This PR is to add a process.py file for converting a directory of gold JSON annotations into a single CSV file. This satisfies the first two objectives of #89.
additions
role-filler-binding/process.py
gold files generated from process.py on 231117-aapb-annotations-44 data
discussion before merging
is this a suitable format for the data (newline-separated CSV)?
do we want to discard any of the non-RF data from this batch? do we actually want all these duplicates?
For the second discussion point, I think we should keep the duplicate labels, since they are also human generated (valuable) data that have potential usages.
This PR is to add a
process.py
file for converting a directory of gold JSON annotations into a single CSV file. This satisfies the first two objectives of #89.additions
role-filler-binding/process.py
process.py
on231117-aapb-annotations-44
datadiscussion before merging