This repository contains utility code to process labeled text data from the Tagalog labeling software, and shows how this data can be used to train NLP models.
scripts/json2conll.pyconverts a Tagalog JSON file to a CONLL-type file, where every line contains a token and its label, separated by a tab. Takes the path to the Tagalog JSON file as its only argument, and produces a file with the same name and path, but with the extension.conll.