Fork me on GitHub

Wikimedia IRC logs browser - #gsoc-catimages

Filter:
Start date
End date

Displaying 320 items:

2016-07-08 12:21:44 <jayvdb> hey ;-)
2016-07-08 12:22:55 <jayvdb> have you looked at the video copyright code yet?
2016-07-08 12:28:16 <AbdealiJK> Hi DrTrigon DrTrigon_ jayvdb
2016-07-08 12:28:30 <jayvdb> perfect timing ;-)
2016-07-08 12:28:34 <DrTrigon_> Hi!
2016-07-08 12:28:46 <AbdealiJK> :D
2016-07-08 12:28:54 <AbdealiJK> I didn't have time today to do it on the recent files though :(
2016-07-08 12:29:23 <AbdealiJK> Had to go out suddenly and just came back. I wil do it in another 2-3 hrs and post results for that in conpherence
2016-07-08 12:30:24 <jayvdb> ok
2016-07-08 12:30:24 <DrTrigon_> about time; the other meeting today is at a bad timing for me - any change that we find a replacement?
2016-07-08 12:31:06 <jayvdb> I havent heard back from Ty. If he can make it, I'll talk with him, and then organise another meeting with us all
2016-07-08 12:31:26 <AbdealiJK> Tomorrow same time or day after same time is good for me.
2016-07-08 12:31:39 <DrTrigon_> that would be perfect - if not possible ping me and I'll try to follow...
2016-07-08 12:31:41 <AbdealiJK> I'm not available on Mon-Wed at that time though
2016-07-08 12:32:18 <jayvdb> ok. AbdealiJK is that time today good for you?
2016-07-08 12:32:24 <DrTrigon_> can we share a google calendar with availability? or a doodle?
2016-07-08 12:32:24 <AbdealiJK> Yes
2016-07-08 12:32:26 <jayvdb> ok
2016-07-08 12:32:59 <DrTrigon_> if this is of any use?
2016-07-08 12:33:02 <jayvdb> If Ty cant make it in a few hours, a Doodle would be good
2016-07-08 12:33:19 <DrTrigon_> perfect - so let's see what happens
2016-07-08 12:33:32 <DrTrigon_> are you fine guys? do we want to start?
2016-07-08 12:33:40 <jayvdb> ok, do we intend to skype now?
2016-07-08 12:34:06 <AbdealiJK> I think IRC is fine - not much to discuss this week
2016-07-08 12:34:14 <AbdealiJK> But anythings good for me
2016-07-08 12:34:15 <DrTrigon_> no oppinion
2016-07-08 12:34:41 <AbdealiJK> jayvdb, Which would you prefer ?
2016-07-08 12:35:56 <jayvdb> maybe try with IRC first and then a quick skype for anything that we want to chat about
2016-07-08 12:36:08 <AbdealiJK> Alrighto
2016-07-08 12:36:15 <DrTrigon_> alrighty
2016-07-08 12:36:32 <AbdealiJK> So, regarding updates - I looked at EXIF data a bit more to see if we can use any more of it
2016-07-08 12:37:14 <AbdealiJK> I found that we can detect whether an image is a screenshot using one bit of information (If "gnome-screenshot" was in some EXIF data it was a screenshot from GNOME's default tool)
2016-07-08 12:37:40 <AbdealiJK> I tried checking other tools too - cheese, digikam, imagemagick, etc. But none of them left any trace on the exif data
2016-07-08 12:38:13 <jayvdb> very useful
2016-07-08 12:38:28 <jayvdb> screenshots can be copyrighted, and thus need to be deleted
2016-07-08 12:38:29 <AbdealiJK> I then also tried to add more softwares but found popular things like matplotlib, octave, matlab, etc don't mess with the exif data :(
2016-07-08 12:38:54 <AbdealiJK> jayvdb, Nod. Plus there's a Category:Screenshot where we can move it to get the right people looking at it
2016-07-08 12:39:02 <jayvdb> yup
2016-07-08 12:39:08 <DrTrigon_> jayvdb: > 3000 files in cat screenshots
2016-07-08 12:40:17 <DrTrigon_> AbdealiJK: but you have some detection for matlab, right?
