[06:38:47] <drbean> I don't see certify_V in IrregEng.
[07:07:25] *** Quits: proteusguy (~proteus-g@2405:9800:b408:bc31:d29:6aae:a9bf:ab8d) (Remote host closed the connection)
[09:07:13] <inariksit> freeside: let foo : SomeParam => SomeValue = table { foo => bar ; ... }
[09:07:38] <inariksit> but I don't think you can use the category Det as a left-hand side of a table
[09:13:33] <inariksit> daherb: have you managed to complete a course in duolingo?
[09:13:50] <inariksit> yesterday I had only the last 2 lessons left, and now they go like "Notice something different?
[09:13:51] <inariksit> We’ve updated your skill tree to teach you Dutch more effectively."
[09:14:21] <inariksit> and I suddenly have a few new lessons such as spirituality and bathroom, and most of my previously golden skills are at 0 again
[09:14:27] <inariksit> and the last 2 lessons are greyed out
[09:18:03] *** Joins: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151)
[09:20:04] <freeside> mmm. it worked when i moved the table to a standalone oper.
[09:20:06] <freeside> sadTable : Det => Conj; sadTable = table { some_Det =>  or_Conj ; any_Det  => and_Conj };
[09:20:19] <inariksit> okay, cool!
[09:20:28] <freeside> > cc sadTable!any_Det
[09:20:28] <freeside> {n = ParamX.Sg; s1 = []; s2 = "or"; lock_Conj = <>}
[09:22:07] <inariksit> cool, so then any finite type works as a left-hand side of a table ^^
[09:22:54] <inariksit> does it work if you just name it in a let or where?
[09:23:00] <freeside> sadly, no
[09:23:41] <freeside> wait, i am wrong
[09:24:26] <freeside>     listvpi2v2 : ListVPI -> Det -> V2;
[09:24:26] <freeside>     listvpi2v2 listvpi det =
[09:24:26] <freeside>       let sadTable : Det => Conj = table { some_Det =>  or_Conj ; any_Det  => and_Conj } in
[09:24:29] <freeside>       let conj = sadTable ! det in
[09:24:31] <freeside>       let vpi  = ConjVPI conj listvpi in
[09:24:34] <freeside>       mkV2 (vpi.s ! RE.VVPresPart ! RE.AgP3Pl RE.Neutr);
[09:24:45] <freeside> i was trying to squish the sadTable ! det into a case statement
[09:24:56] <freeside> that was what didn't work.
[09:25:09] <inariksit> yay
[09:26:19] <inariksit> you can btw just do this
[09:26:19] <inariksit> let sadTable : Det => Conj = table { some_Det =>  or_Conj ; any_Det  => and_Conj }  conj = sadTable ! det  vpi  = ConjVPI conj listvpi
[09:26:19] <inariksit> in mkV2 (vpi.s ! RE.VVPresPart ! RE.AgP3Pl RE.Neutr);
[09:26:31] <freeside> well, to be honest, that whole thing there doesn't work, because i'm trying to do for a VPI what ExtensionsEng.GerundNP does for VP
[09:26:40] <inariksit> sorry, forgot ;s
[09:26:41] <inariksit> let sadTable : Det => Conj = table { some_Det =>  or_Conj ; any_Det  => and_Conj } ;
[09:26:41] <inariksit> conj = sadTable ! det ;
[09:26:41] <inariksit> vpi  = ConjVPI conj listvpi ;
[09:26:42] <inariksit> in mkV2 (vpi.s ! RE.VVPresPart ! RE.AgP3Pl RE.Neutr);
[09:26:57] <freeside> oh, i can do that?
[09:27:11] <inariksit> yeah, as long as you separate with ; the different variables in a let
[09:28:36] <freeside> ah, it is there in the reference manual.
[09:30:03] <freeside> anyway, after wrestling with my GerundNP : VPI -> NP problem for some time yesterday at a japanese restaurant, i swallowed a large quantity of wasabi and gave up, but now i am ready to try again
[09:31:44] <freeside> here is my problem statement: ExtensionsEng.GerundNP knows how to VP -> NP, but now I want to VPI -> NP, where the VPI is a ConjVPI of a ListVPI.
