13:39 PM
x3YTjJV6asPhskrZ
i can work on a workaround if you want
13:40 PM
samj1912
you can describe the algorithm
13:40 PM
arbenina_ has quit
13:40 PM
and show how it is better
13:40 PM
x3YTjJV6asPhskrZ
okay
13:40 PM
samj1912
although that would be very qualitative
13:41 PM
x3YTjJV6asPhskrZ
tell you what, ill copy this code and play around with it in python
13:41 PM
samj1912
cool
13:41 PM
x3YTjJV6asPhskrZ
and if i find something i show you
13:41 PM
samj1912
you might also need some data from the WS to play around with
13:41 PM
x3YTjJV6asPhskrZ
yes please :3
13:41 PM
samj1912
well :P
13:42 PM
you can query it
13:42 PM
x3YTjJV6asPhskrZ
okay sure...how do i do that?
13:42 PM
samj1912
13:42 PM
all yours :D
13:42 PM
13:43 PM
a list of really dense(metadata wise) release groups :)
13:43 PM
x3YTjJV6asPhskrZ
ahh okay
13:43 PM
samj1912
anything else I can help you with?
13:43 PM
x3YTjJV6asPhskrZ
nah thats it for now, thanks!
13:43 PM
samj1912
also, as for tagging singles, I did suggest you look at beets
13:43 PM
x3YTjJV6asPhskrZ
yeah
13:44 PM
samj1912
picard is not a mass tagger :P
13:44 PM
x3YTjJV6asPhskrZ
hehe i see that now :)
13:44 PM
samj1912
its more of a customize your media library how you need it tagger
13:46 PM
Slurpee joined the channel
13:54 PM
x3YTjJV6asPhskrZ
hey could i use python-musicbrainz3 for the lookup?
13:54 PM
or well version 2
13:55 PM
oh wait its deprecated
14:03 PM
hey samj1912 can i use python-musicbrainzngs?
14:03 PM
samj1912
yup
14:03 PM
x3YTjJV6asPhskrZ
noice
14:04 PM
okay do you have a release group that has more than one release type/
14:05 PM
samj1912
x3YTjJV6asPhskrZ: I posted the list above
14:05 PM
x3YTjJV6asPhskrZ
but theyre all just showing album
14:05 PM
samj1912
you need to add relations
14:05 PM
x3YTjJV6asPhskrZ
oh?
14:06 PM
samj1912
inc=releases
14:07 PM
14:07 PM
x3YTjJV6asPhskrZ
im not sure im following
14:07 PM
samj1912
x3YTjJV6asPhskrZ: click the link above for example
14:07 PM
14:09 PM
x3YTjJV6asPhskrZ
okay
14:10 PM
agentsim joined the channel
14:14 PM
okay so i have to extract the release type from each release in the release group?
14:14 PM
samj1912
from the medium of each releases from the releases in the release group
14:17 PM
x3YTjJV6asPhskrZ
remeber im using the python api
14:18 PM
how would i do the include releases?
14:24 PM
Rotab joined the channel
14:26 PM
ohrstrom joined the channel
14:29 PM
MajorLurker has quit
14:35 PM
saifulbkhan has quit
14:35 PM
saifulbkhan joined the channel
14:36 PM
saifulbkhan has quit
14:36 PM
ohrstrom has quit
14:41 PM
hey samj1912 where do i get this dict? config.setting["release_type_scores"]
14:49 PM
drsaunders joined the channel
15:00 PM
arbenina_ joined the channel
15:04 PM
*sigh*
15:05 PM
samj1912
x3YTjJV6asPhskrZ: its dynamically generated from you picard config
15:06 PM
x3YTjJV6asPhskrZ
so how do i call that from the test script
15:06 PM
samj1912
do you want the default values?
15:06 PM
x3YTjJV6asPhskrZ
yes please
15:06 PM
samj1912
[('Single', 0.5), ('Compilation', 0.5), ('Spokenword', 0.5), ('Remix', 0.5), ('Mixtape/Street', 0.5), ('Album', 0.5), ('Live', 0.5), ('Demo', 0.5), ('Audiobook', 0.5), ('Other', 0.5), ('Interview', 0.5), ('DJ-mix', 0.5), ('Broadcast', 0.5), ('Soundtrack', 0.5), ('EP', 0.5)]
15:07 PM
basically for singles you would want singles at 1
15:07 PM
and everything else at 0
15:07 PM
x3YTjJV6asPhskrZ
and what if i want albums without compilations (since compilations have albums as a primary type)
15:08 PM
samj1912
just make albums 1, compilations 0
15:08 PM
x3YTjJV6asPhskrZ
and thats not working, hence the bug
15:08 PM
samj1912
not working as in?
