[02:06:01] *** Quits: drbean (~drbean@124.219.83.228) (Quit: ZNC - http://znc.in) [02:07:42] *** Joins: drbean (~drbean@124.219.82.59) [06:14:54] *** Quits: vin-ivar (~vinit@122.170.43.81) (Quit: WeeChat 1.0.1) [06:15:24] *** Joins: vin-ivar (uid74719@gateway/web/irccloud.com/x-vgvloxmaefmronur) [09:16:54] *** Joins: jmvanel (~jmvanel@66.0.88.79.rev.sfr.net) [09:40:23] *** Quits: jmvanel (~jmvanel@66.0.88.79.rev.sfr.net) (Quit: Quitte) [10:51:00] *** Joins: chru (~cunger@2001:638:504:20e6:3137:1d15:8489:8aa9) [11:00:51] *** Joins: jmvanel (~jmvanel@78.193.21.40) [12:07:59] *** Quits: spectre (~fran@115.114.202.84.customer.cdi.no) (Ping timeout: 246 seconds) [13:08:39] *** Quits: vin-ivar (uid74719@gateway/web/irccloud.com/x-vgvloxmaefmronur) () [13:14:02] *** Joins: spectre (~fran@dhcp89-eduroam2.wifi.uit.no) [13:17:10] *** Joins: vin-ivar (~vinit@122.170.43.81) [14:15:21] www.grammaticalframework.org/lib/doc/synopsis.html [14:15:42] hmm, we could build a sort of mapping from apertium monodices to this [14:19:33] *** Quits: spectre (~fran@dhcp89-eduroam2.wifi.uit.no) (Ping timeout: 256 seconds) [14:21:25] like, the symbols could be converted to categories [14:21:44] the problem is that categories in resource grammars are a lot more complex than sdefs [14:32:36] inariksit, what do you think^? [14:41:14] *** Joins: spectre (~fran@129.242.93.107) [15:32:18] spectre, have you had a look at that? pretty interesting [15:43:57] look at what ? [15:44:02] night! [15:44:02] * Desconnectat (L’argument passat no és vàlid). [15:44:05] this was the last thing i saw [15:45:52] whoops [15:45:59] www.grammaticalframework.org/lib/doc/synopsis.html [15:46:14] we could map sdefs from apertium monodixes to this^ [15:46:23] but these are more complex [15:46:26] yeah, i've seen that before [15:47:42] yeah [15:47:50] a lot of these seem to resemble apertium paradigms, like no_geV in german [15:47:50] but you could use some magic (read unsupervised) to distinguish) [15:48:25] yeah [15:48:31] like what I was thinking was [15:48:46] choosing something based on pos-tagged contexts [15:50:01] so like mkN2 for English [15:50:21] yeah [15:51:05] which is N -> Prep -> N2, I could parse the apertium stream and pick out an (N,Prep) chain and pass that as an N2 [16:06:49] *** spectre is now known as spectie [16:06:53] *** Quits: spectie (~fran@129.242.93.107) (Changing host) [16:06:53] *** Joins: spectie (~fran@unaffiliated/spectie) [16:28:21] *** Quits: spectie (~fran@unaffiliated/spectie) (Ping timeout: 246 seconds) [17:15:22] *** Joins: spectie (~fran@unaffiliated/spectie) [17:19:05] *** Parts: chru (~cunger@2001:638:504:20e6:3137:1d15:8489:8aa9) () [17:42:55] inariksit: can you ping when you have a minute? got a couple of questions [17:43:04] vin-ivar: hey! [17:43:16] I've been marking exams today and now I'm going to choir rehearsal [17:43:21] but after I get home, yeah [17:43:23] well that was lucky [17:43:24] ah, cool [17:43:48] choir is cool, you have any recordings or something up? [17:43:53] hmm [17:44:17] let's see if they have put the recordings from christmas concert online [17:44:50] this one is from a choir party just 3 weeks ago, recorded with mobile: http://listenmaa.fi/kuoro/diga_janeta/diga.mp3 [17:44:52] i've tried ghetto arrangements of old church music at home :D [17:44:56] and we don't know the song super well [17:44:58] ooh cool! [17:44:59] ^^ [17:45:38] nice! [17:45:51] do you guys