If possible, please check this quickly, and note the following:
to avoid using the replacments and to keep things very clean here, I placed some irregular forms in lexemes, please do that in your datasets as well, if you have less than, say 50 instances, to avoid overloading the code-based
use the ortho-profile to handle specific things, e.g., # cannot be handled in orthoprofiles, but there are only five contexts, so add each context with preceding sound, like a# -> a, and this is circumvented without exceeding usage of replacements.
you see that the code is drastically reduced now, at the cost of lexemes.tsv, but I prefer that over too many statements that are scattered over a python script.
If possible, please check this quickly, and note the following:
#
cannot be handled in orthoprofiles, but there are only five contexts, so add each context with preceding sound, likea# -> a
, and this is circumvented without exceeding usage of replacements.lexemes.tsv
, but I prefer that over too many statements that are scattered over a python script.