Menu
▾
▴
#800 Ignore 'Fragments that should be removed' on fuzzy matches
If you have set a 'Fragments that should be removed' regular expression in the tag-options dialog, and you have different segemtns that are equal, except for the part that should be removed, they might not show up in the fuzzy matches pane.
e.g, you want to remove some 'tip' that is prefixing some segments: e.g. '[tip: abbreviation of weekday]su' should be translated as if it was 'su'.
say you have
su
[tip: abbreviation of weekday']su
[tip: abbreviation of sunday]su
[tip: sunday]su
when you translate any of these sources to 'su', you want it as fuzzy match for the other entries.
Because the [tip xxx] part is messing up the editing distance of the match, the 'su' translation does not show up. The '[tip: xxx] should be ignored on matching.
Solved. If there is a remove-pattern defined in tag-options, then that part is ignored on the fuzzy match.
A fixed penalty of 5 is added if the removed part does not match some other segment, so the match does not show up as 100%
Closing...
This feature was implemented in the released version 2.6.1 of OmegaT.
This is mixing two features into one. For tag validation, the "Fragment(s) that should be removed" setting applies to the target text, but for fuzzy match ignoring it applies to the source text.
In Martin's examples above it was assumed that whatever the user wants to be blacklisted from the target text would be the same as what the user wants to be ignored from the source text. And that is not necessarily so.
If I want \[tip:.+?\] to be ignored when doing fuzzy matching, but I do want it to be present in translations (i.e. not cause tag validation to fail), then because of this mixing of two separate features I would not be able to do it.
If I specify \[tip:.+?\] in "Fragment(s) that should be removed" because I want it to be ignored during fuzzy matching, then all such segments will also fail tag validation (and will be flagged in red during translation).
I suggest that these two features be split into two:
* Fragment(s) prohibited in target text during tag validation
* Fragment(s) ignored in source text during fuzzy matching
The mixing of the two features can lead to some interesting gymnastics:
http://i48.tinypic.com/2qmepi1.jpg