I managed to write an .ecv with different annotation conventions as the “languages” of the multilingual setting. However, I encounter two problems when applying this .ecv to .eaf files that have already been annotated using the values of one “language” but from an internal CV:
-
Some files do not seem to automatically switch between the language options, even though all values are in the .ecv and all the .eaf files have been updated with “CVE_REF” set to the correct CVE_ID in the .ecv (all entries have unique ids). I can’t seem to figure out why this is the case, as some files happily toggle between the languages, others are fixed (and the default language setting for the tiers and tier types are identical across files).
-
Some values in one language A correspond to the same value in language B, but these have still been assigned separate CVE_ID numbers in the .ecv, for example:
LanguageA,LanguageB,CVE_ID
XXX,n/a,1
ZZZ,n/a,2
…
However, if the .eaf file allows toggling between the languages (see point 1 above), such that XXX–>n/a and ZZZ–>n/a when changing from language A to B, it seems such values do not return to their unique values when going from language B to A. It is as if the conversion between the languages is done partly on the annotation values rather than the CVE_REF link to the .ecv CVE_ID. Is this the case?