Quesito: I was wondering if we should walk through the process of adding a couple of customers and sending them an invoice...
2017-04-05 09519, 2017
lazka joined the channel
2017-04-05 09541, 2017
Quesito
ruaok: yes I believe so--currently on the way to Dr apt. But perhaps after?
2017-04-05 09517, 2017
drsaunder joined the channel
2017-04-05 09550, 2017
ruaok
Quesito: great!
2017-04-05 09533, 2017
d4rkie has quit
2017-04-05 09501, 2017
D4RK-PH0ENiX joined the channel
2017-04-05 09549, 2017
deepti has quit
2017-04-05 09546, 2017
Freso
samj1912: Heh. I was actually looking at webbrowser2 myself last night and was going to make a PR to simply remove the file and replace calls to webbrowser2 in Picard code with regular Python webbrowser ones.
2017-04-05 09534, 2017
Freso
samj1912: I'd be happy to make that PR right away. (It doesn't need to wait for Py3 port.)
2017-04-05 09546, 2017
drsaunder has quit
2017-04-05 09501, 2017
ruaok
another service that might be useful for our customers: collect and download wikipedia artist pages and put them into a convenient download
2017-04-05 09506, 2017
samj1912
Freso: you can :)
2017-04-05 09523, 2017
samj1912
I'll have to rebase though :p
2017-04-05 09500, 2017
Freso
samj1912: Actually, the error catching in open() and the Picard specific goto() might be nice to keep around, so maybe https://github.com/metabrainz/picard/pull/664/com… should just go into its own PR. (Possibly with a rewritten docstring rather than a completely removed one. ;)).
2017-04-05 09517, 2017
Freso
samj1912: I can make a new PR doing this if you don't feel like it.
2017-04-05 09534, 2017
Freso
("webbrowser2" seems like the wrong name for this though, but eh. :))
2017-04-05 09531, 2017
samj1912
I wouldn't mind picard.webbrowser
2017-04-05 09553, 2017
samj1912
Freso, I'll be done with the py3 port in a couple of days though
2017-04-05 09505, 2017
samj1912
Since the webbrowser code is changed in py3 even more I think it should be included in the PR
2017-04-05 09514, 2017
Jormangeud has quit
2017-04-05 09518, 2017
Freso
But all webbrowser stuff you're doing is webbrowser.open(), which (at an API level) is the same between Py2 and Py3. But okay.
2017-04-05 09534, 2017
Freso
Grr. I can't comment on code in the expanded parts. :(
2017-04-05 09517, 2017
samj1912
Hehe, it also registers a few browsers depending on the OS to run at a higher priority if they don't exist in the python registry
very good. then we have no problems to deal with in that regard, right?
2017-04-05 09525, 2017
iliekcomputers__
yes, seems that way to me,
2017-04-05 09513, 2017
ruaok
great. I hope to have some time this afternoon to go over the bugs in jira and figure out what needs to be done before we go to beta.
2017-04-05 09530, 2017
ruaok
I also need to figure out why we have a different number of listens in BigQuery vs in Influx.
2017-04-05 09557, 2017
ruaok
BQ has 1M less than influx. We may need to ditch the data in both and start over again, but I'll likely start with a smaller test.
2017-04-05 09509, 2017
samj1912
Freso: done :)
2017-04-05 09540, 2017
iliekcomputers__
ruaok: okay, we should probably ditch the data before beta anyways because a lot of data in additional_info has probably been dropped before LB-139 was fixed
let me investigate the disparity between BQ and Influx and once I have that fixed (or whatever) we an drop the data again.
2017-04-05 09518, 2017
iliekcomputers__
okay great
2017-04-05 09531, 2017
ruaok
iliekcomputers__: want a "fun" task to work on in the meantime?
2017-04-05 09540, 2017
iliekcomputers__
sure
2017-04-05 09558, 2017
ruaok
I'd like to see if influx is up to the task that we're setting out for it.
2017-04-05 09535, 2017
ruaok
I would like you to write a script that inserts 1B listens into LB.
2017-04-05 09552, 2017
alastairp
one BILLION DOLLARS!
2017-04-05 09556, 2017
ruaok
lol.
2017-04-05 09535, 2017
ruaok
I imagine that this script would be invoked with:
2017-04-05 09540, 2017
ruaok
1) total number of listens to enter
2017-04-05 09549, 2017
ruaok
2) number of simultaneous threads to run
2017-04-05 09502, 2017
ruaok
3) the average number of listens per user to add.