arrange the parts yourself, or find scores? [17:45:53] http://choir.chs.chalmers.se/choir_web/sites/default/files/uploaded_files/13%20-%20Chalmers%20S%C3%A5ngk%C3%B6r%20-%20A%20Patre%20Unigenitus.mp3 [17:46:12] so the choir has around 120 members, but most of the year we sing in 3 subchoirs [17:46:20] and christmas concerts with all [17:46:44] actually my subchoir is right now practicing a song arranged by me ^^ [17:47:00] damn, that sounds hard. what voices do you normally have? [17:47:04] ooh [17:47:11] the original is this https://www.youtube.com/watch?v=gbpfdQioeHI , I don't have a good recording of it [17:47:23] inariksit, ooh, which one ? [17:47:25] I'll definitely want to record it as soon as we know it well [17:47:29] spectie: bogoroditse [17:47:37] nice ! [17:47:42] i have choir tonight too [17:47:48] we are rehearsing with the byorkestr [17:47:51] vin-ivar: if you have your arrangements online, I'd be interested to see [17:47:52] cool ^^ [17:47:54] so much choir [17:48:00] i don't, haha, they're awful [17:48:09] the teacher of the course I was marking labs today also has choir rehearsal now [17:48:16] i do it for the satisfiction of knowing I'm not totally shit at singing anymore :D [17:48:19] (speaking of which I should start leaving soon :-D) [17:48:21] vin-ivar: ^^ [17:48:28] we need to sing in GF summer school! [17:48:36] count me in o/ [17:48:45] spectie: vintervisa http://choir.chs.chalmers.se/choir_web/sites/default/files/uploaded_files/07%20-%20Chalmers%20S%C3%A5ngk%C3%B6r%20-%20Vintervisa.mp3 [17:48:50] vin-ivar: yes! [17:49:04] i may do a legit recording soon though, girlfriend bought a decent mike which I'm planning to steal [17:49:44] haha that's nice to have [17:50:11] anyway, have fun at choir! ping me when you get back and have some time [17:50:20] is choir stuff very common in northern europe? [17:52:31] yeah I'd say [17:52:39] my uni in finland had 11 choirs [17:53:19] cool [17:53:33] of varying quality and purpose; ranging from the music society of the university to a choir of the students of african and asian studies, who just wanted to sing in non-western european languages [17:53:50] some had high requirements, some were just "come and join every now and then" [17:54:26] now I'm leaving [17:54:29] ---> [17:54:39] sounds really nice [17:54:40] bye! [18:04:56] *** Quits: jmvanel (~jmvanel@78.193.21.40) (Ping timeout: 256 seconds) [18:07:18] *** Joins: jmvanel (~jmvanel@78.193.21.40) [18:36:26] *** Quits: spectie (~fran@unaffiliated/spectie) (Ping timeout: 244 seconds) [18:58:05] *** Joins: spectie (~fran@unaffiliated/spectie) [19:58:15] *** Quits: spectie (~fran@unaffiliated/spectie) (Ping timeout: 265 seconds) [20:46:09] *** Quits: jmvanel (~jmvanel@78.193.21.40) (Quit: Quitte) [21:14:55] vin-ivar: I'm back [21:18:48] hey, great timing, I was just about to crash [21:18:49] so [21:19:00] what I was thinking was [21:19:39] now that i'm done with the analyser and disambiguator, I could work on a script to add to the GF lexicon, using apertium's lexicon [21:19:47] yeah! [21:19:50] that'd be nice [21:20:20] I have some crappy python script I wrote while staying up all night waiting for Flammie to catch his plane :-D [21:20:30] now the problem here is that apertium's symbols are a lot 'simpler' than GF's symbols, there's no complex N2 and stuff in apertium [21:20:33] ooh, awesome [21:21:07] so I was thinking of first just doing a really basic parse