Public Repository

Last pushed: a year ago
Short Description
This is an open-source sentiment analysis tool for Hungarian language, written in Pyhon.
Full Description

Sentiment Analysis for Hungarian language

This is open-source sentiment analysis tool for Hungarian language, written in Python. The application has a REST API for easier access, and also accessible via Docker's container technology.

Docker & container installation:

Usage:
1. Launch application with:

  • Docker: docker exec -it sentanalysishun python $HOME/SentimentAnalysisHUN-master/src/Application.py

2. Usage of REST API:

  • Based on a HTTP POST request.
  • Receive an url (substitute to example): http://ip_addr_of_docker:5000/sentiment
  • Please use sentence tag for adding input like this {'sentence': '<write your input here>'} to enter your input.
  • Tool has two different HTTP POST requests:
    /sentiment: for overall score
    /sentiment_verbose: for more detailed scores
  • Example usage on Linux/Mac with console curl:
    curl -i -H "Content-Type: application/json" -X POST -d '{"sentence": "Budapest az egyik legszebb város."}' http://<ip_of_your_machine>:5000/sentiment
  • For Windows use a REST client like https://github.com/wiztools/rest-client

Sources
Following external open-source tools were applied, some of their installer files are collected at /resources folder. All of their rights are owned by their creators.

  1. NLP tools - HunToken, HunPoS, Ocamorph & HunMorph, Polyglot NER
  2. OpinHUBank sentiment corpus
  3. Python Flask REST API - http://flask.pocoo.org/
  4. Python packages - NumPy, SciPy, Sklearn, NLTK
Docker Pull Command
Owner
dhuszti

Comments (0)