posegment takes a Gettext PO or XLIFF file and segments the entries, generating a new file with revised and smaller translation units.

This is useful for the creation of a file that can be used as a Translation Memory as you should get better matching after you have exposed translated sentences that might occur elsewhere in your work.

Posegment won’t do very advanced sentence boundary detection and alignment, but has customisations for the punctuation rules of several languages (Amharic, Afrikaans, Arabic, Armenian, Chinese, Greek, Japanese, Khmer, Oriya, Persian). For the purpose of increasing your TM (as described below), it is already very useful. Give it a try and help us to improve it even more for your language.


posegment [options] <input> <segmented>



translations to be segmented


translations segmented at the sentence level



show program’s version number and exit

-h, --help

show this help message and exit


output a manpage based on the help


show progress as: dots, none, bar, names, verbose


show errorlevel as: none, message, exception, traceback

-i INPUT, --input=INPUT

read from INPUT in po, pot, tmx, xlf formats

-x EXCLUDE, --exclude=EXCLUDE

exclude names matching EXCLUDE from input paths

-o OUTPUT, --output=OUTPUT

write to OUTPUT in po, pot, tmx, xlf formats

-S, --timestamp

skip conversion if the output file has newer timestamp

-P, --pot

output PO Templates (.pot) rather than PO files (.po)

-l LANG, --language=LANG

the target language code


the source language code (default ‘en’)


Disable automatic stripping of whitespace


Removes units where sentence number does not correspond


You want to reuse all of your Pidgin translations in another Instant Messenger:

posegment pidgin-af.po pidgin-af-segmented.po

Now all of our Pidgin translation are available, segmented at a sentence level, to be used as a Translation Memory for our other translation work.

You can do the same at a project level. Here we want to segment all of our OpenOffice.org translation work, a few hundred files:

posegment af/ af-segmented/

We start with all our files in af which are now duplicated in af-segmented except files are now fully segmented.


  • If the toolkit doesn’t have segmentation rules for your language then it will default to English which might be incorrect.

  • Segmentation does not guarantee reuse as your TM software needs to know how to segment when matching. If you use software that doesn’t do segmentation, you can consider joining the original and the segmented files together with msgcat, to get the best of both worlds.

  • You cannot (yet) use the tool to break a file into segments, translate, and then recreate as the segmented file does not know which parts should be joined together to recreate a file.