Tired of reading dependency structures as text in your jupyter notebook? This jupyter notebook extension visualizes NLP output using brat1, encapsuled by annodoc2.
Installation:
- clone this repository
- run
jupyter-nbextension install jupyter-annodoc, add--useror--symlinkif you like - run
jupyter-nbextension enable jupyter-annodoc
Open a jupyter notebook with a conll-x string as output, select that cell and click on the new jupyter-annodoc icon to convert the output.
This plugin tries to be smart about the output you provide. If the output is enclosed in quotes it assumes jupyter shows a variables's content and will unescape the output. Otherwise it will use the output as-is.
Currently, conll-x format is hardcoded. Patches welcome for automatic or manual selection of the right backend!
If you use it, drop me a short mail for motivation :-)