Hi, following the doc the only option to export annotations is using db-out, but it exports to JSONL format. Exist a way to export annotations into UIMA CAS xmi format directly?
We don’t have a direct converter, no, but if you end up writing one, feel free to share it!
One reason we chose JSON / JSONL as the default output format is that it’s generally pretty easy to parse in most languages and frameworks, and as a result, allows users to write their own converters. For example, NER annotations include the "text"
and a list of entity spans as the "spans"
property, annotated in the character offset format with a "start"
and "end"
index and a "label"
. Converting this to most other formats should hopefully be pretty straightforward.