contribute

[ludwig-loth]

Hi,
I am considering continuing this project.
Does this make sense at all from your consideration or should I create a completely new fork of compromise?

This project seems very outdated and cannot be used anymore due to very outdated dependencies.

The slack-channel is not available anymore. That's why I'm contacting this way

[spencermountain]

Hey @dollinvans thats great to hear, i could really use some help.
Yes, the current version is not recommended.
Compromise (en) v14 is coming out in a few weeks and allows new languages without a hard-fork. This is a huge improvement, and I'm keen to restart work on de-compromise, especially if i have some help from a fluent speaker.

I suppose that's a non-answer. I could setup some boilerplate using a v14 prerelease, if you wanted. Otherwise this summer is better timing for a reboot.
Cheers!

[ludwig-loth]

I'm glad to hear that.
The v14 sounds very promising. If it's not that much of work for you I would be happy if you could ´give me some kind of a prerelease. But it would be only be for a little training for me for the real releaseversion, I guess.

So is there some channel where I can participate? The Slack channel seems to be abandoned

[spencermountain]

hey @dollinvans i've started a new branch called v14 for us to play on.
check out the branch, run npm install then you should see the results of 'node scratch.js'

happy to help sort things out - right now there is very basic tagging via a small lexicon lookup, and a quick title-case check. Feel free to go crazy, change things however you'd like.
cheers

[spencermountain]

so the biggest things to tackle are adding prefix or suffix rules for the tagger -
I've been using the tiger dataset for some machine-learning of german, and it's amazing. I setup a quick script in ./scripts/tiger if you wanted to have a go at it.

another big task is noun inflection and verb-conjugation. This helps us make the lexicon much bigger. I can tackle this one, if you'd like to try the first one.

I guess number-parsing is a big one too.

up to you!
there's a bunch of people who are willing to help, if this gets off the ground.
cheers

[spencermountain]

hey i got some verb conjugation working.
haha! just a rough past/present right now.

if you add 'sprechen' to ./lexicon/infinitives.js and then run npm run pack
you will see it now also adds the past and present tenses of it

let text = 'er spricht'
var dok = nlp(text)
dok.debug()
/*
  │ 'er'       - Pronoun
  │ 'spricht'  - PresentTense
*/

🥳