process for submitting corrections

[jtauber]

I am compiling a list of lemmatisation errors, mostly through applying my own rule-based morphology parsing.

What should the process be for submitting corrections?

[gcelano]

Good question. I think you could generate a pull request here, for example

[jtauber]

when you generate new lemmatisations of text, do you use the uniqueTokens?

The corrections I'm compiling are not context- / text- specific but are just corrections to the unique tokens / morpheus output. For example: when verbs are lemmatised with the wrong preverb.

If I make pull requests against the uniqueTokens, will those improvements be picked up next time you run lemmatisation on actual texts? Or will they get overwritten unless I also make a pull request against the Morpheus data?