The Divvun and Giellatekno teams build language technology aimed at minority and indigenous languages
View GiellaLT on GitHub divvungiellatekno/giellalt.uit.no
Giellatekno-møte 24.6.
fst:
Vi hadde hatt diskusjon i Canada, m.a. med Måns Huldén, om dette.
I staden for referat legg vi her ut skisse frå diskusjonen:
--------------
samis sápmi+Hum+N+Sg+Loc+AErr
L sápmi+N+Sg+Loc +AErr !!
sápmi^WGs
T sápmi^WGs
sámis
S sámis á (->) a ;
samis
--------------
define Lex {ääliö} ("+MIS":"^") ; <--- lexc +AErr:+AErr
def Rule1 ä (->) a X , ö (->) o X | _ ?* "^" ;
def Rule2 "^" => X ?* _ ;
def Rule3 X -> 0 .o. "^" -> 0;
regex Lex .o. Rule1 .o. Rule2 .o. Rule3;
------------------------
heile sme-lexc + kopi.sme-lex med ("+MIS":"^");
define Lex lexicon.xfst ("+MIS":"^")
define
business as usual [sme.lexc kopi&+MIS] .o. sme.bin --> sme.fst
Rule 123
------------------------
heile sme-lexc med nytt ENDLEX:
LEXICON ENDLEX +MIS:%^ # ; # ;
def Rule1 ä (->) a X , ö (->) o X | _ ?* "^" ;
def Rule2 "^" => X ?* _ ;
def Rule3 X -> 0 .o. "^" -> 0;
regex lexicon.xfst .o. Rule1 .o. Rule2 .o. Rule3;
------------------------
foma
source aalio.foma
up aalio
ääliö+MIS
up ääliö
ääliö
re operator: => context restriction, e.g.
A => B _ C, D _ E
"LangA"
á:a => _ :* ( %>: ) ( »: ) ( »: »: »: ) %+AErr:0 ;
Cns:+ _ ( %>: ) ( »: ) ( »: »: »: ) (¤:) (StemCns:*) ( %>: ) ( »: ) ( »: »: »: ) X7: ;
"OnlyA"
%+AErr:0 => á:a :* _ ;
--------------
guesser
foma
source guesstest.foma
---------------------------------
guesstest.foma
------------
def C [b|c|d|f|g|h|k|l|m|n|p|q|r|s|t|v|w|x|z](%@);
def V [a|e|i|o](u);
def PhonWord [C* V (V) C*]+;
read lexc guesstest.lexc
substitute defined PhonWord for "^GUESSNOUNSTEM"
def Grammar;
def Phonology [..] -> e | [s h | c h | s | x ]( z) _ s .#. ;
def Grammar Grammar .o. Phonology ;
def GuessGrammar $[ "+GUESS" ] .o. Grammar ;
def CleanGrammar ~$[ "+GUESS" ] .o. Grammar ;
regex CleanGrammar .p. GuessGrammar ; # Lower-side priority union
----------------------
guesstest.lexc:
----
Multichar_Symbols +N +Sg +Pl ^GUESSNOUNSTEM +GUESS
LEXICON Root
Nouns ;
LEXICON Nouns
cat NInfl ;
dog NInfl ;
mouse NInfl ;
horse NInfl ;
^GUESSNOUNSTEM+GUESS:^GUESSNOUNSTEM NInfl ;
LEXICON NInfl
+N:0 Infl;
LEXICON Infl
+Pl:s #;
+Sg: #;
---------------------
apply up> horse
horse+N+Sg
apply up> blabla
blabla+GUESS+N+Sg
apply up> blablas
blablas+GUESS+N+Sg
blabla+GUESS+N+Pl
apply up> blablass
???
apply up> blablases
blablases+GUESS+N+Sg
blablase+GUESS+N+Pl
blablas+GUESS+N+Pl
blablass+GUESS+N+Sg
apply up>
---------------------
giellabeassi
giellabeassi girji+N+SgNomCmp+Cmp#busse+N+G3+Sg+Nom = 1
giellabeassi girjebusse+N+G3+Sg+Nom = 0
==> 0 er mindre enn 1
a. giellabeassi
b. giella#beassi
a - b = the winner: giellabeassi
a.
giellabeassi <========
g#i#ellabeassi
giella#beassi
...
b.
giella#beassi
g#iella#beassi
...