[12:58:27] Hi all [12:58:44] hi [12:58:49] I send a email to the other Kamsuri and DJFF to thank them for their proposals [12:59:24] hi [12:59:58] oh that's good. [13:00:20] Meanwhile I was trying create a tool. [13:00:46] I am not sure what the Service group name should be. [13:01:21] *I was trying to create a tool. [13:01:41] the service group name is the name of the tool [13:01:57] as in tools.wmflabs.org/ [13:02:06] I put 'sibu' as the name. [13:02:20] But I got 'failed to create service group'. [13:03:08] I guess we can create a test tool during the community bonding period before the name is finalized [13:03:34] yeah sure. [13:03:49] and toolname 'sibu' unfortuately is already taken by djff https://tools.wmflabs.org/?tool=sibu [13:04:04] oh I see. [13:04:41] There was some critique on the name [13:04:48] yeah [13:04:53] and I understood the points made [13:04:59] I don't like the name either [13:05:06] so an option would also be to select a different name [13:07:13] as of now for the test tool shall we call it glam2commons? [13:07:27] that matches with video2commons also. [13:07:35] lol my tool [13:07:45] haha :) [13:08:50] how is the flickr one called? [13:08:53] flickrripper? [13:08:59] or flickr2commons? [13:09:34] flickrripper is the flicker bot of pywikibot iirc [13:09:43] flickr2commons is a tool by magnus [13:09:46] there is flickr2commons yes [13:10:02] url2commons [13:10:06] geograph2commons [13:10:25] maybe it is the most clear when we go and fit into that theme [13:11:09] but it doesn't show how this project is different from many other glam tools [13:11:35] where the preprocessing is done and the upload can be done by anybody [13:11:47] well the 2commons tools all share that theme right? [13:12:03] url2commons can be well-suited. It is short and cryspy. [13:12:15] url2commons is an existing tool [13:12:52] ok [13:15:41] zhuyifei1999_: do you see a better way to show what it exactly entails? [13:21:41] (idk) [13:32:12] I've to do quite a bit today so I'd prefer to discuss some of the things I'm needed for now [13:32:22] I think a good start is to make some structure in phabricator [13:32:52] what I was thinking of is the following: we have the main task (either your proposal or the main task) [13:32:59] below that we make 4 subtasks [13:33:23] one for each month of the project (including the evaluation [13:33:38] and below that we list all the things to do according to how they depend on other task to get a working tool [13:34:20] How do you think about that [13:34:43] hmm [13:34:49] alright [13:34:57] (not too bad) [13:35:47] don't be afraid to be direct if you don't like things/say it if you have other ideas ;) [13:36:01] I'm from the netherlands we can handle direct [13:36:15] and forgive me if I'm too direct [13:36:28] haha that was hilarious. [13:36:42] I think four tasks idea is good. [13:37:00] Since it will allow a tree like structure to the entire project. [13:37:10] yes some order to the chaos [13:38:14] I had some questions about the test tool. [13:38:30] sure [13:39:10] For the National Archief GLAM api we could give a searchstring and upload images in batches. [13:39:36] Are we going to do the same flow for other glams as well? [13:40:04] What are the assumptions about the GLAM api? [13:40:08] we've to look into some suitable GLAMs I think [13:40:16] and see how the API's look [13:40:22] its a good thing for this month to do that [13:40:39] if they have an API some search is possible [13:41:02] some structure to URLs based on an identifier or another field is another option [13:41:47] for National Archief too? [13:42:45] they have an identifier, but it's not shown on the website [13:42:54] however you can search for the specific field in the API [13:44:01] ok [13:45:33] also if we upload based on the searchstring only, then how do we give enough flexibility to the user to make his choices? [13:48:09] I don't really see the issue. Initially and if possible we do it based on a unique identifier [13:48:22] there should be a clear description how to find this identifier for a certain image [13:48:38] and we should thus make sure the identifier is indeed in the identifier field [13:48:51] if you want to do multiple uploads based on a search string [13:49:07] which is something listed for August as possible extension, and which would be nice [13:49:16] we would have the user type in a search string [13:49:29] and then pre-show the images and they have to select which to upload [13:49:35] something like that would work I think [13:50:23] This leads me to a related point: somewhere we should show the user which post-processing steps need to be done [13:50:34] sometimes categories have to be added [13:50:41] or descriptions have to be fixed [13:50:59] sometimes not all images claimed to be free by the GLAM are truly free [13:51:05] and they should check for that [13:51:49] how to check if it's free? [13:52:31] well it's abit hypothetical [13:52:52] one glam had objects photographed [13:52:56] the photos are free [13:52:59] but not all objects are [13:53:16] so they'd have to check whether the object is more than 70 years old at least [13:53:34] but the thing is: the tool is useful when some work by hand has to be done [13:53:44] that is what's preventing us from just uploading the complete collection [13:53:58] ah yes. [13:54:02] so we have to state somewhere in the user-flow which work has to be done [13:54:47] https://phabricator.wikimedia.org/T164668 the task for May, we can make subtasks for the bullet points (I'll leave that to you infobliss) [13:55:56] ok sure. [13:57:31] I had compiled a list of GLAMs. [13:57:41] yep, but it was just names right? [13:57:54] yes [13:58:00] we should locate their API's and look whether their collection can still be moved to Commons [13:58:04] I can help a bit on that one [13:58:25] great. [14:00:36] For the current tool lab tests: keep it simple, just try to upload something (using an URL) to commons as a first step [14:00:44] with a hardcoded link and everything [14:01:16] that already has some complicated parts I think, because you have to connect it to an account on Commons [14:01:34] but zhuyifei1999_ and tom29739 have more expertise on this point + I should test this out a bit myself as well [14:02:18] make sure it's a tool :P (as in with a web frontend) [14:02:41] ok [14:02:52] aah ok, well then it gets a bit more complicated already [14:02:58] ? [14:03:11] may be. but it's ok. [14:03:38] since in the long run a frontend would be necessary anyway. [14:03:55] Hi tom29739, I was just saying you 2 know more about how toollabs precisely work [14:07:55] infobliss: Anything I should be answering now? I'll go on to do some things and will read here in a few hours again [14:07:59] or maybe in between [14:08:45] Ok nothing for now. [14:08:50] ;) [14:08:58] I will make the necessary tasks in Phabricator. [14:09:04] thanks [14:09:08] yes, they will allow us to follow what you do [14:09:15] sure. [14:09:17] and I'll try to be here as often as possible as well [14:09:36] ok great. [14:09:56] and be sure to ask any questions you have or when you get stuck with something to indicate so [14:10:11] sure I will.