Yes, but since BB doesn't have a stable API, could I include revamping the the API in my project as well?
Gentlecat
I'd prefer not having to have a separate instance of BB running
that's one more thing to maintain and keep in sync
API of what? BB?
muesli has quit
muesli joined the channel
akashn97
Or how about caching the database indefinitely and then, check if it's time to invalidate it. This should reduce the requests, and we can increase frequency of the data dumps
Gentlecat
sure, caching can improve the performance, but that's not the issue
where are these data dumps and how often they are updated? is there an easy way to import them regularly?
akashn97
We could import only the invalidated ones, so there would definitely be an easy way around.
Where would you suggest we store the data dumps?
Gentlecat
when you say "the invalidated ones", what do you mean?
jasperkirton has quit
akashn97
The ones which have a change in them and the cached version is no longer the same as the most up-to-date version.
It's similar to the dirty bit used in cache
Gentlecat
ok, but what data are you talking about importing? BB data?
I don't think I understand how you see parts connecting together
the issue is having access to BB data in the first place
idk if BB has a reliable way to run a separate replica of the main database
since other options of accessing are out of the question currently
and if it doesn't then I don't see this happening at all
"this" being having BB data in CB
samj1912
zas: also updated the py3 PR, it isn't in the state to be merged though
LordSputnik
Gentlecat: the only way to keep an up to date replica at the moment is to drop the database and reimport
Which takes about a minute
UmkaDK has quit
We'd welcome a replication system as part of a proposal though
UmkaDK joined the channel
Gentlecat
it wouldn't be a CB project anymore though
and I don't see both parts being doable within one GSoC
arbenina_ joined the channel
deeptibaghel joined the channel
agentsim joined the channel
saifulbkhan joined the channel
agentsim has quit
madmouser1 joined the channel
justharshal
LordSputnik: Shared my draft on GSOC site. Feedback?
justharshal: ok, needs more technical detail, particularly on how the rate limiting will work
I'd like to know also what the authentication process will be and how it'll work
ruaok
LordSputnik: we should use the same method that LB uses.
which is the twitter way of doing.
Slurpee has quit
LordSputnik
justharshal: and try to introduce some of your own ideas, because I'm not seeing a lot of thought on there beyond what was in the initial suggestion and the ideas I've discussed with you on IRC
ruaok: I'll investigate later, unless you have time to give a brief description :)
ruaok
when you make a request, it gives you a window of time.
usually for X seconds and it tells you have many more requests you can make inside that window.
it is up the client to decide how to space the requests. they can make them all at the beginning or space evenly.