As you suggested @rlskoeser, moving the regex to solr was absolutely the way to go here! Not only did it simplify the indexing code and reduce the index size, but it also fixed a bug where a search term without sigla would not be highlighted in a transcription where that term only appears with sigla.
As you suggested @rlskoeser, moving the regex to solr was absolutely the way to go here! Not only did it simplify the indexing code and reduce the index size, but it also fixed a bug where a search term without sigla would not be highlighted in a transcription where that term only appears with sigla.