#936 Make several identical results in Search Results window appear as one result (similarly to the now-implemented function in the Fuzzy Matches window)

[Roman Mironov]

Problem:

When during search you find several identical results they appear all and need more place to display (e.g. several variants of segment "movie trailer"). Thease duplicates may be two or may be twenty.

Suggestion:

To make identical search results appear as one as in the Fuzzy Matches window where duplicates look as one segment with (+X more) addition.

[Didier Briel]

Why not simply uncheck the Display all results box?

That way, you have no duplicates.

Didier

[Didier Briel]

I close as "work for me" until my question has been answered.

Didier

[Roman Mironov]

Thanks, Didier. The Display all results applies to the segments in the project only, not to the search results from the /tm. If you have ten identical search results from the /tm, they all appear separately, whereas it could be beneficial to “aggregate” them into one similar to what happens in the Fuzzy Matches pane.

[Didier Briel]

The Display all results applies to the segments in the project only, not to the search results from the /tm

Good point. I reopen.

Didier

[cienislaw]

Display all results in search window affects also entries outside project - implemented in /trunk.

+X will need some more changes, but I can do this. Just tell me where to display merged count to make it usable.

[Didier Briel]

The way it is implemented now, it just removes any duplicate translation if Display all results is not checked.

This might be an issue, when you need to check whether a translation comes from an approved TMX, not just from memory.

Didier

[cienislaw]

I can add additional hashset for all entries not from project memory. Will be this enough or you suggest more 'classes' for TM sources?

[Didier Briel]

Grouping any entry not from project memory (similarly to what was done for the Fuzzy Match pane), as suggested by Roman, would be fine.

Didier

[cienislaw]

From what I see Fuzzy Match pane groups the same results from PM and project tmx and marks them as This project (+X more).

Anyway, I've added additional hashset which is used to group only entries outside PM (glossary, orphans, tm). +X functionality will be added later.

[Didier Briel]

Fuzzy Match entries and Search results have different purposes.

Didier

[cienislaw]

+X done. I've exchanged set with map, which counts occurrences during search. Then, before results are presented, is additional step which modifies preamble of each entry with more than one occurrence by adding "+%s more" to it. and this is place where 'problem' is. for PM entries there is no preamble which means additional line. for non-PM there is defined text or tmx file name which in last case can give wrong impression that 5 more entries are in for example "extra.tmx +5 more", when those 5 are from all other non-PM entries. please let me know if it can stay that way.

[Didier Briel]

In the current /trunk, I have a result from the first TMX with a given occurence when "Display all results" is not checked, and no other indication that there could be more.

please let me know if it can stay that way.

Do you mean, without any indication that there could be more entries?
(I don't understand your message above.)

Didier

[cienislaw]

'mytmxfilename.tmx +5 more'

when those 5 more may be from this tmx or any other non-PM source associated with project.

[Didier Briel]

'mytmxfilename.tmx +5 more'

when those 5 more may be from this tmx or any other non-PM source associated with project.

That would be fine, but I don't see it it /trunk.

Is it committed?

Didier

[cienislaw]

now commited.

[Didier Briel]

Implemented in SVN (/trunk).

When there is more than one result from external TMXs, + x more (x being the number of additional results) is displayed after the name of the first TMX where this result is encountered.

Didier

[cienislaw]

All non-PM entries were treated as one, which lead to unwanted results for orphans, alternative translations and so one.

I've added identification for each source:

    private final int ENTRY_ORIGIN_PROJECT_MEMORY = 0; // and above 0
    private final int ENTRY_ORIGIN_TRANSLATION_MEMORY = -1;
    private final int ENTRY_ORIGIN_ORPHAN = -2;
    private final int ENTRY_ORIGIN_ALTERNATIVE = -3;
    private final int ENTRY_ORIGIN_GLOSSARY = -4;
    private final int ENTRY_ORIGIN_TEXT = -5;

From now only PM and TM entries are grouped (separately) when 'show all results' is not checked. I can add non-grouped sources to any existing groups or create new ones just for them, just let me know if there is such need.

[Didier Briel]

Implemented in the released version 3.1.2 of OmegaT.

Didier