2016-07-08 12:40:23 <jayvdb> I see "Number of grey shades used: 246" - I guess that can be used to determine if a image is b&w only
2016-07-08 12:41:04 <AbdealiJK> DrTrigon_, Right you are. Sorry - matlab was written wrongly in that message
2016-07-08 12:41:16 <DrTrigon_> :)
2016-07-08 12:41:26 <AbdealiJK> jayvdb, Is there a reason to detect BnW images ? I found that the Category didnt have all sorts of BnW images
2016-07-08 12:41:37 <jayvdb> there is a separate category for browns only , like https://commons.wikimedia.org/w/index.php?oldid=198123175
2016-07-08 12:41:38 <DrTrigon_> yea it's a pitty not all software use the exif tags ...
2016-07-08 12:42:11 <jayvdb> https://commons.wikimedia.org/wiki/Category:Black_and_white_photographs
2016-07-08 12:42:28 <jayvdb> and other https://commons.wikimedia.org/wiki/Category:Monochrome_photographs
2016-07-08 12:43:03 <AbdealiJK> Ahh, ok. I was just looking at Category:Black and white
2016-07-08 12:43:12 <jayvdb> your tool should be able to process all of the files in https://commons.wikimedia.org/wiki/Category:Monochrome_photographs , and move them into a subcategory
2016-07-08 12:43:49 <DrTrigon_> (straight forward...)
2016-07-08 12:44:08 <AbdealiJK> Not very sure if it is straight forward - but yes will look into it
2016-07-08 12:44:14 <jayvdb> hehe
2016-07-08 12:44:18 <AbdealiJK> Especially the sepia color checking
2016-07-08 12:44:18 <DrTrigon_> ;))
2016-07-08 12:44:41 <DrTrigon_> You don't have to do ALL subcats PERFECT
2016-07-08 12:44:47 <AbdealiJK> Cool, will do that :+1:
2016-07-08 12:44:59 <AbdealiJK> DrTrigon_, nod
2016-07-08 12:45:24 <AbdealiJK> I think converting RGB -> HSV and then thresholding the Hue should work ...
2016-07-08 12:45:44 <DrTrigon_> or peak detection?
2016-07-08 12:45:47 <AbdealiJK> will try it out later
2016-07-08 12:46:03 <DrTrigon_> may be also with summing/integrating?
2016-07-08 12:46:03 <AbdealiJK> Peak detection to find whether an image is sepia color ?
2016-07-08 12:46:35 <jayvdb> regarding models, it would be useful to publish a list of makes/models that you cant categorise, because the category doesnt exist
2016-07-08 12:46:52 <jayvdb> someone probably needs to create a category, or create a redirect to an existing category
2016-07-08 12:47:50 <AbdealiJK> What do you mean by publish a list ?
2016-07-08 12:47:51 <DrTrigon_> jayvdb: yes important point! good note!
2016-07-08 12:48:03 <DrTrigon_> e.g. if there is a color cat missing...
2016-07-08 12:48:19 <DrTrigon_> ...like you have red, green but blue is missing
2016-07-08 12:48:55 <DrTrigon_> jayvdb: can we just use/add thes cats?
2016-07-08 12:49:12 <DrTrigon_> not too extensively of course...
2016-07-08 12:49:34 <jayvdb> yes, possibly, but I am worried about junk EXIF data resulting in junk categories.
2016-07-08 12:50:07 <AbdealiJK> jayvdb, I think DrTrigon_ meant can we add it manually when we find it ourselves rather than making a list
2016-07-08 12:50:14 <jayvdb> maybe have a threshold - if the bot has seen the same make/model 3 times, add the category to the file page even if the category does not exist
2016-07-08 12:50:14 <DrTrigon_> yes, we should only do that in very restricted cases
2016-07-08 12:50:49 <AbdealiJK> Nooo, the bot should never add it.
2016-07-08 12:50:54 <DrTrigon_> AbdealiJK: some of them are not very disputed - but some will be - thus a list beforehand might be useful
2016-07-08 12:50:57 <jayvdb> then it is easy for a human reviewer to create the category
2016-07-08 12:51:24 <AbdealiJK> Ah - right. If there's a human reviewer then it's fine.
