ianmcorvidae: goodies can't do web requests, so they don't make sense for MB
2013-11-06 31049, 2013
ruaok
er responses.
2013-11-06 31017, 2013
ianmcorvidae
JonnyJD_: no, but point stands -- I think longtail is probably the main one that's workable for MB data
2013-11-06 31035, 2013
ianmcorvidae
JonnyJD_: only included it because it's keyword-oriented, like all the others that aren't longtail
2013-11-06 31052, 2013
navap
Responses? What do they have to do with the logs?
2013-11-06 31025, 2013
ianmcorvidae
hm, possibly the longtail ones are also keyword-based
2013-11-06 31036, 2013
misterswag joined the channel
2013-11-06 31004, 2013
ruaok
navap: I thought you were talking about responses, since you said "compact"
2013-11-06 31019, 2013
ianmcorvidae
I'd still go with fathead or longtail, anyway, since I don't know how they throttle their requests to our WS, if they do -- though we'd need to find a source that could be used for it (I *guess* the data dumps, but that'd be hairy)
2013-11-06 31033, 2013
njh
Twitter URLs seems to be stored in MusicBrainz using a mixture of http and https. Is this a known issue? Can I raise a bug?
2013-11-06 31018, 2013
navap
ruaok: I was thinking about the possibility of keeping logs for a longer time period so we have more to analyze. 1 web visit -> several hits, 1 ws visit -> 1 hit. That's what I meant by more compact
2013-11-06 31020, 2013
ianmcorvidae
njh: they changed it at some point (twitter, I mean); afaik the notion is that https is preferred but they need to be changed over time
2013-11-06 31047, 2013
nikki
ianmcorvidae: no, they've already been fixed
2013-11-06 31051, 2013
ruaok
navap: ah.
2013-11-06 31006, 2013
rvedotrc
I can't get used to ruaok being in the same(ish) time zone as me.
2013-11-06 31007, 2013
nikki
there's exactly two left, which probably slipped past the cleanup
2013-11-06 31014, 2013
ianmcorvidae
ah, lol
2013-11-06 31018, 2013
ruaok
rvedotrc: :)
2013-11-06 31027, 2013
rvedotrc
It's like: replies, but faster! :-D
2013-11-06 31051, 2013
nikki
haha
2013-11-06 31057, 2013
nikki
yeah, it's still weird to me too
2013-11-06 31005, 2013
ruaok
navap: we could double the number of days we keep, but then we'd be outta space on that server.
2013-11-06 31020, 2013
ruaok
nikki: even after 6 months?
2013-11-06 31042, 2013
nikki
njh: anyway, there's your answer :P our data was already fixed quite a while ago, if yours isn't then it's out of date
2013-11-06 31049, 2013
ruaok
ok, I gotta head out. getting dark here in b-town.
2013-11-06 31019, 2013
nikki
ruaok: yes. 6 months isn't much after 8 years or so of knowing you were in california :P
2013-11-06 31026, 2013
ruaok
ha, true that
2013-11-06 31037, 2013
ruaok
bbiab
2013-11-06 31033, 2013
JonnyJD_
ianmcorvidae: sounds like the duckduckhack stuff will again be exposed in an api, so one could use the duckduckgo api to query musicbrainz indirectly. Sounds like the number of requests can be considerable.
2013-11-06 31000, 2013
ianmcorvidae
JonnyJD_: yeah -- looking further they do cache results for a period
2013-11-06 31014, 2013
JonnyJD_
https://api.duckduckgo.com/api : Our long-term goal is for all of our instant answers to be available through this open API.
2013-11-06 31024, 2013
ianmcorvidae
(which you can change/configure -- but we might still have to give them a special ratelimit bucket)
2013-11-06 31030, 2013
misterswag joined the channel
2013-11-06 31058, 2013
MBJenkins joined the channel
2013-11-06 31001, 2013
ianmcorvidae
yeah, it is
2013-11-06 31007, 2013
ianmcorvidae
(going off [off], doesn't matter much)
2013-11-06 31050, 2013
ianmcorvidae
some things using fathead are e.g. the arch packages thing, which downloads a list of packages and processes it into their tab-separated format
2013-11-06 31037, 2013
ianmcorvidae
so I guess their intention is to use spice where there's an API to use and fathead when you need to process a downloaded file, but
2013-11-06 31040, 2013
JonnyJD_
yep, I also found the arch thing
2013-11-06 31058, 2013
JonnyJD_
but the code is just some metadata, couldn't find the rest
