Okapi Filters Plugin for OmegaT

From OkapiWiki

Jump to: navigation, search

Contents

Overview

OmegaT is a free and open-source translation tool that offers support for many file formats. It also provides a plugin mechanism to use addition filters.

Several of the Okapi filters have been packaged into a plugin that works with OmegaT's plugin interface. This allows you to use the filters seamlessly directly from OmegaT.

Filters Included

Currently support for the following formats are included:


Warning: Note that the TTX filter is set by default to auto-detect pre-segmented files.
  • If no segments are detected, the filter extract all text by creating its own TTX segmentation.
  • If one or more segments are detected only the existing segments are passed to OmegaT. So if a file is only half segmented you will not get the un-segmented text in OmegaT. In those cases you can:
    • define your own filter settings file for TTX
    • use Rainbow to create an OmegaT project where the TTX filter is forced to extract the non-segmented text.

Note that several of the formats supported by the plug-in are also supported by OmegaT native filters. You should select which filter to use by enable/disable it in the File Filters dialog. If several filters are set for a given format, the first one in the list is used by default.

Filter Parameters

Starting in m24 you can specify a custom filter parameters file for each of the filter that supports options.

Use OmegaT's Options button in the File Filters dialog box to select whether you want to use the default settings, or a custom filter parameters file (.fprm extension) where you have stored your options.

You cannot create or edit the filter parameters file from OmegaT, but you can use Rainbow to do this (menu Tools > Filter Configurations).

Warning: All filter parameters files you use in OmegaT must be in the same directory.

Download and Installation

Download the file okapi-pluginForOmegaT_all-platforms_<version>.zip from:

To install the plugin:

Macintosh

On the Macintosh platform, the main OmegaT directory is OmegaT.app/Contents/Resources/Java. You can open the directory by right-clicking OmegaT.app and selecting the command Show Package Content.

You may also have to change the default version of the Java virtual machine used to run OmegaT. To do this, open the file OmegaT.app/Contents/info.plist and replace:

<key>JVMVersion</key>
<string>1.5</string>

by:

<key>JVMVersion</key>
<string>1.7</string>

Segmentation

For the file formats that represent segments, such as TTX, be aware that the segmentation created by OmegaT is not carried back into the translated document. For example, an unsegmented paragraph of two sentences may be translated as two separate segments in OmegaT (and produce two TM entries), but it is merged back as a single paragraph (between segment markers because that is the only way to store translation) in the translated TTX file.

You can use the Segmentation Step in Rainbow, or the Segmentation command of Tikal to create a pre-segmented TTX file before opening it in OmegaT.

Note that any line-break in the source text is considered a segment break by Trados TagEditor, even when it is within an existing segment. Opening a segment that includes a line-break with TagEditor results in a segment withing segment.

Pre-Translation

TTX documents may contain segments that are already translated. The translation of such segments is loaded as the current translation in OmegaT.

Note that the target language of the OmegaT project must match the target language specified in the TTX file. The target language of a TTX file is defined in the attribute TargetLanguage of the <UserSettings> element.

Testing a Filter

Some file formats are difficult to extract and merge. If you want to be sure that the translated file merges back properly and is a valid file. One step toward verifying this is to re-extract the merged file and compare the first extraction with the second.

Support for XLIFF 2

Note: This feature is available in the M25-SNAPSHOT only and is experimental.

You can find examples of valid XLIFF 2 documents in the SVN repository of the XLIFF TC

Personal tools
Namespaces
Variants
Actions
Navigation
Toolbox