2016-07-08 12:51:26 <jayvdb> AbdealiJK, the bot shouldnt create the category, but it can add `[[Category:Foo]]` even if the category doesnt ecist
2016-07-08 12:51:27 <DrTrigon_> the bot checks whether it exists, and if start to populate
2016-07-08 12:51:50 <jayvdb> *exist
2016-07-08 12:51:51 <DrTrigon_> jayvdb: that's what I consider "creating"
2016-07-08 12:51:58 <DrTrigon_> ... ;))
2016-07-08 12:52:15 <AbdealiJK> Just a note - even if the category doesnt exist, the bot does print it (just without the [[ ]] for the link)
2016-07-08 12:52:20 <jayvdb> ya, it does implicitly create a category
2016-07-08 12:52:48 <AbdealiJK> ok, got it
2016-07-08 12:52:54 <jayvdb> AbdealiJK, but to reach 5% (or another goal %), you need to modify the file pages
2016-07-08 12:53:19 <jayvdb> so linking to categories, even if they do not exist, is still categorisation
2016-07-08 12:53:38 <AbdealiJK> ok
2016-07-08 12:53:54 <DrTrigon_> jayvdb: which is good?!?
2016-07-08 12:53:59 <jayvdb> yup
2016-07-08 12:54:29 <DrTrigon_> AbdealiJK: about "Camer/Scanner" ...
2016-07-08 12:54:38 <AbdealiJK> DrTrigon_, yes ?
2016-07-08 12:54:53 <DrTrigon_> ... I think you can try to suggest categories ...
2016-07-08 12:55:11 <DrTrigon_> ... again be bit careful and do only obvious cases ...
2016-07-08 12:55:21 <DrTrigon_> ... that is better than nothing, right?
2016-07-08 12:55:31 <AbdealiJK> I am suggesting categories right ?
2016-07-08 12:55:34 <jayvdb> DrTrigon , that is what I was talking about above. The bot can add categories to the page, which is a suggestion
2016-07-08 12:55:35 <AbdealiJK> I don't think I understand
2016-07-08 12:55:48 <DrTrigon_> "Because of this, it cannot "suggest" categories that do not exist yet"
2016-07-08 12:55:55 <AbdealiJK> Ahhh
2016-07-08 12:55:56 <DrTrigon_> (may be I got it wrong?)
2016-07-08 12:56:13 <AbdealiJK> ok. Got what you meant
2016-07-08 12:56:21 <jayvdb> we call them 'red categories'
2016-07-08 12:56:32 <AbdealiJK> Based on the above discussion that it's ok to add red-categories, I will add it.
2016-07-08 12:56:42 <DrTrigon_> cool! ;)
2016-07-08 12:56:42 <jayvdb> ok
2016-07-08 12:56:48 <AbdealiJK> BTW: One issue here is that
2016-07-08 12:57:07 <AbdealiJK> I cannot know whether it should be "Taken with <make>" or "Scanned with <make>"
2016-07-08 12:57:27 <DrTrigon_> in case of doubt just "<make"
2016-07-08 12:57:32 <AbdealiJK> Hence we still cannot suggest whether it is "scanned" or "taken"
2016-07-08 12:57:58 <DrTrigon_> nice info but not the main part...
2016-07-08 12:58:14 <DrTrigon_> ...as soon as you look up the software you should know it anyways.
2016-07-08 12:58:15 <jayvdb> "Created with ..." ?
2016-07-08 12:58:16 <AbdealiJK> I think I'll just make "... with <make>" so that the human who is checking can easily move that category and decide what "..." is
2016-07-08 12:58:33 <AbdealiJK> "Created with ..." is for softwares jayvdb - Like ImageMagick, Matlab, etc.
2016-07-08 12:58:57 <AbdealiJK> "Taken with ..." is for camera photographs
2016-07-08 12:58:57 <jayvdb> a camera is 80% software these days ;-)
2016-07-08 12:59:13 <DrTrigon_> yes but a scanner or a generator?
