13:10 PM
ruaok
no
13:11 PM
(yes, should not be removed)
13:23 PM
pristine___: hang on.
13:23 PM
where do you need to deal with this? I sense things might be more complicated than they need to be.
13:23 PM
pristine___
ruaok: so if & comes in name as well, and in join phrase too,
13:23 PM
Okay
13:23 PM
reosarevok
You'll have plenty of join phrases that are only punctuation FWIW
13:24 PM
pristine___
ruaok: Converting artist and track names to matchable
13:24 PM
The 4 steps you told me that day
13:24 PM
Remove whitespace, punctuation, accent and convert to lower
13:24 PM
ruaok
yes, clearly. it sounds like you are constructing an artist_credit_name from its components. is that right?
13:27 PM
pristine___
I will have a df of listens, I will convert the artist_name and track _name to Matchable texts so that I can join listens with the mapping on these Matchable fields.
13:27 PM
ruaok
ok, where in that process do you need to construct an artist credit name?
13:30 PM
travis-ci joined the channel
13:30 PM
travis-ci
13:30 PM
travis-ci has left the channel
13:31 PM
pristine___
Isn't artist credit name is same as artist name in listens?
13:31 PM
ruaok
yes and there are no join phrases in listens, which is why I am confused as to what you're doing.
13:32 PM
pristine___
13:33 PM
we are joining here on msid, we want to join on matchable texts, no?
13:33 PM
ruaok
correct.
13:33 PM
pristine___
13:33 PM
here too.
13:34 PM
ruaok
ok, when I say "constructing an artist credit name", I meant constructing a text from these three tables
13:34 PM
13:34 PM
pristine___
so I need to convert the `artist_name` and `track_name` col of listens to matchable (remove whitespace, punctuations etc)
13:35 PM
ruaok
those three tables do not exist in the LB listens.
13:35 PM
so, ignore artist credits. simply carry out the 4 steps on the artist_name and track_name fields without worrying about join phrases.
13:36 PM
pristine___
so that i can join like this: `listens_df.artist_name_matchable == 'msb_artist_credit_name_matchable'
13:36 PM
yeah, so that was what i was asking, should i need to keep join phrases while removing whitespaces
13:36 PM
okay
13:54 PM
13:55 PM
13:55 PM
ruaok: ^^
13:55 PM
The labs API returned 500
14:03 PM
ruaok
pristine___: what is your exact query?
14:03 PM
14:05 PM
travis-ci joined the channel
14:05 PM
travis-ci
14:05 PM
travis-ci has left the channel
14:07 PM
kori has quit
14:08 PM
kori joined the channel
14:16 PM
v6lur has quit
14:17 PM
pristine___
ruaok: yup, got the issue. Opening a PR.
15:03 PM
_lucifer
15:03 PM
if these changes are ok, i'll add the same for other entities as well
15:05 PM
alastairp
that looks perfect. exactly the series of tests that are needed
15:06 PM
reosarevok
Whoo
15:06 PM
Tests
15:06 PM
That I didn't need to write
15:06 PM
The best kind of tests
15:08 PM
_lucifer
👍
15:08 PM
alastairp
_lucifer: so the only reason that we were getting the weird behaviour was because the mbid was cached?
15:09 PM
now if we use the wrong entity it'll return 404?
15:09 PM
_lucifer
yes
15:09 PM
alastairp
right
15:09 PM
sneaky bug
15:10 PM
_lucifer
i saw that the sqlalchemy logs were not fetching the artist so i came to probe the cache and after adding entity to the mbid it went away
15:10 PM
so i think this should be the only issue
15:13 PM
alastairp
pristine___: btw, "API didn't send a valid responce due to Internal Server Error." -> should be response
15:14 PM
pristine___
yeah, thanks
15:14 PM
I corrected this in the PR i am writing rn :p
15:14 PM
alastairp
cool
15:33 PM
BrainzGit
15:33 PM
BrainzBot
15:38 PM
reosarevok
Whoops :D
15:38 PM
I did wonder for a moment why we even had that there if it was useless, but :D
15:38 PM
bitmap
sshhh
15:39 PM
actually, I'll add a comment there too saying when it's NaN, because it already confused me once
15:40 PM
reosarevok
Probably smart
15:46 PM
bitmap: this is added by you and currently set to decision required. What do we need to decide, other than "yes/no"?
15:46 PM
15:46 PM
BrainzBot
MBS-2704: Return release labels/catalog numbers in /ws/2 release group request
15:49 PM
reosarevok
15:49 PM
BrainzBot
MBS-2962: Enable MB server to run under proxy.