15:09 PM
x3YTjJV6asPhskrZ
im still getting compilations
15:09 PM
samj1912
15:10 PM
x3YTjJV6asPhskrZ
which im looking at as we speak
15:11 PM
i think it should take the secondary types into account
15:11 PM
and the role of the primary type should be minimized
15:13 PM
samj1912
how do you want to do that?
15:13 PM
secondary-types might be absent in a lot of cases
15:13 PM
x3YTjJV6asPhskrZ
how bout we add primary type and secondary types to a single list
15:14 PM
and then determine the score based on that list
15:14 PM
gimmie a sec
15:14 PM
do you have a release id with both primary and secondary types set?
15:15 PM
reosarevok
15:16 PM
samj1912
15:17 PM
drsaunders has quit
15:18 PM
x3YTjJV6asPhskrZ
this is a release right, not a release group
15:18 PM
samj1912
yes
15:19 PM
x3YTjJV6asPhskrZ has quit
15:20 PM
x3YTjJV6asPhskrZ joined the channel
15:21 PM
x3YTjJV6asPhskrZ
back
15:21 PM
gimmie a sec
15:22 PM
arbenina_ has quit
15:22 PM
github joined the channel
15:22 PM
github
[critiquebrainz] gentlecat opened pull request #141: Upgrade requirements (master...req-update)
https://git.io/v7VRM
15:22 PM
github has left the channel
15:25 PM
x3YTjJV6asPhskrZ
okay so since this is a release i cant do the release-group include
15:25 PM
which means that this algorithm only looks at release groups when determining the weights
15:26 PM
samj1912
x3YTjJV6asPhskrZ: huh?
15:26 PM
x3YTjJV6asPhskrZ
the weighting algorithm in the metadata.py only looks on release groups, not releases
15:26 PM
samj1912
default picard query for releases include release-groups
15:27 PM
we query releases, which have an inc=release-groups with them
15:27 PM
for every query
15:27 PM
15:29 PM
x3YTjJV6asPhskrZ
so every release has a release group?
15:29 PM
samj1912
yes
15:31 PM
x3YTjJV6asPhskrZ
maybe we should look at the type rather than the primary type
15:31 PM
why are they both there anyways?
15:32 PM
i think this can be solved easily by changing the primary type to type!
15:33 PM
i want to test it. how do i compile the picard source code?
15:33 PM
samj1912
x3YTjJV6asPhskrZ: 1 little problem
15:33 PM
x3YTjJV6asPhskrZ
hmm?
15:33 PM
samj1912
I think you are looking at the xml output
15:33 PM
15:33 PM
json output only has primary-type and seconary-type
15:33 PM
no "type"
15:33 PM
x3YTjJV6asPhskrZ
ahh okay
15:33 PM
samj1912
bitmap: ^^
15:34 PM
any reason for this?
15:34 PM
and can we get "type" included in the output?
15:34 PM
x3YTjJV6asPhskrZ: which picard version are you on?
15:34 PM
x3YTjJV6asPhskrZ
1.4.0
15:34 PM
okay so when picard itself polls for the release, which format does it use? json or xml?
15:34 PM
samj1912
2.x uses json
15:34 PM
1.4.x uses xml
15:35 PM
x3YTjJV6asPhskrZ
okay
15:35 PM
samj1912
15:35 PM
x3YTjJV6asPhskrZ
we could solve this by including a check for both type and primary type
15:35 PM
samj1912
ill fix the above for json output
15:36 PM
but it should be fixed for the xml one
15:36 PM
x3YTjJV6asPhskrZ
okay
15:36 PM
so which version is on the main github page?
15:36 PM
1.4 or 2.0
15:36 PM
samj1912
the master is at 2.0
15:37 PM
x3YTjJV6asPhskrZ
ohh
15:37 PM
samj1912
the current stable is 1.4.2
15:37 PM
x3YTjJV6asPhskrZ
i cant make a pull request for the current stable can i? only for the dev?
15:37 PM
samj1912
you can
15:37 PM
its available at the 1.4.x branch
15:38 PM
x3YTjJV6asPhskrZ
ahh okay
15:38 PM
ohrstrom joined the channel
15:38 PM
so to clarify, xml has type while json only has primary-type
15:39 PM
and 2.0 uses json while 1.4 uses xml
15:39 PM
samj1912
yes
15:39 PM
x3YTjJV6asPhskrZ
okay
15:40 PM
then its a simple matter of checking for all three variables, primary type secondary type and type, and then creating a merge request for both branches once they check out
15:40 PM
now its just to find the right order
15:40 PM
samj1912
I want to know how the type, secondary type and primary type is populated though
15:40 PM
reosarevok: any idea?
15:41 PM
bitmap: yvanzo ^^
15:41 PM
CallerNo6
'type' is derived from primary and secondary. it's a legacy field.