Add Basque support to the Segmenter

[zuazo]

This adds Basque support to the segmenter.

• abbr.tsv:
  • Added common abbreviations from different sources:
    • Elhuyar: http://www.euskara.euskadi.net/appcont/elhuyar/laburdura_zerrenda.pdf
    • Euskaltzaindia: https://www.euskaltzaindia.eus/index.php?option=com_ebe&view=bilaketa&Itemid=1161&task=bilaketa&id=985
    • Euskaljakintza: https://euskaljakintza.com/kontsultategia/laburdurak/
  • Added two regex rules to avoid splitting sentences with . (-garren) in ordinal numbers.
• validate.tsv: normalize diacritics from Spanish and French that are frequent in Basque texts due to common names or foreign words.

With this segmenter, I re-trained your previous models adding the Wikipedia corpus here: coqui-ai/STT-models#25

[ftyers]

Thanks! :)