15:49 PM
reosarevok
The comment suggested "There is some remaining work here updating a few of the javascript files to not use absolute URLs."
15:54 PM
bitmap
that is going to be all broken if MBS isn't running at /
15:56 PM
though I'm not sure how well that worked before
15:56 PM
we didn't port uri_for to JS anyway, so any React code will assume / is the server root
15:58 PM
reosarevok
Ok, so basically that's a lot further away than before rather than the other way around
15:58 PM
Good to know. I'll leave a comment
15:59 PM
bitmap
re: MBS-2704 I'm not sure, but probably whether it makes sense to return the info there
15:59 PM
BrainzBot
16:00 PM
reosarevok
Ok
16:00 PM
Maybe we can decide :)
16:00 PM
I understand we don't want to add even more stuff to ws responses - this won't usually be *a lot* per release but...
16:00 PM
bitmap
tbh I think the inc=releases thing (which only returns a max of 25 iirc?) should be deprecated anyway
16:01 PM
the proper way is via a browse request, which if it doesn't return that info we should add it there, but
16:01 PM
reosarevok
And just use browse only?
16:01 PM
Well, that's certainly the way we should do it for ws3 at the very least
16:01 PM
bitmap
yeah
16:02 PM
reosarevok
I *assume* a browse shows this, but let me check
16:02 PM
bitmap
if it does I wouldn't bother improving inc=releases tbh
16:02 PM
reosarevok
It does with inc=labels, yes
16:03 PM
Well, it's your ticket, you close it ;p
16:03 PM
bitmap
fair enough :)
16:11 PM
re: MBS-1575 we could change it to /remove while still supporting /delete
16:11 PM
BrainzBot
16:11 PM
reosarevok
Just detach delete to remove?
16:12 PM
bitmap
yeah
16:12 PM
or s/detach/forward/
16:16 PM
reosarevok
Guess we might as well
16:19 PM
SothoTalKer_
the documentation does not mention that you can inc multiple values
16:19 PM
and it also does not mention which seperator shall be used then
16:19 PM
reosarevok
The examples do, but I'll take a look
16:19 PM
SothoTalKer_
and " " is a strange delimiter, when used in urls
16:20 PM
i'd love to have maybe "," as additional one
16:21 PM
and the URL it spits out when the query is incorrect is outdated aswell
16:22 PM
reosarevok
Is it? give me an incorrect query so I can check and maybe fix it, please :)
16:25 PM
SothoTalKer_
16:25 PM
it is luckily a redirect
16:25 PM
v6lur joined the channel
16:26 PM
but nowhere it is mentioned that you can use multiple inc statements in a single query ^^
16:26 PM
ah, there:
16:26 PM
To include more than one subquery in a single request, separate the arguments to inc= with a + (plus sign), like inc=recordings+labels.
16:26 PM
but it says +
16:26 PM
not space
16:29 PM
bitmap
in www-form-urlencoded data + is interchangeable with %20
16:35 PM
BrainzGit
16:35 PM
BrainzBot
16:35 PM
reosarevok
Oh, the mmd link has always been a redirect
16:35 PM
I've always used +, I didn't know that you could use space, even
16:36 PM
SothoTalKer_
ok, its all settled then ^^
16:39 PM
reosarevok
bitmap: heh, all our entities have /delete, not just release
16:40 PM
bitmap
oh, really?
16:40 PM
maybe I was thinking of collections
16:40 PM
something used add/remove
16:42 PM
reosarevok
cdtoc and entities-from-collection, yes
16:42 PM
I guess we also use "Remove whatever" for other entities, but
16:43 PM
bitmap
do we use 'Delete whatever' anywhere still?
16:44 PM
reosarevok
Delete user and delete account only
16:44 PM
It seems
16:44 PM
bitmap
hmm. well, not sure this is a real issue
16:45 PM
unless we want to change everything to /remove
16:45 PM
I suppose we could do that and change the user/account strings, but
16:46 PM
reosarevok
I think I might just close it
16:47 PM
bitmap
fine with that
16:50 PM
thomasross joined the channel
16:51 PM
diru1100
CatQuest: Android 11 has this new feature
16:51 PM
I think you might like it
16:51 PM
It's more an Easter egg 🥚 than a feature ig
16:52 PM
17:07 PM
BrainzGit
17:16 PM
ruaok pays for the summit shirts
17:17 PM
shivam-kapila
Yay
17:19 PM
ruaok
at least you'll get *one* t-shirt that way, shivam-kapila :p
17:22 PM
kori has quit
17:26 PM
kori joined the channel
17:26 PM
kori has quit
17:26 PM
kori joined the channel