Cyna: sorry, I don't see any question about display data (if there was one I didn't get pinged - and I only saw your messages now because amCap1712 pinged me)
2019-06-19 17040, 2019
Cyna
Ohh
2019-06-19 17027, 2019
Cyna
I was wondering if I had to build the display data that is passed to the edit form or would it be done automatically
2019-06-19 17048, 2019
Cyna
When editing edit details it was done by Perl
2019-06-19 17024, 2019
Nyanko-sensei joined the channel
2019-06-19 17029, 2019
ferbncode
spellew: That looks good as well (sorry for the extremely late response). I'll make sure to review your pull request for CB musicbrainz_db module today o/
2019-06-19 17046, 2019
Nyanko-sensei has quit
2019-06-19 17047, 2019
D4RK-PH0ENiX has quit
2019-06-19 17059, 2019
bitmap
right, the server will do it for any existing edit (that's already been entered) if you're viewing it - but if you mean for an "in progress" edit, you'll have to build that yourself in JS
2019-06-19 17059, 2019
bitmap
since the server doesn't know about the edit at all (you haven't submitted it yet), all the pending data is in JS
2019-06-19 17022, 2019
D4RK-PH0ENiX joined the channel
2019-06-19 17051, 2019
Cyna
Okay... I'll try building up the display data in .js. will ping you at night
2019-06-19 17005, 2019
Lotheric_ joined the channel
2019-06-19 17012, 2019
Lotheric has quit
2019-06-19 17003, 2019
amCap1712
thanks bitmap
2019-06-19 17028, 2019
amCap1712
is it possible that a barcode may be associated with 2 releases in mb?
2019-06-19 17041, 2019
reosarevok
amCap1712: sure. I've seen two different CDs be given the same barcode (usually because people are stupid)
2019-06-19 17030, 2019
CatQuest
welcome ot brilliant discs
2019-06-19 17044, 2019
CatQuest
two unrelated thing have the same valid barcode
2019-06-19 17054, 2019
CatQuest
also a release cna have more thna one barcode
2019-06-19 17012, 2019
CatQuest
and individula mediums cna have a barcode each (which we canna store so it goes to the release isntead)
2019-06-19 17018, 2019
CatQuest
(we shoudl fix that at some point)
2019-06-19 17016, 2019
amCap1712
so in that case, when a barcode lookup happens. first a list of releases should be displayed that are associated with it and the user should be allowed from them.
2019-06-19 17028, 2019
reosarevok
amCap1712: I think it's rare enough it might make sense to go automatically to the only result by default, and only show a selection form if we happen to find multiple hits
2019-06-19 17059, 2019
amCap1712
reosarevok: yeah thats a good idea
2019-06-19 17019, 2019
amCap1712
reosarevok: can you share any example of barcode associated with 2releases so that i can test the app
2019-06-19 17049, 2019
reosarevok
Not from the top of my head, but I'll make a query
2019-06-19 17000, 2019
reosarevok
Unless CatQuest remembers one example by heart
2019-06-19 17036, 2019
CatQuest
[10:02] <reosarevok> amCap1712: I think it's rare enough it might make sense to go automatically to the only result by default, and only show a selection form if we happen to find multiple hits <--
2019-06-19 17048, 2019
CatQuest
ohi have noe actulaly. let me find it ..
2019-06-19 17015, 2019
gr0uch0mars joined the channel
2019-06-19 17032, 2019
CatQuest
wait patiently i've almost got it (had to dig a bit becasue i lsot thel ink, got it soon)
i have begun work on integrating barcode scanner but i just found that last release of zbar wasin 2012
2019-06-19 17031, 2019
amCap1712
should i go on with it?
2019-06-19 17037, 2019
gr0uch0mars
let's go with it, it's our best choice so far
2019-06-19 17001, 2019
gr0uch0mars
but hide zbar implementation behind an abstraction (another `BarcodeScan` class, or something like it)
2019-06-19 17016, 2019
gr0uch0mars
so we can change it “easily” in the future
2019-06-19 17006, 2019
gr0uch0mars
1. Create an abstract class `BarcodeScan` with the general methods that should be available publicly.
2019-06-19 17006, 2019
gr0uch0mars
2. Then create a `BarcodeScanWithZbar` class that extends that abstract, and implement those methods using Zbar
2019-06-19 17006, 2019
gr0uch0mars
3. And in the app, use `BarcodeScanWithZbar`, which can be easily replaceable in the future as the methods and results are defined in the abstract class (`BarcodeScan`), not the zbar implementation (`BarcodeScanWithZbar`)
[listenbrainz-server] paramsingh merged pull request #564 (master…spotify-reader-typeerror): Make spotify reader parse listens in a less restrictive way https://github.com/metabrainz/listenbrainz-server…
pristine__: can you open a PR with your code for this as well?
2019-06-19 17021, 2019
pristine__
i will in some time
2019-06-19 17032, 2019
pristine__
ruaok: do you see less of Pink Floyd?
2019-06-19 17021, 2019
pristine__
I have eliminated all the artists from top artists with similar artists count = 0
2019-06-19 17046, 2019
ruaok
k
2019-06-19 17045, 2019
ruaok
wow, that is much improved. I think this is the first one that I really wish I could play!
2019-06-19 17024, 2019
pristine__
The reason why we had too much of pink Floyd or any other artist was because there were songs of only 20 top artists from which CF could pick
2019-06-19 17029, 2019
ruaok
there is still a lot of pink floyd, but it is much better.
2019-06-19 17049, 2019
iliekcomputers
i like it better, all of it is artists I have listened to before, that is intentional?
2019-06-19 17020, 2019
pristine__
There are 2 playlists iliekcomputers
2019-06-19 17028, 2019
pristine__
For each user
2019-06-19 17037, 2019
iliekcomputers
oh whoops
2019-06-19 17039, 2019
iliekcomputers dum
2019-06-19 17043, 2019
pristine__
Scroll and check the other one
2019-06-19 17020, 2019
pristine__
ruaok: these html files have top artists limit set to 50, when I set to 100 recordings are more evenly distributed
2019-06-19 17039, 2019
pristine__
Because CF has more options to choose from.
2019-06-19 17030, 2019
ruaok
please keep notes on how these variable changes affect the recommendations.
2019-06-19 17050, 2019
pristine__
yeah, I have :)
2019-06-19 17018, 2019
pristine__
Also, with around 100 top artists, after eliminating artists with similar artists count = 0, around 40 are left, which means for more than 50% artists we don't have similar artists.
2019-06-19 17048, 2019
akhilesh
Hi Mr_Monkey! I am at market now, I will ping you, if any problem will occur, I will complete relationship lookup today. Please help to add more tests.
2019-06-19 17049, 2019
ruaok
gah, we really need to make these playlists playable!
2019-06-19 17019, 2019
ruaok
this is exciting!
2019-06-19 17024, 2019
pristine__
And one more observation that I would like to share
2019-06-19 17047, 2019
pristine__
If we have less number of top artists, we have too many repetitions in playlist 1 (like Pink Floyd) so we shall increase the number.
On increasing top artists count we have more number of similar artists, and therefore playlist 2 seems a bit off, because it may include weird, new artists names.
2019-06-19 17054, 2019
pristine__
There has to be a balance between the two.
2019-06-19 17050, 2019
pristine__
I will share the html with you, in which top artists count = 100, meaning more similar artists, (10 similar artists for each artists), you will notice that playlist 1 is even and good, but playlist 2 seems a bit off.
