Represents an aligner that aligns analogously to given reference alignment, which means numbers of source and target segments in resulting list will be the same as in reference alignment. This implies that numbers of source and target segments must be identical in reference and input alignments. Additionally scores are copied from reference alignment to output alignments.
Useful when to perform an alignment it is required to modify or destroy segment contents (tokenization, stemming, removal of rare words, etc. - see {@link MooreMacro}). After that the original input can be unified with damaged alignment list to obtain the undamaged result but aligned correctly.
@author loomchild
|
|
|
|