2016-07-08 12:59:16 <DrTrigon_> ;))
2016-07-08 12:59:19 <AbdealiJK> I don't think the commons community would agree though ^
2016-07-08 12:59:38 <DrTrigon_> photographs or coders?
2016-07-08 13:00:11 <AbdealiJK> Ok, 1 moment. I'd like to pause as I find this is a very small issue
2016-07-08 13:00:25 <AbdealiJK> I've already tested it for most of the big brands - Fujifilm, Nikon, Canon, etc.
2016-07-08 13:00:27 <DrTrigon_> sure, go on!
2016-07-08 13:00:46 <AbdealiJK> And so 80% of images should be categorized correctly
2016-07-08 13:01:08 <jayvdb> ok. so not worth worrying about?
2016-07-08 13:01:20 <AbdealiJK> We can increase it later by adding *some sort of* red links. But It's a minor detail and we can just do it later
2016-07-08 13:01:36 <jayvdb> nod.
2016-07-08 13:01:47 <DrTrigon_> agree, the code needs to be ready for that ...
2016-07-08 13:01:57 <AbdealiJK> agree
2016-07-08 13:02:00 <DrTrigon_> ... but addind is some "minor" task for later.
2016-07-08 13:02:11 <jayvdb> Maybe the red-category can be "Taken/Scanned with <make>"
2016-07-08 13:02:37 <AbdealiJK> ^ interesting :)
2016-07-08 13:02:50 <jayvdb> or "Taken or scanned with <make>"
2016-07-08 13:02:54 <DrTrigon_> or we ask on village pump? ;)
2016-07-08 13:03:08 <DrTrigon_> AbdealiJK: do you want to continue with the next point?
2016-07-08 13:03:09 <AbdealiJK> I think we can just choose one. And if someone comments with a better option, we can just use cat-a-lot to move stuff easily
2016-07-08 13:03:20 <AbdealiJK> DrTrigon_, yes
2016-07-08 13:03:21 <jayvdb> nod
2016-07-08 13:03:22 <DrTrigon_> nod
2016-07-08 13:03:49 <jayvdb> now regarding OpenStreetMap vs Google Maps, we really should be using OpenStreetMap if we can
2016-07-08 13:03:51 <AbdealiJK> So, I tried creating an "official" docker file based on our last week conversation
2016-07-08 13:04:10 <DrTrigon_> nod
2016-07-08 13:04:14 <AbdealiJK> AH, ok - skipping the docker point then
2016-07-08 13:04:18 <jayvdb> how much worse is the OSM data ?
2016-07-08 13:04:24 <AbdealiJK> I tried using OSM. I prefer OSM compared to Google too
2016-07-08 13:04:25 <DrTrigon_> ???
2016-07-08 13:04:59 <AbdealiJK> So the issue is that OSM was giving "New No 4, Street ABC, Town, City, Country, PIN" as a string
2016-07-08 13:05:14 <jayvdb> and you need to parse it?
2016-07-08 13:05:26 <AbdealiJK> Can't parse it - it has no standard format
2016-07-08 13:05:45 <jayvdb> hmm. sounds like the geopy package is not helpful
2016-07-08 13:05:58 <AbdealiJK> checks the OSM API
2016-07-08 13:06:30 <jayvdb> looking for the python package I switched to recently ...
2016-07-08 13:06:44 <AbdealiJK> AH, youve tried it already - nice
2016-07-08 13:07:18 <AbdealiJK> jayvdb, But even with that OSM will still have a limit
2016-07-08 13:07:47 <jayvdb> the limit isnt the problem
2016-07-08 13:08:18 <jayvdb> wikimedia has a copy of the OSM data, so we can host our own instance of any query service if we need it
2016-07-08 13:08:33 <AbdealiJK> :o nice
2016-07-08 13:08:34 <DrTrigon_> ahhhh!
2016-07-08 13:08:41 <jayvdb> (and may already have a copy of the Nominatim service running somewhere on tools)
2016-07-08 13:08:56 <AbdealiJK> Wait ! so is OSM a part of wikimedia ?
2016-07-08 13:09:00 <AbdealiJK> Or is this a collaboration ?