2017-04-05 09517, 2017
ruaok
4) the number of users to spread the number from #1 over.
2017-04-05 09505, 2017
iliekcomputers__
3 seems to be dependent on 4
2017-04-05 09542, 2017
ruaok
the script should then fire up a number of threads, one user per thread. the script should insert from 50% less to 50% more of #3.
2017-04-05 09504, 2017
alastairp
I wonder if there's an interesting distribution which you could use here
2017-04-05 09509, 2017
ruaok
we may need to run a script on the server that creates a pile of users ahead of time and records their tokens...
2017-04-05 09505, 2017
ruaok
or just make a branch that randomly accepts auth tokens and if one is not recognized, just make a user for it and accept listens.
2017-04-05 09536, 2017
ruaok
iliekcomputers__ all the figures are vaguely dependent, no?
2017-04-05 09513, 2017
ruaok
my point is to try and simulate a heavy load of users submitting listens in a short time frame.
2017-04-05 09523, 2017
iliekcomputers__
ruaok: right
2017-04-05 09535, 2017
ruaok
oh and the listen data itself should just be random characters.
2017-04-05 09501, 2017
ruaok
not always the same and roughly similar to typical listen.
2017-04-05 09525, 2017
iliekcomputers__
ruaok: okay, I'll try to come up with something tonight or tomorrow
2017-04-05 09534, 2017
iliekcomputers__
probably tonight
2017-04-05 09536, 2017
ruaok
<3
2017-04-05 09552, 2017
samj1912
ruaok: how many of the gsoc applicants were indians this time? :P
2017-04-05 09548, 2017
ruaok
120%, I think.
2017-04-05 09557, 2017
samj1912
lol :P
2017-04-05 09513, 2017
iliekcomputers__
tfw google has reduced the stipend ;_;
2017-04-05 09532, 2017
samj1912
I dont think that affected anyone's enthusiasm :P
2017-04-05 09537, 2017
ruaok
well, the harmonized it.
2017-04-05 09548, 2017
iliekcomputers__
samj1912: it hasn't :)
2017-04-05 09551, 2017
ruaok
so that people all over the world more or less get paid the same.
2017-04-05 09510, 2017
samj1912
but yeah, gsoc pays pays about same or less as an avg. intern now
2017-04-05 09518, 2017
ruaok
out of the 8 proposals we have 7 appear from Indians.
2017-04-05 09541, 2017
jwf
Is it actually adjusted by country now?
2017-04-05 09548, 2017
samj1912
jwf: yup
2017-04-05 09552, 2017
jwf
Oh, wow
2017-04-05 09501, 2017
ruaok
a huge undertaking for them to do that.
2017-04-05 09526, 2017
jwf
I only learned recently that until now, GSoC has been a substantial amount of money for people from smaller countries
2017-04-05 09528, 2017
ruaok
but there is the famous story of an indian guy taking his payments and buying a new machine and finishing his project successfully.
2017-04-05 09539, 2017
ruaok
then he started a company and hired a few people with the rest of the money.
2017-04-05 09549, 2017
samj1912
lol :P
2017-04-05 09552, 2017
samj1912
pretty possible
2017-04-05 09552, 2017
jwf
When I did GSoC last summer, it wasn't much money by American standards, but it was enough
2017-04-05 09557, 2017
jwf
ruaok: …wow.
2017-04-05 09502, 2017
ruaok
A swiss or norwegian person would be fairly poorly paid in comparison.
2017-04-05 09554, 2017
samj1912
I dont think they increased the max limit that much?
2017-04-05 09507, 2017
jwf
If I lived in a bigger city, the total sum would have only just been able to cover rent for the duration of GSoC. I worked another job during the summer too which took away from some of my time to contribute
2017-04-05 09544, 2017
samj1912
oh btw jwf: chhavi says hi :P
2017-04-05 09545, 2017
ruaok
yeah, we really frown on people working jobs during GSoC.
2017-04-05 09550, 2017
ruaok
that has never gone well.
2017-04-05 09524, 2017
Sophist-UK
zas, sam1912, anyone else: PICARD-1043 https://github.com/metabrainz/picard/pull/676 is ready for review. However support for total number of parts really works only for releases with all parts (pending MBS-9304 to provide this in WS XML), and I need some examples of works with >1 level of parts breakdown to add support for this.
