Installing existing Apertium data

From LING073
Jump to: navigation, search

Installing existing Apertium data is pretty simple. The following example walks you through the installation and testing of apertium-srd-ita, an Italian-to-Sardinian translator.

These instructions assume you have the Apertium core tools installed. All lines starting with $ are meant to be run on a linux shell / command line. Usually you can just copy/paste or type everything after the $ symbol and hit enter.

You'll want to start in a fresh directory appropriate for Ling 073 materials, e.g. ~/ling073/.


The easy way

The following command should get and compile the data.

 $ apertium-get srd-ita

If it works, you can skip the next two sections and skip ahead to Testing.

Get the data

Get the monolingual modules. The --depth=1 part is optional, but will ensure you only get the most recent version of the code and not the entire history of revisions.

 $ git clone --depth=1
 $ git clone --depth=1

Get the translation module:

 $ git clone --depth=1

Each one of these commands should show that it's retrieving a repository, and should return you to the prompt when done.

Compiling the monolingual modules

$ cd apertium-ita
$ ./
$ make

The and make commands should both show a certain amount of output, and make take a minute to complete. There may be a couple "Warning" statements, but there should be no "Error" statements.

You'll need to cd ../ to get back to the original directory before continuing. Then repeat, but with the other module:

$ cd apertium-srd
$ ./
$ make

Compiling the translation module

$ cd apertium-srd-ita
$ ./ --with-lang1=../apertium-srd --with-lang2=../apertium-ita
$ make

As before, the autogen and make lines should each return a certain amount of output, and may take a minute to run.


If you run the following from within your apertium-srd-ita/ directory (or replace . in the command with the path to that directory),

$ echo "Il secchio era da mettere in quell'angolo." | apertium -d . ita-srd

then you should should get output something like the following:

Su puale fiat in pònnere in cuddu àngulu.

Some non-Sardinian-like output could mean a number of things. If you get an error message, you may have run the command wrong, or you may have screwed up a step in the installation process. Let the instructor or course assistant know if you get something different.