search

clulab / processors

Natural Language Processors
https://clulab.github.io/processors/
417 stars 101 forks source link

Java interface for processors #755

Open MihaiSurdeanu opened 10 months ago

MihaiSurdeanu commented 10 months ago

I think the adoption of this library will be higher through Java.

We do have an example here: https://github.com/clulab/processors/blob/master/corenlp/src/main/java/org/clulab/processors/ProcessorsJavaExample.java

We should check/improve two things:

  1. Does this code run with the new ONNX models?
  2. Is there a simpler (what is simplest) Java interface for these processors? I think we can focus on simply supporting the mkDocument and annotate methods in Java.
MihaiSurdeanu commented 10 months ago

This is not urgent.

BeckySharp commented 10 months ago

can you make sure to support custom annotation (e.g., when we tag, lemmatize, etc but don't do the more expensive bits?), and also choosing which processor (in case some are still GPL)

kwalcock commented 8 months ago

mkDocument and annotate still work.

OdinStarter runs on Java 8, 11, 17, 18, and 21, including grounding that reads a kryo file.

kwalcock commented 8 months ago

One can still use the FastNLPProcessor or CoreNLPProcessor.

MihaiSurdeanu commented 8 months ago

Nice!!

The new deberta model is still training. ETA tomorrow.