LanguageTool Step

From Okapi Framework
Jump to: navigation, search

Overview

This step calls the LanguageTool style, grammar and spelling checker and adds annotations for the issues found. Optionally bilingual terms and black terms can be checked using LanguageTool's morphological analyzers.

LanguageTool is an open source proofreading software for English, French, German, Polish, Romanian, and many more other languages. It finds many errors that a simple spell checker cannot detect like mixing up there/their and it detects some grammar problems.

Note: This step is distributed as a plug-in and is not included in the general distribution.

This step is only available as part of the snapshots for now:

To install the plugin:

With Rainbow, one way to install it the following:

  • Start Rainbow
  • Select Tools > Plugins Manager
  • In the field under Plugins available for installation enter the URL of the release you want to install (see above)
  • Click Refresh
  • Select the plugin, its name starts with okapiLanguageTool
  • Click Install Checked Plugins

Parameters

Check also the source text — Set this option to check the source content in addition to the target content.

Check for false friends — Set this option to check for false friends when possible.

Check for spelling mistakes — Set this option to check for spelling errors when possible.

Check for bilingual terms — Set this option to check for that the source term is translated with the corresponding target term.

Bilingual terms — Set this option to set the path to the file (csv, tab or tbx) containing the bilingual terms.

Check for black terms — Set this option to check for that the target does not contain black terms.

Black terms — Set this option to set the path to the file (tab) containing the black terms.

Limitations

  • Term and black term checking use the stemmers provided by LanguageTool. Not all languages are supported. The default is to match the whole term, without stemming.
  • The LanguageTool Java API does not support inline codes yet.