MoA's usage
Date : 95.8.13 (Sun), bgjang
This page contains full explanations about MoA itself and the other
utilities provided. For short brief help message,
type executables without options.
moa
Morphological Analyzer. It needs 4 kinds dictionary and 2 tables to be
executed. If anyone is not supplied, then it will say an error and die.
- N-Type dictionary :
a dictionary which contains normal words and its POS and frequency.
- E-Type dictionary :
a dictionary which contains exceptional words or words used in
contraction recovery.
- I-Type dictionary :
a dictionary which contains words can have a irregular form.
- U-Type dictionary :
a dictionary which contains word information used in unknown
word processing step.
- cTable : connectivity table
for normal word searching.
- mTable : connectivity table
for unknown word searching.
The dictionary/table locations is determined by 4 environment variables.
It must be set to execute moa. For example, see
supplied env-moa file.
-
MOA_DICTDIR : directory in which dictionary files are.
In this directory, the above 4 kind dictionary must be located.
In distributed MoA, there is 4 dictionary in dict-test
and dict-datum.
-
MOA_TABLEDIR : directory in which table files are.
In this directory, the above 2 kind table must be localted.
In distributed MoA, only one table TRANTBL is
supplied currently. And one table is used as cTable and mTable
without any difference. In future, I will make two tables and
two tables can be different.
-
MOA_CTABLE : cTable's file name. It must be located in
MOA_TABLEDIR directory.
-
MOA_MTABLE : mTable's file name. It must be located in
MOA_TABLEDIR directory.
And the other two environment variable can be set.
-
MOA_CACHE : must be '0' or '1'. When it is set to
'0', then memory-cache for N-Type
dictionary is not used. When it is set to '1' or is not set,
use memory-cache.
-
MOA_LOGDIR : directory in which the log file is
generated in the case LOG_STATISTIC is defined
when compiling.
If the above necessary 4 environment variables are set, then
moa can be executed. The options are like the followings :
-c or -f data-file : Where input comes from ?
-c means input is from standard input till EOF. And
-f data-file means that moa read data from
data-file line by line and processing read line until
EOF.
-h : for brief help message
-n dict-name : N-Type dictionary, default name is 'nn'.
-e dict-name : E-Type dictionary, default name is 'ee'.
-i dict-name : I-Type dictionary, default name is 'ii'.
-u dict-name : U-Type dictionary, default name is 'uu'.
chkcon : Connectivity Checker
datum/TRANTBL contains connectivity
informations but is not human-readible. So to change this file or
lookup the informations in this file, use chkcon utility.
You can simply type chkcon to get help message.
dictman : Dictionary Manager
It is used to update the contents in 4-type
dictionary.
Home of MoA
Byoung-Gyu, Chang / bgjang@csone.kaist.ac.kr