org.omegat.core.segmentation

Class Segmenter

java.lang.Object

org.omegat.core.segmentation.Segmenter

public final class Segmenter
extends java.lang.Object

The class that sentences the paragraphs into sentences and glues translated sentences together to form a paragraph.

Constructor Summary

Constructors

Constructor and Description
Segmenter(SRX srx)

Method Summary

Modifier and Type	Method and Description
SRX	getSRX()
java.lang.String	glue(Language sourceLang, Language targetLang, java.util.List<java.lang.String> sentences, java.util.List<java.lang.StringBuilder> spaces, java.util.List<Rule> brules) Glues segments back into a paragraph.
java.util.List<java.lang.String>	segment(Language lang, java.lang.String paragraph, java.util.List<java.lang.StringBuilder> spaces, java.util.List<Rule> brules) Segments the paragraph to sentences according to currently setup rules.
void	segmentEntries(boolean needResegment, Language sourceLang, java.lang.String sourceEntry, Language targetLang, java.lang.String targetEntry, java.util.List<java.lang.String> sourceSegments, java.util.List<java.lang.String> targetSegments) Segment source and target entries from TMX when counts are equals.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Segmenter

public Segmenter(SRX srx)

Method Detail

getSRX

public SRX getSRX()

segment

public java.util.List<java.lang.String> segment(Language lang,
                                                java.lang.String paragraph,
                                                java.util.List<java.lang.StringBuilder> spaces,
                                                java.util.List<Rule> brules)

Segments the paragraph to sentences according to currently setup rules.

Bugfix for bug 83 : Sentences are returned without spaces in the beginning and at the end of a sentence.

An additional list with space information is returned to be able to glue translation together with the same spaces between them as in original paragraph.

Parameters:

paragraph - the paragraph text

spaces - list to store information about spaces between sentences (can be null)

brules - list to store rules that account to breaks (can be null)

Returns:

list of sentences (String objects)

glue

public java.lang.String glue(Language sourceLang,
                             Language targetLang,
                             java.util.List<java.lang.String> sentences,
                             java.util.List<java.lang.StringBuilder> spaces,
                             java.util.List<Rule> brules)

Glues segments back into a paragraph.

As segments are returned by segment(Language, String, List, List) without spaces before and after them, this method adds spaces if needed:

• For translation to non-space-delimited languages (Japanese, Chinese, Tibetan) it does not add any spaces.

  A special exceptions are the Break SRX rules that break on space, i.e. before and after patterns consist of spaces (they get trimmed to an empty string). For such rules all the spaces are added.

• For translation from non-space-delimited languages it adds one space.
• For all other language combinations it restores the spaces present before segmenting.

Parameters:

sentences - list of translated sentences

spaces - information about spaces in original paragraph

brules - rules that account to breaks

Returns:

glued translated paragraph

segmentEntries

public void segmentEntries(boolean needResegment,
                           Language sourceLang,
                           java.lang.String sourceEntry,
                           Language targetLang,
                           java.lang.String targetEntry,
                           java.util.List<java.lang.String> sourceSegments,
                           java.util.List<java.lang.String> targetSegments)

Segment source and target entries from TMX when counts are equals.