[09:32:02] <inariksit> okay let's see
[09:32:06] <freeside> do you have any advice for how to do this? after my wasabi incident i manually ripped out mkV2 (vpi.s ! RE.VVPresPart ! RE.AgP3Pl RE.Neutr);
[09:32:47] <freeside> code is at the bottom of https://github.com/legalese/legalese-compiler/blob/master/sandbox-gf1/missouri/MissouriI.gf
[09:33:44] <freeside> i used the LibraryBrowser to construct an analogous sentence, but it just treated all the verbs as nouns
[09:34:41] <freeside> LargeLibraryBrowser> p -lang=Eng "I like to run and hide" | l -lang=API
[09:34:52] <freeside> mkUtt ( mkS ( mkCl ( mkNP i_Pron ) ( mkVPSlash ( mkVPSlash like_V2 ) to_Adv ) ( mkNP ( mkCN and_Conj ( mkListCN ( mkCN run_N ) ( mkCN hide_N ) ) ) ) ) )
[09:35:07] <freeside> arguably incorrect? i don't know.
[09:35:22] <inariksit> oh yeah the run and hide are Ns
[09:35:29] <freeside> they are? wow
[09:35:41] <freeside> i thought they were verbs. i should really just give up on speaking english
[09:35:48] <inariksit> hahah I mean in the parse
[09:35:50] <inariksit> they are totally verbs
[09:35:57] <inariksit> but in here
[09:35:59] <inariksit> mkListCN ( mkCN run_N ) ( mkCN hide_N )
[09:36:01] <inariksit> that is super wrong
[09:36:09] <inariksit> we could suggest this to aarne, now that he is generalising the Extra and Extensions to Extend
[09:36:21] <inariksit> to generalise GerundNP to VPIs
[09:37:00] <freeside> the generalising of making gerunds would be to my liking
[09:37:17] * inariksit is appreciating
[09:37:21] <freeside> haha
[09:37:27] <inariksit> sounds very lawyerish too
[09:38:05] <freeside> yes, maybe Legalese can sponsor an Obfuscated GF contest in future
[09:38:21] <inariksit> ^^
[09:38:34] <freeside> so, what i have so far, produces
[09:38:59] <freeside> > l mkReactionRule repealing MustNot abating
[09:39:00] <freeside> the repealing of the section and laws repealed by this law shall not have the effect of abating , nullifying , suspending or vitiatinging any proceeding .
[09:39:44] <freeside> which is almost a perfect match for the original text, https://github.com/legalese/legalese-compiler/blob/master/sandbox-gf1/missouri/missouri.org
[09:39:58] <freeside> except there is too much inging
[09:40:33] <inariksit> vitiatinging :-D
[09:50:40] <mbrock> hmm, I want to attach a parenthetical text after e.g. a verb phrase, like: "to buy an apple (one euro)"
[09:51:03] <mbrock> after the noun phrase there would be fine too probably
[09:52:57] <freeside> maybe define a new word for_price, and then just in your application do postprocessing to s/for_price (.*? euro)/ ($1)/
[09:53:58] <freeside> i like this postprocessing idea. i have added a hack to my thing and now it works!
[09:54:07] <freeside> > l mkReactionRule repealing MustNot abating
[09:54:11] <freeside> the repealing of the section and laws repealed by this law shall not have the effect of abating , nullifying , suspending or vitiating any proceeding .
[09:54:28] <inariksit> is the hack s/inging/ing/? :-D
[09:54:51] <freeside>     lessing : Str -> Str = \s -> table { x + "vitiating" => x + "vitiate"; _ => s } ! s ;
[09:55:19] <freeside> i should call it lessening
[09:56:07] <freeside> there is probably some more lexically woke way of turning "vitiating" into "vitiate"
[09:56:22] <freeside> i am just doing it as a string-lookup-table
[09:56:32] <inariksit> how come it gets turned into vitiating in the first place in your code? it looks like there's some error there
[09:56:44] <inariksit> in your code or in the english rgl
[09:57:11] <freeside> because i am extracting vpi.s ! RE.VVPresPart ! RE.AgP3Pl RE.Neutr
[09:57:49] <inariksit> okay
[09:57:50] <freeside> see https://github.com/legalese/legalese-compiler/blob/master/sandbox-gf1/missouri/MissouriI.gf#L146
[09:58:15] <inariksit> the type of VPI is { s : VVType => Agr => Str} ;, so one would think that the whole string "xing, ying and zing" would be there
[09:58:24] <inariksit> and not that one of them turns out doubleing
[09:58:40] * inariksit reads it now and yeah
[09:58:41] <inariksit> strange
[09:58:55] <freeside> ah, that is due to the later
[09:58:55] <freeside>  (ExtensionsEng.GerundNP (mkVP
[09:58:55] <freeside>                                           (listvpi2v2 causing_worsening some)
[09:58:55] <freeside>                                           (mkNP some
[09:58:56] <freeside>                                              (mkCN obesity)))))))
[09:59:11] <inariksit> aha yeah!
