Matxin linearisation with the statistical text generator

From Apertium
Jump to navigation Jump to search


You will need Apache Ant, the Java Development Kit (JDK), and Java. On Ubuntu or any other Debian-based operating system, you can simply install the ant package from APT, and APT will automatically install Java and the JDK.


First, clone the repository.

$ git clone

Then, simply run Ant.

$ ant

Ant will compile all the classes to the build directory and jar them to dist/lib/tg.jar. Ant can also generate the documentation to the doc directory.

$ ant doc

However, very little of the code is documented, so this is not particularly useful.


We also use Ant to run the project's programs. Ant, however, cannot directory forward command-line arguments to programs, so we use a simple wrapper script.

./ [CLASS] [ARGS]...

CLASS is the name of the main class—that is, the class that contains the main function that you want to execute. The programs themselves do not appear to be very well-documented, so I have simply compiled a list of all possible main classes:

  • commata.Evaluator
  • commata.Tagger
  • dsynt2synt.Convert
  • dsynt2synt.Evaluator
  • dsynt2synt.Mapper
  • dsynt2synt2.Convert
  • dsynt2synt2.Evaluator
  • dsynt2synt2.Mapper
  • dsynt2syntP.Convert
  • dsynt2syntP.Evaluator
  • dsynt2syntP.Mapper
  • morph2.Main
  • morph2.StringEdit
  • move.Main
  • sem2syn.DSyntConverter
  • sem2syn.Main
  • sem2syn.SemConverter
  • sem2syn2.DSyntConverter
  • sem2syn2.Main
  • sem2syn2.SemConverter
  • syn2lin2.Main
  • syn2lin2b.Main
  • syn2lin3.Lin
  • syn2lin4.Main
  • syn2lin5.Main
  • syn2lin6.Main
  • syn2lin6b.Main
  • syn2lin6c.Main
  • syn2lin7.Main
  • tag7.Evaluator
  • tag7.Tagger
  • util.ConvertDS
  • util.ConvertSSynt
  • util.Converter
  • util.Evaluate
  • util.Evaluator