lt-comp —
augmented letter transducer compiler for
Apertium
| lt-comp |
[-a |
-v | -l |
-r | -m |
-h] lr |
rl
dictionary_file
output_file
[acx_file] |
lt-comp is the application responsible for
compiling dictionaries used by
lt-proc(1) in
Apertium into a compact and efficient representation (a class of finite-state
transducers called augmented letter transducers).
-
-a,
--alt
- Sets the value of the alt
attribute to use in compilation.
Note that if no value is set, all entries containing an alt attribute
are omitted.
-
-v,
--var
- Sets the value of the v
attribute to use in compilation. This should only be used with monodixes;
for bidixes, see -l and
-r.
Note that if no value is set, all entries containing a
v attribute are considered to be
left-to-right.
-
-l,
--var-left
- Sets the value of the vl
attribute for use in compilation of bidixes. “Left” here
refers to the side of the dictionary, so this option is only valid in
rl mode.
-
-r,
--var-right
- Sets the value of the vr
attribute for use in compilation of bidixes. “Right” here
refers to the side of the dictionary, so this option is only valid in
lr mode.
-
-m,
--keep-boundaries
- Keep any morpheme boundaries defined by the '<m/>'
symbol
-
-H,
--hfst
- expect HFST symbols
-
-S,
--no-split
- don't attempt to split into word and punctuation
transducers
-
-j,
--jobs
- Parallelise minimisation by using one cpu core per section.
By default, this also creates a new section after 50.000 entries. You can
override this number by setting the environment variable
LT_MAX_SECTION_ENTRIES to some number. If set to 0, sections are never
split (but kept exactly as in the dix file). You can also set the
environment variable LT_JOBS=true if you always want parallel minimisation
even if lt-comp was called without this option.
-
-h,
--help
- Prints a short help message.
- lr
- The resulting transducer will process dictionary entries
left-to-right.
- rl
- The resulting transducer will process dictionary entries
right-to-left.
- dictionary_file
- The input dictionary.
- output_file
- The compiled dictionary (a finite state transducer).
- acx_file
- Optional XML file of equivalent characters in
monodices.
apertium(1),
apertium-tagger(1),
lt-expand(1),
lt-proc(1)
Copyright © 2005, 2006 Universitat d'Alacant / Universidad de Alicante.
This is free software. You may redistribute copies of it under the terms of
the GNU
General Public License.
Many... lurking in the dark and waiting for you!