Pyglossary exporter

[bergentroll]

Replace tools/stardict.py export with PyGlossary export.

Rationale:

• Less code in the repo, no implicit Simsapa stardict.py dependency
• Multiple output formats available with few code lines e. g. Slob for Andoird Aard 2, dictd, EPUB. (But not MDict support yet, see the comability table write column.)

Notes:

• Slob export implemented, result is good, but execution time is about 7 times worse than stardict export. Because of this it is disabled by default, and may be ran with ./exporter/exporter.py --formats slob.

Other changes:

• StopWatch, Tic classes as alternatives to measure execution time. Used in exporter.py.

I hope to do some work on exporters for DPS and alternative formats in future.

:pray: :pray: :pray:

[gambhiro]

Dear Anton @bergentroll , did you verify that the pyglossary stardict exporter include the necessary formatting, e.g. inflection definitions, html pages, etc. in the output?

I can see it might be useful to export to other formats, and perhaps that is your motivation.

I don't mind if you don't need that little module from Simsapa, but I specifically added it because the pyglossary exporter only wrote the headword and definition fields for stardict format, ignoring the inflections.

Also, the dictzip dependency was a problem when trying to run the export on Windows.

[bergentroll]

Bhante Gambhīro :pray: :pray: :pray:

Dear Anton @bergentroll , did you verify that the pyglossary stardict exporter include the necessary formatting, e.g. inflection definitions, html pages, etc. in the output?

I try to not miss something. For now output seems fine: there is exactly the same amount of word entries and html-formatted definitions including inflections and special entries (like "thanks") as in stardict.py output. If bhante may, please check that :file_folder:build is appropriate indeed. :pray:

... I specifically added it because the pyglossary exporter only wrote the headword and definition fields for stardict format, ignoring the inflections

PyGlossary takes no special synonyms field, instead list of strings must be passed to bind headers as "synonyms". Initially I was stuck with it, but now it is OK.

I can see it might be useful to export to other formats, and perhaps that is your motivation.

Yes, and other motivation is to keep a bit less code to decrease complexity of the project.

Also, the dictzip dependency was a problem when trying to run the export on Windows.

There is an idea to implement compression with idzip for PyGlossary. And if it will not be in PyGlossary, it should be not hard to use it in the pyglossary_exporter.py just before zipping step.

[gambhiro]

The missing synonyms field stood out for me when I was testing pyglossary, but it seems you found a solution for that.

[bergentroll]

No more /usr/bin/dictzip dependency.