neulab / cmulab

CMU Linguistic Annotation Backend
14 stars 1 forks source link

Create command line client to annotate ELAN or Praat files? #3

Closed neubig closed 5 years ago

neubig commented 5 years ago

A nice first step at actually making the backend do something useful might be to create a client that annotates ELAN or Praat files. It looks like this library might be useful in doing so: https://github.com/dopefishh/pympi

If we could create a simple example that, for example, reads in an ELAN tier, adds VAD, and then writes out a new file with this added, that might be a nice proof of concept that is also easy to implement. I'd be happy to help out with creating the groundwork for it if that sounds useful.

antonisa commented 5 years ago

After reading in the ELAN file/tier, what does "sending to the server" mean? Does it mean

  1. (possibly creating a new corpus and) adding those segments, getting annotations with a model, and exporting them back or
  2. send, annotate them, and then delete them from the database
neubig commented 5 years ago

Good question. I think it's a non-trivial design decision. Maybe we could discuss this later today. Will you be around?

antonisa commented 5 years ago

Are you not @ICLR? (I was planning on working from home this morning and coming to the office around 3, but I can chat online whenever)

neubig commented 5 years ago

Yes, I'm at ICLR so it'd be online (sorry "around" was vague). I'll maybe call a little before noon Pittsburgh time, does that work?

antonisa commented 5 years ago

yes, that's fine!