apertium-get

This is a little script to quickly get an Apertium language pair (with its monolingual dependencies) set up for development. It’s meant for developers and people interested in hacking on the language data; if you just want to run the newest translators you should use the nightly repositories.

Prerequisites

You’ll need the newest Apertium development dependencies installed first, follow the “Prerequisites” step at https://wiki.apertium.org/wiki/Installation but just skip installing any language data in the “Minimal installation from SVN” step – that’s what this script handles.

This should give you the “core tools”, ie.:

• apertium
• lttoolbox
• apertium-lex-tools
• apertium-separable
• apertium-recursive
• apertium-anaphora
• vislcg3
• hfst

Not all language pairs depend on all the core tools, but since you can get core tools from the nightly repos, it’s easiest to just get them all at once.

Usage

Assuming you have the core tools installed, you can get the pair “fie-bar” along with its data dependencies apertium-fie and apertium-bar downloaded and compiled by simply doing

wget https://raw.githubusercontent.com/apertium/apertium-get/master/apertium-get.py -O apertium-get
chmod +x apertium-get
./apertium-get fie-bar

The data will be placed under your current working directory. If you later run apertium-get fie-fum from the same directory, it’ll just update apertium-fie, not redownload it (so you can actually use this script as a simple way to do svn up && make for a pair and its dependencies as well).

You can also get a listing of available pairs by passing the -l argument to apertium-get, see apertium-get -h for more information.

If you installed a monolingual dependency through apt-get or similar (or otherwise have it installed in your PKG_CONFIG_PATH), then you can choose to skip that dependency and use the installed version with -x DEP. For example, to avoid compiling the sme package before getting apertium-sme-nob, do:

sudo apt-get install giella-sme
./apertium-get -x sme sme-nob

You can also specify a git clone depth with d DEPTH, since some of these git checkouts can be quite big and slow, e.g.:

./apertium-get -d 1 nno-nob

Giellatekno data

If you have the standard Giellatekno data setup with the SVN trunk checkout in a GTHOME variable, then apertium-get will try to reuse data from your $GTHOME, which should make compilation faster. You should be able to run

./apertium-get myv-mdf

and have it use the gtcore, langs/myv and langs/mdf from your GTHOME SVN checkout.

Using apertium-get without GTHOME

If GTHOME is unset, then apertium-get will download and compile Giellatekno data just like Apertium data.

If you have GTHOME, but want to force apertium-get to not use it for whatever reason, simply run the program with the variable emptied, e.g.

GTHOME= ./apertium-get myv-mdf