Freeling tags for cerca de
olzama opened this issue · comments
The word cerca should often be tagged (by Freeling) as a preposition but it looks like it never is? No successful examples of that in the corpus either. I am not trying to fix it now because it involved retokenization (with respect to the de in cerca de).
Do you mean “cerca de” should be a MWE?
Yes, I think so. It should be a token with a P tag.
SRG lexical entry:
cerca_de_p-np-nm := p_np_i-nm_native_le &
[ STEM < "cerca_de" >,
SYNSEM.LKEYS.KEYREL.PRED _cerca_de_p_rel ].
But Freeling returns:
Cerca de la mesa.
Cerca cerca RG 0.995192
de de SP 0.999961
la el DA0FS0 0.98926
mesa mesa NCFS000 0.992395
. . Fp 1
This is fixed easily by adding cerca_de to the locutions.dat file in Freeling.