2016-07-08 13:09:40 <DrTrigon_> not part ..., right?
2016-07-08 13:09:47 <jayvdb> https://github.com/damianbraun/nominatim
2016-07-08 13:09:56 <jayvdb> https://pypi.python.org/pypi/nominatim
2016-07-08 13:10:10 <jayvdb> this is a collaboration
2016-07-08 13:10:39 <jayvdb> Wikimedia uses OSM data a lot, so we set up our own database and tile server
2016-07-08 13:10:48 <AbdealiJK> I can use `nominatim` +1
2016-07-08 13:10:52 <DrTrigon_> I knew once that wiki draws a lot of data from OSM... :)
2016-07-08 13:10:54 <AbdealiJK> I see jayvdb
2016-07-08 13:11:00 <DrTrigon_> cool!
2016-07-08 13:11:09 <jayvdb> and replicate the data into our local clone fairly regularly
2016-07-08 13:12:47 <jayvdb> Wikimedia also has a license for MaxMind GeoIP, but I have never used that, so I dont know how useful it is. And I am not sure whether Wikimedia Foundation will be able to share their license for this purpose (probably, but not 100%)
2016-07-08 13:13:20 <DrTrigon_> AbdealiJK: briefly comming back to docker - sorry for the time! - do you have anything you can give me in order that I can experiment arond a bit?
2016-07-08 13:13:37 <AbdealiJK> jayvdb, Is there any IRC channel I can ask aboout the OSM + wikimedia stuff ?
2016-07-08 13:13:50 <jayvdb> I know the MaxMind license is basically 'unlimited'
2016-07-08 13:14:22 <DrTrigon_> jayvdb: IF we can avoid using commercial services it would be nice ...
2016-07-08 13:14:35 <DrTrigon_> ...of course only if we don't loose quality
2016-07-08 13:14:40 <jayvdb> https://wikitech.wikimedia.org/wiki/OSM_Tileserver
2016-07-08 13:15:08 <jayvdb> https://lists.wikimedia.org/pipermail/maps-l/
2016-07-08 13:15:26 <AbdealiJK> DrTrigon_ I'm updating my latest dockerfiles
2016-07-08 13:15:48 <jayvdb> https://www.mediawiki.org/wiki/Maps
2016-07-08 13:15:55 <DrTrigon_> just put it somewhere and drop me a link
2016-07-08 13:16:02 <jayvdb> https://maps.wikimedia.org/#4/40.75/-73.96
2016-07-08 13:16:41 <AbdealiJK> DrTrigon_, in https://github.com/AbdealiJK/file-metadata/tree/ajk/docker you can see Dockerfile.ubuntu - that would be what you'd be interested in
2016-07-08 13:16:55 <AbdealiJK> There's also Dockerfile.centos if you prefer that
2016-07-08 13:17:14 <DrTrigon_> whatever you think works better - can use anything inside the VM
2016-07-08 13:17:56 <jayvdb> https://tools.wmflabs.org/locator-tool/index.html#/
2016-07-08 13:18:00 <AbdealiJK> I think Ubuntu would be easier.
2016-07-08 13:18:08 <DrTrigon_> ok
2016-07-08 13:18:16 <AbdealiJK> maps.wikimedia.org is really cool jayvdb
2016-07-08 13:20:06 <DrTrigon_> agree - thanks for that!
2016-07-08 13:20:23 <DrTrigon_> open questions here? do we want to go through the meeting agenda?
2016-07-08 13:21:28 <DrTrigon_> AbdealiJK, jayvdb: ^^^
2016-07-08 13:21:30 <jayvdb> sure
2016-07-08 13:21:47 <AbdealiJK> yes please
2016-07-08 13:22:36 <DrTrigon_> so AbdaliJK as I understand we get "categorization hints" (could not check the code yesterday...), right?
2016-07-08 13:23:28 <AbdealiJK> Yes, that was done last week if I'm not mistaken
2016-07-08 13:23:51 <DrTrigon_> I thought there were minor thing open - even better! :)
2016-07-08 13:24:14 <DrTrigon_> then I was thinking about the to define the goal ...