[09:59:21] <freeside> https://github.com/legalese/legalese-compiler/blob/master/sandbox-gf1/missouri/MissouriI.gf#L193
[09:59:25] <inariksit> hehe yeah so it only takes the last suffix
[09:59:56] <inariksit> *adds the suffix to the end of the string
[10:00:29] <inariksit> so make a new function directly into   gerundVPI : ListVPI -> Det -> NP
[10:01:50] <freeside> well, there are two things going on -- the `mkVP : V2 -> NP` interpolates an "of" in between
[10:02:00] <freeside> if i skip that step i lose the "of"
[10:02:18] <freeside> er, i should have said `mkVP : V2 -> NP -> VP`
[10:03:00] <inariksit> using a hack to add a simple "of" is much less dubious than the previous hack ^^
[10:03:03] <freeside> without mkVP, the adding of the "of" would become "the adding _ the "of""
[10:03:54] <inariksit> how about making effect into N2?
[10:04:17] <inariksit> ComplN2 : N2 -> NP -> CN ;    -- mother of the king
[10:04:35] <freeside> "mothering of the king"
[10:04:36] <inariksit> effect of (the causing_worsening ...)
[10:04:53] <freeside> https://hiddenremote.com/wp-content/blogs.dir/280/files/2015/04/Cersei-and-Joffrey-cersei-lannister-31147145-624-352.png
[10:06:02] <freeside> i need to go learn about complements. brb.
[10:08:41] <inariksit> okay, so that's what you are using already, the second mkCN instance here http://www.grammaticalframework.org/lib/doc/synopsis.html#CN
[10:08:57] <inariksit> so then the of is actually coming from that (it's calling ComplN2 internally)
[10:09:07] <freeside> oh
[10:09:07] <inariksit> so mkVP isn't relevant in that
[10:09:21] <inariksit> so just make gerundVPI : ListVPI -> Det -> NP
[10:09:38] <freeside> even simpler, i should make gerundVPI : VPI -> NP
[10:09:39] <inariksit> and mkCN : N2 -> NP -> CN takes care of placing the of ^^
[10:09:40] <inariksit> yeah
[10:09:57] <freeside> the Conj'ing of the ListVPI belongs outside the Gerunding
[10:11:48] <freeside> the Conj'ing of the ListVPI is belonging outside of the Gerunding
[10:19:14] <mbrock> freeside: ah, yeah good idea
[10:19:18] <mbrock> (the postprocessing)
[10:22:58] <inariksit> by the way, the IRC channel for UD is  #_u-dep
[10:23:02] <inariksit> here on Freenode
[10:29:54] <freeside> cool
[10:30:22] <freeside> http://lindat.mff.cuni.cz/services/udpipe/ parses to a tree that i don't necessarily agree with, semantically
[10:30:48] <freeside> but it does know about gerunds.
[10:31:50] <freeside> btw, maybe abating , nullifying , suspending and vitiating are transitive, not intransitive forms
[10:31:58] <freeside> so should i be Gerunding VPI or some other VP?
[10:32:08] <freeside> remind me -- what is a VPS vs a VPI?
[10:32:56] <freeside> my poor head.
[10:33:35] <inariksit> in the new Extend module, aarne writerrs
[10:33:38] <inariksit> *-rr
[10:33:38] <inariksit> -- $VP$ conjunction, covering both finite and infinitive forms (formerly VPI and VPS).
[10:41:44] *** Quits: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151) (Ping timeout: 260 seconds)
[10:52:12] *** Joins: lillann (uid244962@gateway/web/irccloud.com/x-jtbttimfgskrvtyo)
[11:09:34] <daherb> hi lillann
[11:12:18] <lillann> yello :)
[11:12:26] <lillann> how are tricks?
[11:12:27] *** Joins: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151)
[11:13:07] <lillann> I'll upload my latin grammar in a bit - i got it to compile :P
[11:13:18] <daherb> great
[11:13:58] <daherb> at the moment prasanth is presenting his work on universal dependencies
[11:18:16] <lillann> Ooh, I would have liked to have seen it
[11:18:29] <lillann> I guess JP has left for the airport by now?
