When a /tm/penalty-nnn has been set, the matches appear with the proper percentage, but that modified percentage does not appear in the project statistics.
What I think would be logical is that the penalty figure is taken into account in the match statistics, namely "Match statistics" and "Match statistics per file".
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Penalty is only one aspect of the problem: if I understand correctly, statistics should be based on the same score as the matches pane, correct?
In the actual implementation, before my correction, during statistics calculation a new score is calculated, similar to adjusted score but not taking care of penalties. I do not understand why.
So my idea was to re-use the adjusted score which is already in the NearString: this one does already take care of penalties.
But now the question I ask at the end. In the matches pane you can decide whenever you want to have matches sorted by adjusted, stemmed or non-stemmed score. The question is whenever we should always use adjusted score in statistics, or if we may use stemmed or non-stemmed score instead, in particular when the user selected these scores as first criteria for sorting matches?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
With this patch statistics will be based on adjusted score saved in NearString, rather than based on a re-calculated one.
But maybe you may find logical to use the score which has been set as first sort order instead?
PS. Patch works for OmegaT 5.4.1 as well as for 4.3.2
Last edit: Thomas CORDONNIER 2020-12-16
I'm not sure I understand what you mean.
What I think would be logical is that the penalty figure is taken into account in the match statistics, namely "Match statistics" and "Match statistics per file".
Penalty is only one aspect of the problem: if I understand correctly, statistics should be based on the same score as the matches pane, correct?
In the actual implementation, before my correction, during statistics calculation a new score is calculated, similar to adjusted score but not taking care of penalties. I do not understand why.
So my idea was to re-use the adjusted score which is already in the NearString: this one does already take care of penalties.
But now the question I ask at the end. In the matches pane you can decide whenever you want to have matches sorted by adjusted, stemmed or non-stemmed score. The question is whenever we should always use adjusted score in statistics, or if we may use stemmed or non-stemmed score instead, in particular when the user selected these scores as first criteria for sorting matches?