--- Behdad Esfahbod [EMAIL PROTECTED] wrote:
Can you please educate us on how these are supposed
to work? I
can't get anything out of them. I choose UTF-8, and
type a verb
in the stemmer, I get back the verb verbatim.
Sorry about the late reply. The perl script is run
from the command-line, taking input from STDIN and
outputing to STDOUT under Unix/Linux/Cygwin.
I just finished version 0.7, which natively supports
input and output to from UTF-8, CP-1256 (aka.
Windows-1256), and ISIRI 3342. The default input
output is romanized text. This version also fixes
some bugs (eg. --root). A tentative name for the
stemmer is Perstem. If any of you think of a better
name, please let me know. Also, I'll try to get the
web page version updated sometime soon, which
hopefully will fix the problem that Behdad mentioned.
http://students.cs.byu.edu/~jonsafar/perstem
Sample usage might include (after removing the single
quotation marks around commands):
Input a UTF-8 webpage and output to CP-1256,
preserving only the roots of words, and remove HTML
tags:
'perstem -i utf8 -o cp1256 --root --noroman
my_utf8.html my_cp1256.txt'
Input romanized sentence from the command-line, output
to UTF-8, show the morphological links, remove short
vowels, and tokenize punctuation:
'echo man ketAb-hAie tu rA nemi-binam. | perstem -o
utf8 --links --unvowel --tokenize my_utf8.txt'
For a full list of commands, try the -h or --help
option.
The stemmer and syntax parser were well recieved last
week at the First International Conference on Aspects
of Iranian Linguistics in Leipzig, Germany.
Thanks for all your feedback so far,
-Jon Dehdari
__
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
___
PersianComputing mailing list
PersianComputing@lists.sharif.edu
http://lists.sharif.edu/mailman/listinfo/persiancomputing
