Search for a user lexicon is conducted in the directories specified by the flag --path; if the flag is not set, then in the directories specified by the environment variable PATH. The directories are searched in the order they are listed. First, the user lexicon called et.usr.dct is looked for, then et.usr.dct.utf8, and finally et.usr.dct.uc. Only the first found lexicon is actually used.
User lexicon is a text file.
Lines starting with symbol # are comments, i.e. they are ignored by the program.
Other lines follow the format of the output of the morphological analyzer vmeta.
NB! An error in number or position of space characters, commas or slashes will result in abortion of the program and an error message.
The encoding depends on the name of the user lexicon:
- et.usr.dct - Windows Baltic.
- et.usr.dct.utf8 - UTF-8.
- et.usr.dct.uc - little endian two-byte unicode.
Let the current directory contain a two-word lexicon et.usr.dct.utf8:
l@xid mine+sid //_V_ sid, //
r66mu rõõm+0 //_S_ adt, // rõõm+0 //_S_ sg p, // rõõm+0 //_S_ sg g, //
... then the morphological analysis with vmeta will result in:
echo '<s> R66mu rullid l@xid koju . </s>' | vmeta -p .:$PATH
<s>
R66mu rõõm+0 //_S_ adt, // rõõm+0 //_S_ sg g, // rõõm+0 //_S_ sg p, //
rullid rull+d //_S_ pl n, // rulli+d //_V_ d, //
l@xid mine+sid //_V_ sid, //
koju kodu+0 //_S_ adt, //
. . //_Z_ //
</s>
... and morphological analysis with vmeta, followed by disambiguation with vmety with output set to GT-schema, will result in:
echo '<s> R66mu rullid l@xid koju . </s>' | vmeta -p .:$PATH | vmety --gt
<s>
R66mu rõõm+0 //_S_ Sg Gen, //
rullid rull+d //_S_ Pl Nom, //
l@xid mine+sid //_V_ Pers Prt Ind Pl3 Aff, // mine+sid //_V_ Pers Prt Ind Sg2 Aff, //
koju kodu+0 //_S_ Sg Ill, //
. . //_Z_ //
</s>
Commandline morphological analyzer vmeta.
Copyright © 1997-2022 OÜ Filosoft