[YouTube] Fix parsing short relative date formats (English only)

[Theta-Dev]

• I carefully read the contribution guidelines and agree to them.
• no API changes

I added support for the new, short date format (e.g. 1wk ago).

Fixes #1067

[TobiGr]

LGTM, but what about the other languages? I'd guess we need to update them, too. Should we ask the community for help here?

[Theta-Dev]

Parsing YouTube data in other languages is currently disabled and it would require more extensive changes to the dictionary and the parser.

I have implemented (and tested) a parser that works with all languages. Here is the dictionary for it. Also note that there are some special cases that have to be handled seperately (e.g. in French a is both an article and the short form of "years").

https://code.thetadev.de/ThetaDev/rustypipe/src/branch/main/testfiles/dict/dictionary.json

[AudricV]

Parsing YouTube data in other languages is currently disabled

Yes, but the timeago parser can be also used for something else than YouTube by clients, even if that's its main goal.

Is the data you provided extracted from YouTube? Do you wish to do other languages support using your similar approach to us? I think we should have a parser separating date units (seconds, hours, days, ...), digits (1, 2, 3, ...) and number units (tens, hundreds, thousands, ...) for each language we want to add support.

in French a is both an article and the short form of "years"

Did you mean the verb have at the third singular person and the short form of years? There is not a article, but a à one.

[Theta-Dev]

This is the French term in question (5 years ago):

il y a 5 a

I currently have a special case for the French language which checks if the string ends with a.

The data from the parsing dictionary is a combination of data extracted from YouTube and the CLDR repository.

[AudricV]

Merging this PR, as we would need to overhaul the timeago parser system to work with other languages on short time units. Thanks for the fix!