August 21, 2019

174 words 1 min read

Varal7/ieturk

Varal7/ieturk

Intuitive Annotation Tool for Information Extraction / Named Entity Recognition using localturk / Amazon Mechanical Turk

repo name Varal7/ieturk
repo link https://github.com/Varal7/ieturk
homepage
language JavaScript
size (curr.) 1039 kB
stars (curr.) 246
created 2018-12-17
license MIT License

IE-Turk: Annotation tool for Information Extraction with localturk / Amazon Mechanical Turk

Annotate

Annotate using Mechanical Turk

  • Create an new project in Amazon Mturk interface.
  • Paste the content of annotate.html in the second tab Design layout.
  • Insert the scripts config.js, annotate.js and the CSS file style.css into the document as well
  • Prepare a tokenized version of the entry, splitting characters with >>.
  • Submit batch

Annotate locally

Requires localturk. Install using

npm install -g localturk

Modify config.js with the name of the fields of interest. Then simply run it using the same tokenized csv file as for Mechanical Turk.

localturk annotation.html input.csv output.csv

Screenshot

Visualize

Simply open visualize.html with any modern browser. Then choose the .csv that came from either localturk or Mechanical Turk.

Screenshot

This file is also served at Visualization Demo. You can try uploading this CSV file.

comments powered by Disqus