README

Tokenizer for indic languages.

Tokenizer convert a sentence into word level tokens and return sentence marker for each sentence of input text.

Requirements:

Operating System	: LINUX/UNIX System
Compiler/Interpreter	: perl

For installation on Linux, please refer to the file INSTALL.

How to use ??

perl  tokenizer_indic.pl --lang=hin --input=input-file --output=output-file

-l, --lang=[ hin | tel|... ]	: select the language 3 letter code (ISO-639)
-s, --str_input=<input-string>	: give input string
-i, --input=<input-file>	: give input file
-o, --output=<output-file>	: give output file
-j, --jflag=[yes|no]		: give input to print -JOIN in between multiwords. default is yes
-h --help			: print the detail useage

e.g.

perl tokenizer_indic.pl --lang=hin --input=tests/input_hin_utf.txt

where input_hin.txt is the text input file which need to be tokenized
output of above command is STDOUT and can be redirect into output file.

NOTE: For better result use UTF input to tokenizer.

Directory Structure:

tokenizer-indic
     |
     |---lib (source code of the convertor library)
     |
     |---tests (contains the referenece input and output)
     |
     |---doc (documentaion)
     |
     |---data (acronym data file)
     |
     |---tokenizer_indic.pl (main file)
     |
     |---INSTALL (How to install the module)
     |
     |---README (How to run/use the module)
     |
     |---ChangeLog (version information)
     |
     |---ISSUES.txt (listed issues still persist)
     |
     |---tokenizer_indic_run.sh (How run using shell script)


Contact :
Rashid Ahmad
Expert Software Ltd.
rashid101b@gmail.com