text annotation with BRAT

Author

Oren Bochman

Published

Tuesday, January 16, 2018

BRAT is a rapid annotation tool. Its old no long maintained - the original developers seem to have moved on. Requests for RTL support have been largely ignored. Some advantages:

  1. Small code base (Python backend via cgi !? and JQuery front end)
  2. Fairly clean
  3. based on stav projecter
  4. nice data model
  5. live embedding http://brat.nlplab.org/embed.html
  6. many annotation types.

looking into modernising the UI to work with RTL.

Get brat in a docker container.

plan A: Recreate the svg element using d3.js

plan B

Breaking up the element into smaller web-components in polymer. Organize things better using using modern javascript

Brat and RTL support

references

  • https://github.com/nlplab/brat
  • https://github.com/UniversalDependencies/docs/issues/52
  • https://github.com/nlplab/brat/issues/774
  • https://github.com/spyysalo/conllu.js/issues/11
  • https://github.com/nlplab/brat/pull/1150

Citation

BibTeX citation:
@online{bochman2018,
  author = {Bochman, Oren},
  title = {Text Annotation with {BRAT}},
  date = {2018-01-16},
  url = {https://orenbochman.github.io/posts/2018/2018-01-16-BRAT/2018-01-16-BRAT.html},
  langid = {en}
}
For attribution, please cite this work as:
Bochman, Oren. 2018. “Text Annotation with BRAT.” January 16, 2018. https://orenbochman.github.io/posts/2018/2018-01-16-BRAT/2018-01-16-BRAT.html.