bitmap: plus, "1) clean URL display as a link allows to visually check it (whereas input field may hide part of it) 2) dialog can be extended with options: apply to unclean, bypass cleanup (whereas bubble doesn’t allow that easily)" from https://chatlogs.metabrainz.org/libera/metabrainz…
2021-07-14 19555, 2021
ritiek has quit
2021-07-14 19508, 2021
tandy[m]
<Lotheric "@ListenBrainz devs - I amde a ti"> does anyone still use opera?
2021-07-14 19507, 2021
lucifer
param: hi! a quick question for last.fm importer we currently have handwired logic to disable submit button till username has not been filled. is it fine to use the required attribute instead?
i am adding another radio button here so wanted to know if there is some reason behind this.
2021-07-14 19550, 2021
alastairp
lucifer: yeah, that's probably fine
2021-07-14 19502, 2021
lucifer
👍 thanks
2021-07-14 19515, 2021
alastairp
although, does `required` only trigger on form submit? do we use a submit, or do we just read the value?
2021-07-14 19559, 2021
lucifer
we do a submit and but do not use that value.
2021-07-14 19524, 2021
lucifer
i'll change it just use that the submitted value as its simpler.
2021-07-14 19548, 2021
alastairp
ok, cool. as long as the functionality still works
2021-07-14 19504, 2021
lucifer
+1
2021-07-14 19530, 2021
Lotheric_ joined the channel
2021-07-14 19554, 2021
Lotheric has quit
2021-07-14 19541, 2021
ruaok
moin!
2021-07-14 19545, 2021
ruaok
lucifer: did you try the import?
2021-07-14 19527, 2021
lucifer
ruaok: yes, the dump imported fine. we can also try playing with it in the repl if we want.
2021-07-14 19507, 2021
ruaok
one thing I changed was to move from sorting by created to sorted by listened_at. does it even make sense to sort by created for spark?
2021-07-14 19516, 2021
ruaok
that we need to consider.
2021-07-14 19545, 2021
ruaok
can we run some recs or similarities with this dump? or is that too much a hassle?
2021-07-14 19557, 2021
lucifer
we'll need to rewrite the recs/similarity/stats code each to use the new dump. so not possible to run those just yet.
2021-07-14 19506, 2021
ruaok
oh duh.
2021-07-14 19508, 2021
ruaok
right.
2021-07-14 19545, 2021
lucifer
the sorting can help with not reading the entire data in memory if we just want to calculate the last week stats.
2021-07-14 19508, 2021
ruaok
but that would be using listened_at, not created, yes?
2021-07-14 19522, 2021
lucifer
that was the reason we sorted it by created, because then using incremental dumps becomes simpler.
2021-07-14 19551, 2021
ruaok
should I go back and sort by created?
2021-07-14 19527, 2021
ruaok
there are listens with no created date, so 1970-1-1
2021-07-14 19529, 2021
lucifer
no need. i think from user perspective it makes sense to use listened_at anyways. i am thinking of loading all inc dumps since last full dump in memory and just what we need the full dump.
2021-07-14 19502, 2021
ruaok
ok, tell me if that changes. sorting on a column that is only partially populated presents some problems.
2021-07-14 19535, 2021
lucifer
yes makes sense.
2021-07-14 19523, 2021
ruaok
wow. hetzer fucked up AGAIN.
2021-07-14 19554, 2021
lucifer
that's the usual now it seems
2021-07-14 19543, 2021
BilleeKhoj
tandy[m]: and vivaldi which is a lot of it (and more people should use vivaldi)
2021-07-14 19544, 2021
BilleeKhoj
!!
2021-07-14 19526, 2021
BilleeKhoj
wait what happened at hetzer now? i thoguht an "adult" was looking at it
yes, i have reviewed the picard addition PR again, will merge after one final test. same for onboarding tweaks pr.
2021-07-14 19534, 2021
akshaaatt[m]
Thanks a lot!
2021-07-14 19526, 2021
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged; see https://musicbrainz.org/doc/IRC for details | Agenda: Reviews, Fix brainzbot in #MB & #BB (CQ)
2021-07-14 19509, 2021
lucifer
ruaok: there are 368958011 listens in the dump. does that look accurate?
2021-07-14 19544, 2021
lucifer
lb website count puts the number at 443,723,619 so not sure.
2021-07-14 19513, 2021
lucifer
oh! the dump has only listens with mbids. so probably right.
2021-07-14 19534, 2021
ruaok
Possibly. I'll have to check.
2021-07-14 19534, 2021
ruaok
Zas: i see that boingo has been replaced with ax-51, the Williams replacement. Can you have a look?
yvanzo: What is the reason to keep links with empty URL but non-empty type, can we just delete it?
2021-07-14 19541, 2021
ritiek joined the channel
2021-07-14 19540, 2021
yvanzo
yyoung: To allow to undo blanking an URL I guess. This is probably not needed anymore if URL has to be edited in popover.
2021-07-14 19535, 2021
yyoung
OK, so I may remove links that are emptied.