[11:18:50] <inariksit> I don't see him here
[11:18:52] <inariksit> so probably yes
[11:19:34] <lillann> good :)
[11:56:24] *** Joins: koom (~sirdancea@109.107.211.244)
[11:57:48] *** Joins: root (~root@174.138.71.241)
[11:57:56] *** root is now known as Guest13080
[11:58:23] *** Quits: Guest13080 (~root@174.138.71.241) (Client Quit)
[11:59:17] *** Joins: ayberkt (~root@174.138.71.241)
[11:59:25] *** Joins: koo5 (~koom@109.107.211.244)
[12:01:43] <ayberkt> I think it would be a good idea to GF development on feature branches and not commit directly onto master.
[12:02:31] <ayberkt> I now have a server by the way, so I will keep connected, hopefully!
[12:03:38] <inariksit> cool
[12:03:46] <ayberkt> It is common practice on GitHub to disable committing directly onto master, and allow merging only on GitHub requiring review of PRs by others.
[12:03:58] <ayberkt> I don't know what the GF developers think about this, though.
[12:04:07] <inariksit> on this channel we have nobody who actually is involved in the core development
[12:04:28] <inariksit> the mailing list (or talking directly to thomas and krasimir) is the best channel to talk about such things
[12:05:50] *** Quits: koom (~sirdancea@109.107.211.244) (Quit: Leaving)
[12:06:05] <inariksit> I think GF developers are not very familiar with github, so if you make your case convincingly enough, they could do the change
[12:54:44] *** Quits: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151) (Ping timeout: 260 seconds)
[14:48:26] <inariksit> I'm going to meet a friend of a friend who is a translator and lives in riga
[14:48:30] <inariksit> today at 17:15 at Ģertrūdes iela 45
[14:48:47] <inariksit> anyone who wants to join is welcome, I told her I'll bring with me 0-∞ linguists
[15:04:16] *** Joins: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151)
[15:43:23] <venicecat> yays! A baby step forward i hope!
[15:43:29] <venicecat> i now know how to make a sentence compl
[16:15:59] *** Quits: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151) (Ping timeout: 260 seconds)
[16:27:15] *** Joins: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151)
[16:33:19] <inariksit> woo!
[16:38:58] *** Joins: livy (57f68898@gateway/web/freenode/ip.87.246.136.152)
[16:45:25] <inariksit> hello livy!
[16:46:10] <inariksit> venicecat, livy, esg, ayberkt, freeside, mbrock, anyone of you interested in coming to meet my translator-friend-of-a-friend? daherb and I are leaving in 15 mins
[16:46:14] <inariksit> or in 14 actually
[16:48:57] <mbrock> I'm preoccupied with some manual labor, but thanks!
[16:52:02] <esg> sure, i'll come along!
[17:58:28] *** Quits: livy (57f68898@gateway/web/freenode/ip.87.246.136.152) (Quit: Page closed)
[18:14:24] *** Quits: venicecat (05b30897@gateway/web/freenode/ip.5.179.8.151) (Ping timeout: 260 seconds)
[18:18:29] *** Quits: lillann (uid244962@gateway/web/irccloud.com/x-jtbttimfgskrvtyo) (Quit: Connection closed for inactivity)
[18:24:50] *** Quits: koo5 (~koom@109.107.211.244) (Ping timeout: 240 seconds)
[18:38:45] *** Joins: venicecat (c30dd254@gateway/web/freenode/ip.195.13.210.84)
[18:49:02] <freeside> tonight, at 10:00pm, i will be serving a late-night supper of fresh, free-range Latvian eggs prepared two ways – deep-fried Thai style in clarified browned butter, and sous-vide poached-in-shell at 64.5º for 1 hour in the manner of Thomas Keller, over a minimalist ratatouille of honey-glazed organic Valmieran farmhouse green and yellow courgettes
[18:49:16] <freeside> if you would like to join us, please indicate accordingly; we are located at the eastern corner of the park, on the corner of Elizabetes and Krišjāņa Barona iela
[18:51:04] <freeside> actually, change of plans. 63.5 degrees, not 64.5
[18:57:02] <freeside> sorry i missed your message earlier, inariksit ... i was busy bothering aarne about ParadigmsEng.mkAdv vs ConstructorsEng.mkAdv
[19:10:33] *** Joins: venicecat_ (c30dd254@gateway/web/freenode/ip.195.13.210.84)
[19:11:34] *** Quits: venicecat (c30dd254@gateway/web/freenode/ip.195.13.210.84) (Ping timeout: 260 seconds)
[19:29:04] *** Quits: venicecat_ (c30dd254@gateway/web/freenode/ip.195.13.210.84) (Ping timeout: 260 seconds)
[19:33:51] *** Joins: proteusguy (~proteus-g@2405:9800:b408:bc31:5cb3:2d60:b355:f07c)
[20:40:31] *** Joins: jstar (~jstar@unaffiliated/jstar)
[20:40:48] <freeside> is there some notion of selectivly omitting these words?