2016-07-08 13:24:33 <DrTrigon_> ... would it be possible to start gathering and outputing stats now?
2016-07-08 13:24:48 <DrTrigon_> in a nice table / overview
2016-07-08 13:25:05 <AbdealiJK> DrTrigon_ what sort of stats are you thinking about ?
2016-07-08 13:25:35 <DrTrigon_> e.g. when you do a run for https://commons.wikimedia.org/wiki/User:AbdealiJKTravis/logs/Category_Male_faces ...
2016-07-08 13:26:00 <AbdealiJK> I find automatically adding it is irritating - because the script gets a bit messy
2016-07-08 13:26:25 <DrTrigon_> ... can you output a table at the end summarizing e.g. 1000 processed, 100 categorized, 10 issues, like ...
2016-07-08 13:26:27 <AbdealiJK> Stats need structured data, and right now that is just dumping wiki lines
2016-07-08 13:26:38 <DrTrigon_> something you did in the beginning but as table
2016-07-08 13:27:01 <AbdealiJK> I can do "1000 processed" and "100 categorized"
2016-07-08 13:27:05 <DrTrigon_> automatically doe not need to be ...
2016-07-08 13:27:09 <AbdealiJK> "10 issues" cannot be done easily right ?
2016-07-08 13:27:21 <DrTrigon_> ... may be in a seperate script ...?
2016-07-08 13:28:02 <AbdealiJK> DrTrigon_ I tried that in the beginning too. But it seemed unnecessary ... I would need to parse the wikicode that the first script writes and again process data, etc.
2016-07-08 13:28:04 <DrTrigon_> by "issues" I mean "uncertain categorization"
2016-07-08 13:28:12 <AbdealiJK> I found it was just easier to do manually
2016-07-08 13:28:42 <DrTrigon_> it not about we need that in the bot - it's about to get an idea how it performs overall
2016-07-08 13:29:42 <jayvdb> or maybe a table, with number of categories added per file. i.e. 5+ cats added: 10 images ; 4+ cats added: 50 images ; 3+ cats added: 100 images ...
2016-07-08 13:30:20 <AbdealiJK> that can be done
2016-07-08 13:30:30 <DrTrigon_> we need a tool to keep control of the bot quality
2016-07-08 13:30:41 <AbdealiJK> I think uncertain categories can be done too
2016-07-08 13:30:50 <DrTrigon_> and that has 2 parts; how good the bot finds data from files ...
2016-07-08 13:31:07 <DrTrigon_> ... and how good it can relate this info to categories on commons
2016-07-08 13:31:56 <jayvdb> right. we dont want a bot adding "Black and white" and "photograph" when it can be adding only "Black and white photographs"
2016-07-08 13:32:16 <jayvdb> the most valuable categorisation is leaf categories in the category tree
2016-07-08 13:33:55 <jayvdb> perhaps that is a way to decide on quality categories : the bot must add three categories that are leaf categories
2016-07-08 13:34:01 <jayvdb> ^ just an example
2016-07-08 13:34:07 <DrTrigon_> so we need to find a way to measure perfomance of (A) alogrithms like face detection, etc. and (B) relating that info to commons categorization
2016-07-08 13:34:35 <jayvdb> if it does that, for 5% of the images, the tool is very useful
2016-07-08 13:35:28 <DrTrigon_> of categorized images, right?
2016-07-08 13:35:47 <DrTrigon_> (that could be a goal... correct)
2016-07-08 13:36:03 <jayvdb> 'three leaf categories' is probably too hard. maybe only 'one leaf category, and two non-leaf categories'
2016-07-08 13:36:48 <jayvdb> ok, I need to go soon
2016-07-08 13:37:08 <DrTrigon_> I might be go for 1 leaf cat in 5% as MVP and 3 leaf cat in 5% as bulls eye
2016-07-08 13:37:16 <DrTrigon_> :)
2016-07-08 13:37:29 <jayvdb> nod
2016-07-08 13:37:43 <DrTrigon_> AbdealiJK: what do you think about measuring performance?
