Well, my week was kind of a mess, spent way too much time travelling between the city I come from and the one I study in but at least I got to see a concert in the large hall of https://musicbrainz.org/place/8da8cd3d-6162-471....
On the SpamBrainz side of things, I've gotten most of the API so far done but I'm struggling a bit with how to use Flask Blueprints (just takes a while to figure stuff out).
reosarevok
That's the super cool looking one, right? With all the levels
I also forgot to commit my WIP work so I couldn't work on the train which had a lovely 50 minute delay at 1 in the night, yay!
fin.
Freso
(Only rsh7 and bukwurm / bukwurm_ left on my list.)
Leo__Verto
rsh7, go!
Freso
(Anyone else, speak up ASAP.)
rsh7
hey
Last week I completed the work of importing the MB metadata in AB database. Then worked on the script to update the musicbrainz schema by importing the new data corresponding to a recording whenever there is any new recording insertion in acousticbrainz db.
And read about replication packets and figured out a little on how would i write a script to apply the packets to keep the musicbrainz data in AB db updated with the actual MB database
I have some questions regarding replication packets. bitmap: if you are free for a short discussion after this meeting, please let me know :)
bitmap
yup, after the meeting :)
rsh7
fin! bukwurm?
bukwurm
Hi
rsh7
bitmap: thanks :)
bukwurm
This week on mostly code refactoring and improving mapping.
I set up test environment, wrote some tests.
I am wrapping up last phase objectives.
Documentation and improvement of mapping of some dumps is on radar for this week.
iliekcomputers
last phase of gsoc?
bukwurm
last == previous
So that's it for this week. :)
iliekcomputers
oh. ok.
Freso
bukwurm: So… fin?
bukwurm
Freso: Yeah
:)
Freso
Alright, great!
No one else is left on my list and I haven't seen anyone else asking to go, so…
Thanks bukwurm and everyone else for your reviews!
it seems to be an "art instalment thing" to me which could be served fine with Instrument credits (see my comments on the ticket)
However, I am not certain certain, so incase someone has any counteropinions, _now_ is the time to speak up! :)
freso, you commented on it al ong time ago, do oyu have any imput at all?
Freso
CatQuest: Not really. I looked at the ticket again, but I don't know either which way.
CatQuest
a long time ago, you input*
i fear is that it will basically become a dumping ground for any "guitar played some other way" credits
"altered guitar" i belive we eiher already have or there is a ticket for it
reosarevok
I think I'm happy with ACs (ICs?), unless it becomes common enough, in which case we can review
Freso
Same.
CatQuest
any one else?
Freso
If there is a generic "guitar derivative"/"altered guitar" then the actualy "backwards guitar" instrument can be put there with AC/IC, and guitars played backwards can be AC/IC'd to regular "guitar"/"electric guitar".
CatQuest
yes
Freso
And as reosarevok said, we can review later with more data.
CatQuest
it's an archived rquest and noone else has looked at is since atleast before 2014
so deff. if suddenly it becomes popular or something. reopen or make an ew ticket
a new*
Freso
I'm sure we have plenty of ew tickets already. :)
CatQuest
extra props to HibiscusKazeneko for also replying ot this ticket
Freso: deff 😅
Freso
So done with this DR?
CatQuest
yup, seems liek it
like*
Freso
Alright!
CatQuest
thnak you everyone
Freso
Thanks for your time everyone!
</BANG>
CatQuest
\o/
iliekcomputers
thanks Freso
CatQuest
yes, thanks freso, threso
kartikeyaSh
iliekcomputers: hey!
bitmap
rsh7: you had questions about replication?
kartikeyaSh
I'm going to delete that table now. Won't take much time
rsh7
bitmap: yeah, hi!
bitmap
hey
rsh7
actually I was working on importing the Musicbrainz data in Acousticbrainz database and now i need to add a functionality to keep musicbrainz schema in AB updated with the actual Musicbrainz database
and I would use replication packets to apply in musicbrainz schema tables
iliekcomputers
kartikeyaSh: awesome.
rsh7
bitmap: so, i was thinking that from which replication packet should i start to apply? I thought to go with the latest packets but I read that the new replication packet is applied only when the previous one is already applied to the database. So, from where should i start?
bitmap
you start by importing a data dump, and the last replication sequence applied to that dump is stored in that dump (in the replication_control table)
the tools provided by mb (admin/replication/LoadReplicationChanges) automatically know what replication packet to download and apply
I assume you're using mb's tools, or mbslave, and not writing your own :)
rsh7
oh okay.
No, actually I didn't use mbslave
We needed few tables from MB, so I created a schema for only those tables and wrote a script to import the data
bitmap
from the fullexport dumps?
rsh7
and I was testing with sample MB dumps, I found that the replication_control table is empty there.
bitmap
right
it seems like it would be better to add an option to mb's import script to specify which tables you want to import, instead of reinventing the wheel, but maybe you have other requirements
iliekcomputers
we only wanted those rows that were relevant to us (for ex: only info about recordings that are in AB)
rsh7
yeah
Alright! So, now I am thinking if I create one more table in musicbrainz schema which will store replication sequence number and stuff and then downloads the new packet and script to apply it on the tables and then update the sequence number. But the problem is, what replication packet to start from.
what happens when a recording is added to ab but it doesn't exist in the musicbrainz schema because you didn't import it?
iliekcomputers
bitmap: we plan to have a script running that finds such recordings and imports their data.
kartikeyaSh: thanks.
bitmap
where will you import it from?
iliekcomputers
production MB, the way CB reads it rn.
bitmap
cool
soo this import_mb_data.py script is also reading from the production mb database, not importing from a data dump
if that's the case then you can get replication_control.current_replication_sequence at the beginning of the script, and that's the sequence you start from
there will likely be conflicts (data already existing) when the next packet is released, and you'll need to expect and handle those
rsh7
Alright, got it!
Thank you :)
bitmap
seems like a good argument for keeping all meb projects in one database though
iliekcomputers
yeah.
we came across some issues that are hard to solve efficiently still in CB because we can't join tables across projects.