UIT The arctic university of Norway > Giellatekno
 

Getting Started With The New Infra

Installation and setup

The required steps are described in our Getting Started guide

Check out the source code

Everything

If you are working with many languages, or with both linguistic analysis and e.g. language learning (Oahpa) or dictionaries, you want to check out everything. That process is documented on a separate page.

Only the GT core and the wanted language(s)

If you only do linguistic analysis for a limited number of languages you will do fine by only checking out the core modules, and then each language separately. It will save you a lot of download time and disk space.

It can be wise to create a separate directory for holding the core and language modules. The command to do this is (with the name of the directory set to "langtech"): mkdir langtech. Then cd langtech and continue.

There are two required core modules:

giella-core
core scripts and technical data required for the build system
giella-shared
resources shared among all languages or language groups

Check out the core modules

To check out the core modules and prepare them, do this:

svn co https://gtsvn.uit.no/langtech/trunk/giella-core
svn co https://gtsvn.uit.no/langtech/trunk/giella-shared
cd giella-core
./autogen.sh
./configure
make
cd ../giella-shared
./autogen.sh
./configure
make

To make the core modules accessible, there are two methods: one intended for single users, and one intended for multiuser servers (but it works fine for single users as well).

EITHER: The first method uses environmental variables, one called GIELLA_CORE and the other GIELLA_SHARED. They must be manually set. Add the following line to your .profile file:

export GIELLA_CORE=/path/to/your/giella-core/dir
export GIELLA_SHARED=/path/to/your/giella-shared/dir

Replace /path/to/your/giella-core/dir with the real path on your system, and the same for /path/to/your/giella-shared/dir. You can get them by using the command pwd after you did the ./configure command above. Then open a new terminal window. Now you are done.

OR: The server method is to install some of the components of the core modules. For each module, after you did make above, do:

sudo make install

That is it. NB! This method requires that you redo the installation every time the core modules are updated, although should not happen too often.

Check out the languages you want

Then, for each language you want to work with, do the following:

svn co https://gtsvn.uit.no/langtech/trunk/langs/LANGCODE --username <your_username>

where LANGCODE is the ISO 639-3 three-letter code for the language. You can see a list of the languages we currently have by looking at https://gtsvn.uit.no/langtech/trunk/langs/.

When you have checked out the language(s) you want to work with, then you are ready to...

Actually getting started with your language

When everything is installed and checked out, then do the following:

cd $GTLANG
./autogen.sh -l
./configure

where $GTLANG is the directory name of the language you want to work with. Now you are ready to start working. More info about where to find the different pieces of source code can be found on this page.

To build the transducers and other tools for linguistic analysis, egdo:

make

To run the preinstalled tests, do:

make check

WARNING

You may encounter troubles with your CLASSPATH. We are working on it, here is a fix if the compiler complains it is not set: In the langs directory, write export CLASSPATH=.

Happy linguistic coding!