[20:41:48] <freeside> i'm guessing the thing I don't know enough about is a Slash or a Complement or a Relative Something
[20:50:32] *** Joins: jstar (~jstar@unaffiliated/jstar)
[21:01:38] *** Quits: proteusguy (~proteus-g@2405:9800:b408:bc31:5cb3:2d60:b355:f07c) (Ping timeout: 246 seconds)
[21:03:14] *** Joins: proteusguy (~proteus-g@2405:9800:b408:bc31:5cb3:2d60:b355:f07c)
[21:11:43] <freeside> okay, i have learned about Advs.
[21:11:50] <freeside> next question.
[21:12:59] <freeside> "i like my food grilled."
[21:13:00] <freeside> mkUtt ( mkS ( mkCl i_NP like_V2 ( mkNP ( mkNP ( mkQuant i_Pron ) food_N ) grill_V2 ) ) )
[21:13:16] <freeside> "i like my food roasted."
[21:13:17] <freeside> mkUtt ( mkS ( mkCl i_NP like_V2 ( mkNP ( mkNP ( mkQuant i_Pron ) food_N ) roast_V2 ) ) )
[21:14:08] <freeside> "I like my food both grilled and roasted"
[21:14:13] <freeside> "I like my food either grilled or roasted"
[21:14:21] <freeside> p -lang=Eng "I like my food grilled or roasted"
[21:14:27] <freeside> The parser failed at token 7: "roasted"
[21:29:23] <inariksit> freeside: I'm kinda drunk and considering joining
[21:29:37] <inariksit> I had dinner already
[21:29:58] <inariksit> but your dinner description sounds like from a grammar
[21:32:41] <inariksit> but now I'm doing my daily duolingo
[21:32:46] <inariksit> with lessons about bahtroom and spirituality
[21:32:48] <inariksit> bathroom*
[21:32:58] <inariksit> ooh wow I got a romance lesson too!
[21:33:15] <inariksit> now I will be able to say "deine augen sind wie sterne" and "zu mir oder zu dir" in dutch too ^^
[21:33:17] <inariksit> thanks duolingo
[21:34:28] <inariksit> "she likes to hug her pillow"
[21:35:26] <inariksit> "I am straight"
[21:37:07] <inariksit> ok at least now I got "I am gay"
[21:52:21] *** Joins: koo5 (~koom@79.98.72.195)
[21:52:25] <ayberkt> How do I write tests for a resource grammar?
[21:54:43] <inariksit> you can use something like this https://github.com/GrammaticalFramework/gf-contrib/blob/master/testsuite/resource.gfs
[21:55:13] <inariksit> write the output it produces into a file, manually correct all the sentences, and use that as a gold standard
[21:55:26] <inariksit> and everytime you change the grammar, run the test again and diff with the gold standard
[21:55:38] <inariksit> or even without a gold standard, it can still show you if something changes
[21:55:55] <inariksit> of course these trees are just a random selection, and some of them are not super relevant and some more relevant things are missing
[21:56:17] <inariksit> I just have for myself different small scripts, like one for relative clauses, one for a few transitive verbs, etc.
[21:56:37] <inariksit> I'm afraid there's no unified best practice for it
[21:57:02] <inariksit> so a file with .gfs ending, you use it like
[21:57:07] <inariksit> gf < MyTestFile.gfs
[21:57:15] <inariksit> and it outputs the linearisations
[22:09:31] <inariksit> freeside: I'm going to sleep now, enjoy your dinner! ^^
[22:10:26] <ayberkt> inariksit: thanks, this helps a lot!
[22:11:28] <ayberkt> It would be nice to have an `assertLinearization` command so we would not have to diff things manually
[22:11:58] <ayberkt> It shouldn't be really hard to implement, right?
[22:25:54] <ayberkt> inariksit: https://github.com/GrammaticalFramework/GF/issues/10
[23:15:45] <freeside> goodnight, sorry we missed you! i was occupied in the kitchen.
[23:24:32] *** Joins: venicecat (6d6e1376@gateway/web/freenode/ip.109.110.19.118)