2021-07-14 19555, 2021
yvanzo
Yes :)
2021-07-14 19537, 2021
zas
ruaok: I'm around, yes the machine has 2 disks Seagate ST12000NM002G 12TB
2021-07-14 19507, 2021
zas
it has 2 network interfaces as expected
2021-07-14 19515, 2021
zas
10 gb on /home
2021-07-14 19534, 2021
zas
I will configure it tomorrow morning
2021-07-14 19521, 2021
ruaok
zas: ok, great. lets call it 'aretha'. :)
2021-07-14 19553, 2021
zas
👍
2021-07-14 19533, 2021
zas
I'll probably re-partition
2021-07-14 19548, 2021
ruaok
yeah, good. I dislike that partition scheme. confusing.
2021-07-14 19555, 2021
ruaok
server names wiki page updated.
2021-07-14 19500, 2021
agatzk has quit
2021-07-14 19503, 2021
agatzk joined the channel
2021-07-14 19535, 2021
samthursfield has quit
2021-07-14 19540, 2021
samthursfield5 joined the channel
2021-07-14 19541, 2021
lucifer
ruaok: which of the fields in the parquet dumps are nullable?
2021-07-14 19510, 2021
reosarevok
Don't dump your parquet, that shit is expensive!
2021-07-14 19503, 2021
lucifer
😆
2021-07-14 19554, 2021
lucifer
i think you mentioned created is nullable or is it just set to 1970-1-1. other than that i am unsure about release_name and release_mbid fields. all other are non null acc to my understanding.
2021-07-14 19530, 2021
BilleeKhoj
wait. is parquet = parkett?
2021-07-14 19544, 2021
BilleeKhoj
that stuff that's floor and not stone or wood or vinyl
2021-07-14 19552, 2021
BilleeKhoj
but kinda like vinyl
2021-07-14 19514, 2021
lucifer
i think so
2021-07-14 19558, 2021
BilleeKhoj
wow
2021-07-14 19500, 2021
BilleeKhoj
TIL
2021-07-14 19549, 2021
kepstin
parquet refers to wood floors made of small rectangular pieces of wood in patterns.
2021-07-14 19542, 2021
kepstin
you might be thinking of linoleum flooring?
2021-07-14 19553, 2021
kepstin
(around here, vinyl flooring is sometimes called "linoleum", but actual linoleum is a distinct material)
2021-07-14 19547, 2021
BilleeKhoj
yes i was thinking about lineolieum, but parkett is the wood panneling stuff
2021-07-14 19559, 2021
BilleeKhoj
its hot and my brain is mush.
2021-07-14 19526, 2021
BilleeKhoj
(i will neer understad how indians can een think in 30-40+ degrees)
2021-07-14 19542, 2021
BilleeKhoj
:| f my v key
2021-07-14 19541, 2021
BilleeKhoj
bitmap: I was advised to talk to you about this. on https://beta.musicbrainz.org/statistics/timeline/… is it possible to have a "uncheck all" ticketbox or something? it'd be less work and redrawing to uncheck all and the recheck the 1-2 if you need only that many (wich is often the case)
2021-07-14 19517, 2021
bitmap
BilleeKhoj: sounds like a good idea to me, I can make a note to add it to the PR I have converting it to React
2021-07-14 19504, 2021
BilleeKhoj
that's what reo told me you where doing :)
2021-07-14 19558, 2021
ruaok
lucifer: currently none are nullable. because we export only matched tracks, which have an album looked up.
2021-07-14 19517, 2021
lucifer
makes sense, 👍
2021-07-14 19524, 2021
ruaok
but, skipping non-identified tracks is actually a bad thing -- bad for stats, bad for id'ing missing data.
2021-07-14 19544, 2021
ruaok
I think I should fix the PR to allow MBIDs and release_name to be nullable. what do you think?
2021-07-14 19521, 2021
lucifer
i think so yes, otherwise we can't calculate missing mb data. for all other purposes, its probably fine.
2021-07-14 19533, 2021
ruaok
ok, I'll go do that. its and easy change.
2021-07-14 19557, 2021
lucifer
you can also remove created field. since we are sorting on listened_at we won't be needing it.
2021-07-14 19546, 2021
ruaok
k
2021-07-14 19557, 2021
ruaok
makes the dumps smaller, woo.
2021-07-14 19537, 2021
lucifer
ruaok: btw, 368,958,011/443,723,619 means 83% listens had matches :D \o/
the new process will work out quite well i believe. this is with the old dumps and mapping.
2021-07-14 19527, 2021
lucifer
with the new dumps we can simplify these queries a lot
2021-07-14 19545, 2021
ritiek has quit
2021-07-14 19503, 2021
ritiek joined the channel
2021-07-14 19553, 2021
lucifer
another thing, is there a future use case where we would want to load just the listens between two specified dates. the way i am writing it currently is that it only takes a start_date and loads everything upto now. all the current use cases fit this.
2021-07-14 19547, 2021
lucifer
on second thought, will handle it if and when such a use case comes up.
2021-07-14 19549, 2021
ritiek has quit
2021-07-14 19531, 2021
ruaok
83%? hmmm.
2021-07-14 19517, 2021
ruaok
oh right. 60% of messybrainz IDs have been looked up, but I guess that equates to 83% of all listens.
