auth-free [17:02:28] No, Auth is already in live. [17:02:41] ok then if you have auth then mail should be doable [17:02:54] Only authenticated users can re-upload vydeos back to commons. [17:03:36] zhuyifei1999_: yes but IAmNetx `oauth scope changed` to ? [17:04:17] zhuyifei1999_: Did you try logging in video-cut-tool ? [17:04:34] IAmNetx: what do you mean by oauth scope change ? [17:04:49] change the scope to allow the application to access the user's email (or private information, forgot what it's called) and then use that for sending notifications if they closed the tab? [17:04:56] I think sending from role / bot account is probably easier than changing oauth grants. the latter might require submitting a few consumer [17:05:15] oh you can access private? forgot that [17:05:16] IAmNetx: :) [17:05:37] Gopa: yeah just logged in [17:06:09] private is the `privateinfo` grant [17:06:23] *a few => a new [17:06:49] zhuyifei1999_: Aah so lets notify by bot acoount then ? [17:07:56] to me, submitting a new consumer and having the privateinfo grant added seems trivial in comparison to creating a bot account and sending mail through that [17:09:23] I would be more reluctant to approve the consumer if I were the oauth-manager-whatever-they-are-called, and be mre reluctant to give the tool the grant if I am using the tool [17:10:08] oh and... https://wikitech.wikimedia.org/wiki/Wikitech:Cloud_Services_Terms_of_use#If_my_tools_collect_Private_Information... [17:11:13] could be a minefield [17:11:51] or you might get the sendemail grant make the user send to themselves. it's how google does it I think [17:12:34] like when you use a google tool and tell it to email you something, the sender of the email is yourself [17:13:02] Aah just like a reminder themselfs :) [17:14:37] lol [17:15:16] sendemail grant [17:15:30] * Gopa by mistake [17:15:56] * zhuyifei1999_ brb [17:36:28] * zhuyifei1999_ back [17:41:34] * Gopa zhuyifei1999 waves [18:15:23] Hi [18:15:46] Hi is there a way to query https://en.wikipedia.org/wiki/Special:Log through API or MySQL directly? [18:16:06] https://en.wikipedia.org/w/api.php?action=help&modules=query%2Blogevents [18:16:20] And yes, there's a logging table in mysql [18:29:41] Hey all, is it possible to get raw html title from mwapi? [18:30:00] For example this page has an italic text in the title: https://en.wikipedia.org/wiki/Homo_sapiens [18:31:10] mateusbs17: yes, it's called "displaytitle" [18:31:52] MatmaRex: would that be the following URL? https://en.wikipedia.org/w/api.php?action=parse&prop=displaytitle&page=Homo_Sapiens [18:32:12] https://en.wikipedia.org/wiki/Special:ApiSandbox#action=parse&format=json&page=Homo%20sapiens&prop=displaytitle [18:32:13] yes [18:32:32] i was just looking that up :) [18:32:40] mateusbs17: you have a typo – uppercase "S" [18:32:48] mateusbs17: so your query refers to a different page [18:33:01] it should be like this: https://en.wikipedia.org/w/api.php?action=parse&prop=displaytitle&page=Homo_sapiens [18:33:16] * mateusbs17 🤦‍♂️ [18:33:20] MatmaRex: Thanks! [18:33:29] :D