on the apertium stream [21:21:24] is that what your script does? [21:21:51] mine just takes apertium vblex and outputs GF V, or apertium n and outputs GF N [21:21:56] so nothing fancy at all :-P [21:22:03] cool, same thing in principle :P [21:22:31] so I could get more complex GF ones like N2 by parsing context [21:22:42] in the stream [21:23:14] a noun and then a preposition could combine, give me GF N2 [21:23:15] that would be interesting next step to try ^^ so would you just do it for a corpus, and find lexicon entries from there? [21:23:30] yeah [21:23:36] yeah [21:23:38] i'd do the morph analysis through apertium [21:23:42] (for the corpus) [21:24:05] and then some sort of longest-path thing? [21:24:19] if I have N and Prep, return N2 and not separate N and Prep [21:24:29] something like that [21:24:43] ahahah nevermind, I'm an idiot, my script was from GF to apertium :-D [21:25:00] ok I must've been so confused that it didn't even register to my mind what I was writing [21:25:04] ah well [21:25:05] haha [21:25:19] (but yeah in any case, it wouldn't have been very helpful, you've done the baseline already :) [21:25:20] apertium has a bunch of wrappers for python now iirc [21:25:47] is there a mini resource grammar I can get for german? [21:25:52] hmm [21:26:10] i know there's a full fledged resource grammar but it'd be easier if I work on a smaller one for starters :P [21:26:16] yeah, definitely [21:26:20] not in gf-contrib at least [21:26:39] haha there's gothic mini resource [21:26:55] every_Det = mkDet "hwazuh" "hwizuh" "hwammeh" "hwanoh" "hwah" "hwizuh" "hwammeh" "hwah" "hwoh" "hwizozuh" "hwizaih" "hwoh" Sg Strong ; [21:27:38] wut [21:27:49] just browsing this https://github.com/GrammaticalFramework/gf-contrib [21:28:02] this is awesome, isn't gothic extinct? :D [21:28:06] yeah [21:28:20] doesn't mean you can't write a GF grammar for it ^^ [21:28:43] i know, it's pretty cool [21:28:58] hm, maybe I could roll my own basic german grammar [21:29:22] or find a mini RG in GF that has a decent monodix in apertium [21:29:37] does italian have one? [21:30:13] yep [21:30:18] it isn't huge but it's pretty large [21:30:25] 10,388 stems, 443 paradigms [21:30:28] ok [21:30:45] yeah I could do italian, I think [21:31:50] cool [21:32:04] so i'll give this a shot tomorrow, then [21:32:11] brain is about to shut down now [21:32:22] btw if you are working with the grammars, this page might be helpful: http://www.cse.chalmers.se/~aarne/computationalsyntax/ [21:32:25] 2. Extended mini resource grammar [21:32:27] Copy the latest sources of grammar assignment [21:32:30] Place your own modules Grammar, Lang, Lexicon, Paradigms, Res in the resulting directory extmini/ [21:32:33] Go to extmini/ [21:32:35] Run the lab test script by: [21:32:38] gf -run LangXXX.gf replacing XXX with your own language code. [21:32:41] haha it must be pretty late there? [21:32:44] Inspect the resulting file compsyntax-test.txt to decide whether it is OK; it should look similar to compsyntax-test-Ita.txt. [21:33:39] i think I cloned a pre-existing extmini from github [21:33:48] thanks, this is useful [21:34:08] :) [21:34:31] anyway, i'm off to bed now, I'll ghetto something together tomorrow [21:34:40] cool! [21:34:41] good night [21:34:43] thanks for the help inariksit [21:34:44] night! [21:55:00] *** Joins: jmvanel (~jmvanel@66.0.88.79.rev.sfr.net) [22:21:38] *** Joins: spectie (~fran@unaffiliated/spectie)