Skip to main content

Entity extractioN, Translation and Evaluation using BabelFy

Project description

This is an entity extractor, translator and evaluator that uses BabelFy . Initially developed for the TraMOOC project. It is written in Python 3.

https://github.com/proycon/babelente/blob/master/logo.jpg?raw=true

Installation

(not yet ready; to appear soonish)

pip3 install babelente

or clone this github repository and run python3 setup.py install, optionally prepend the commands with sudo for global installation.

Usage

You will need a BabelFy API key, get it from BabelNet.org .

See babelente -h for usage for now.

To evaluate a translation (english to portuguese in this example), output wil be JSON to stdout:

$ babelente -k "YOUR-API-KEY" -f en -t pt -S sentences.en.txt -T sentences.pt.txt > output.json

To re-evaluate:

$ babelente --evalfile output.json -S sentences.en.txt -T sentences.pt.txt > newoutput.json

You can also use BabelEnte to just extract entities a single language, without evaluation:

$ babelente -k "YOUR-API-KEY" -f en -S sentences.en.txt > output.json

Evaluation

The evaluation produces two metrics, computer per sentence/line pair and final score aggregated as macro-average:

  • Precision -How many of the target synsets are correct? (|matchingsynsets| / |targetsynsets|)

  • Recall - How many of the source synsets are found? (|matchingsynsets| / |sourcesynsets|)

License

GNU - GPL 3.0

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

BabelEnte-0.3.0.tar.gz (8.9 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page