El 2017-07-01 09:18, Kevin Brubeck Unhammer escribió:
Jaume Ortolà i Font
<[email protected]> čálii:
Hi,
I need to match a Spanish contraction ("del" or "al") in Constraint
Grammar, but I am unable to do it.
The contraction is:
^del/de<pr>+el<det><def><m><sg>$
The only think that seems to be visible in the Constraint Grammar
syntax is the second element of the
contraction:
el<det><def><m><sg>
It matches ("el") or (det def m sg), but not ("de"), ("del") or (pr).
Does anybody know how to match the first element of the contraction
(the preposition)?
Yes, with subreadings. See all of chapter 18 of the manual:
https://visl.sdu.dk/cg3/chunked/subreadings.html
E.g. if the lexical unit at position 0 is
^del/de<pr>+el<det><def><m><sg>$, then all of these context conditions
should match:
(0/1 ("de"))
(0/-1 ("de"))
(0/0 ("el"))
(0 ("el"))
(0/* ("el"))
(0/* ("de"))
Also (0 ("<del>")) of course too :)
This could be solved differently. I think these contractions should be
tokenized earlier in the pipeline as two
tokens. This way we would avoid a lot of exceptions and workarounds
when dealing with them. Is it feasible?
These contractions are extremely frequent and now they cause a lot of
undesired results.
Yeah, you could split them early too. If "del" isn't ambiguous then you
don't really gain much by keeping it as one lexical unit.
How would you do that ? :)
F.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