2016-07-08 13:37:55 <AbdealiJK> 3 leaf is probably not very feasible. Currently, the only analysis which gives us a leaf is Scanned with/Created with/Taken with
2016-07-08 13:38:17 <DrTrigon_> and the black whit photographs, right?
2016-07-08 13:38:20 <jayvdb> geo should give you a leaf cat
2016-07-08 13:38:23 <AbdealiJK> Nope
2016-07-08 13:38:33 <AbdealiJK> jayvdb, Not really, the place name can be of any granuality ..
2016-07-08 13:39:00 <AbdealiJK> For example, I can detect that it was in India, Chennai. But maybe not that it was in IIT Madras (my college)
2016-07-08 13:39:03 <DrTrigon_> AbdealiJK: sry, the monochromes
2016-07-08 13:39:08 <jayvdb> geo should give you a leaf cat , or red leafcat
2016-07-08 13:39:29 <jayvdb> ok . going now. ill be around in 15 mins
2016-07-08 13:40:17 <AbdealiJK> DrTrigon_, https://commons.wikimedia.org/wiki/Category:Black_and_white_photographs has lots of subcats
2016-07-08 13:40:42 <DrTrigon_> ;) the "Nope" was not for me...
2016-07-08 13:40:57 <DrTrigon_> so do you think measuring perfomance is feasble?
2016-07-08 13:41:44 <DrTrigon_> jayvdb: bye - see ya!
2016-07-08 13:42:04 <AbdealiJK> DrTrigon_ I am not sure how it can be done actually ...
2016-07-08 13:42:04 <DrTrigon_> AbdealiJK: so do you think measuring perfomance is feasble?
2016-07-08 13:42:39 <AbdealiJK> ^ The measuring performance
2016-07-08 13:43:40 <DrTrigon_> may be first step is to summerize info per bot run - the we can think about merging all thoses together
2016-07-08 13:44:29 <AbdealiJK> nod. I will try creating a summary report based on what jayvdb mentioned (Number f images with 3 cats, etc)
2016-07-08 13:44:35 <AbdealiJK> Is there any other stat ?
2016-07-08 13:45:17 <DrTrigon_> I would include anything you want to learn about or does not work as you expect ...
2016-07-08 13:45:34 <DrTrigon_> ... you mentioned face-detection to be worse than you expected, etc.
2016-07-08 13:46:06 <AbdealiJK> In the stats ?
2016-07-08 13:47:04 <DrTrigon_> they are mostly for us developers, right? we want to know what parts of the code can or need to be improved...
2016-07-08 13:47:33 <DrTrigon_> also I would ouput info like 5 sure categorizations, 3 experimental ones (relates to the face-detection etc.)
2016-07-08 13:47:47 <AbdealiJK> Ok, so I think "Number of distinct categories"
2016-07-08 13:47:52 <DrTrigon_> ...respective to the 2 run modes
2016-07-08 13:47:58 <AbdealiJK> "Number of files for each category"
2016-07-08 13:48:21 <AbdealiJK> But we currently don't have a distinction for experimental vs sure - so, I'd like to procrastinate that if possible
2016-07-08 13:48:52 <DrTrigon_> that is ok - but keep it in mind, for design etc.
2016-07-08 13:49:02 <AbdealiJK> yep
2016-07-08 13:49:25 <DrTrigon_> also it should be clear why they are distinct or categorized etc
2016-07-08 13:49:57 <AbdealiJK> I didnt follow your last message, could you elaborate ?
2016-07-08 13:50:30 <DrTrigon_> if you do stats like "Number of distinct categories" we also need to know WHY they are distinct e.g.
2016-07-08 13:50:48 <DrTrigon_> what result / algorithm made the bot deciding like that
2016-07-08 13:51:09 <DrTrigon_> (was it due to facedetection or anything else etc.)
2016-07-08 13:51:32 <AbdealiJK> Apologies. Let me clarify what I meant by "number of distinct categories"
2016-07-08 13:51:55 <DrTrigon_> please :)
2016-07-08 13:52:04 <AbdealiJK> I meant overall in the whole batch, how many different categories were used
2016-07-08 13:53:20 <DrTrigon_> I see.
