Order of pronunciation variants should be like on Wiktionary

[dietmar]

Hi and thanks a lot for this repository!

I'd like to suggest to keep the ordering of pronunciation variants for the same word/entry in the same order as they appear on Wiktionary, because that typically reflects how common that pronunciation is (most common first).

The example where I came across this topic is the word "Storys" https://de.wiktionary.org/wiki/Storys
On that Wiktionary page we have this:
IPA: [ˈstɔːʁɪz], [ˈstɔʁis], [ˈstoːʁis], [ˈstɔʁiːsː], [ˈstoːʁiːsː], auch: [ˈʃtɔːʁɪz], [ˈʃtɔʁis], [ˈʃtoːʁis], [ˈʃtɔʁiːsː], [ˈʃtoːʁiːsː]

This repository's de_dewikt.csv has:

$ egrep '^Storys' de_dewikt.csv
Storys,/ˈstoːʁis/
Storys,/ˈstoːʁiːsː/
Storys,/ˈstɔʁis/
Storys,/ˈstɔʁiːsː/
Storys,/ˈstɔːʁɪz/
Storys,/ˈʃtoːʁis/,auch
Storys,/ˈʃtoːʁiːsː/,auch
Storys,/ˈʃtɔʁis/,auch
Storys,/ˈʃtɔʁiːsː/,auch
Storys,/ˈʃtɔːʁɪz/,auch

To my ear (Austrian German), the variants with oː feel very strange. But they come first, because of the final sorting that happens in extract_de_ipa.sh:

LC_ALL=C sort -u -o de_dewikt.csv < de_dewikt_raw.csv

I think it would be a good idea to keep the original order. In my use case of this list, I need one pronunciation for a given word. Just always taking the first one in the list would work better after this change (knowing that of course the information on Wiktionary can still be questionable).