2016-07-08 13:54:55 <AbdealiJK> DrTrigon_ ok. think we can call this to a close ? Let me begin with the current stats we've mentioned and see if we need more over time ?
2016-07-08 13:55:14 <AbdealiJK> And add appropriate stats as needed/feasible
2016-07-08 13:55:59 <DrTrigon_> Yes let's do that. Once we see what we are talking off it will get more clear anyway. :)
2016-07-08 13:56:08 <AbdealiJK> nod
2016-07-08 13:56:13 <DrTrigon_> (clearer?)
2016-07-08 13:56:14 <jayvdb> back
2016-07-08 13:56:24 <DrTrigon_> nice!
2016-07-08 13:56:35 <AbdealiJK> We were just planning on ending :)
2016-07-08 13:56:37 <DrTrigon_> so AbdealiJK: do you want to add anything?
2016-07-08 13:56:44 <AbdealiJK> Nothing from me - no
2016-07-08 13:57:10 <DrTrigon_> jayvdb: anything to add? questions?
2016-07-08 13:57:29 <jayvdb> list of distinct categories suggested (and then 'added' when the bot writes) would be useful
2016-07-08 13:57:35 <AbdealiJK> Minor update: I've begun running the script on new images. Will update with results when done
2016-07-08 13:57:55 <jayvdb> nice
2016-07-08 13:58:23 <jayvdb> ill be around until late tonight, so post an update when you have some useful % stats
2016-07-08 13:58:27 <DrTrigon_> AbdealiJK: wanted to add the Bulk test, Upstream and Software results are nice!
2016-07-08 13:58:56 <DrTrigon_> ...and I got a very strange error: http://dpaste.com/3JA5RK3
2016-07-08 13:59:05 <DrTrigon_> that's all from my side!
2016-07-08 13:59:22 <AbdealiJK> DrTrigon_ I answered that in the conpherence I believe
2016-07-08 13:59:25 <jayvdb> thats all from me
2016-07-08 13:59:46 <DrTrigon_> did not refresh conpherence yet...
2016-07-08 14:00:17 <DrTrigon_> you wrote it as pwb.py already... I see... will check! thanks!
2016-07-08 14:00:42 <DrTrigon_> so thanks and have a nice evening folks!
2016-07-08 15:02:35 <TyLandercasper> Is everyone up?
2016-07-08 15:09:46 <jayvdb> hi
2016-07-08 15:10:01 <TyLandercasper> hey
2016-07-08 15:10:30 <jayvdb> ping DrTrigon DrTrigon2 ;-)
2016-07-08 15:10:47 <DrTrigon> jayvdb: Yeessss? :)
2016-07-08 15:10:55 <jayvdb> AbdealiJK just left this room; i've pinged him on Skype
2016-07-08 15:11:07 <DrTrigon> You wanna meet now?
2016-07-08 15:11:22 <jayvdb> if you are free, TyLandercasper is here and on Skype
2016-07-08 15:11:44 <DrTrigon> I kind of about to leave... let me check...
2016-07-08 15:12:25 <DrTrigon> ok, have about 1.5 hrs - let's go... ;)
2016-07-08 15:16:59 <AbdealiJK> TyLandercasper, Hi !
2016-07-08 15:17:15 <TyLandercasper> hello!
2016-07-08 15:17:30 <AbdealiJK> Is there a meeting going on ? Or is it postponed ?
2016-07-08 15:17:43 <TyLandercasper> they're tying to add you right now
2016-07-08 15:34:47 <DrTrigon> I fell out the call...
2016-07-08 15:34:56 <jayvdb> me too
2016-07-08 16:05:33 <jayvdb> i think ivelost everyone
2016-07-08 16:33:17 <DrTrigon> AbdealiJK: https://phabricator.wikimedia.org/Z441 running `$ python bulk.py -search:'eth-bib' -logname:ETH-Bib` currently ;)
2016-07-09 17:02:34 <jayvdb> AbdealiJK, https://meta.wikimedia.org/wiki/WikiConference_India_2016/Scholarships#Decision
2016-07-09 17:03:11 <jayvdb> have you received an email?

This page is generated from SQL logs, you can also download static txt files from here