tardis: Text Analysis with Rules and Dictionaries for Inferring Sentiment (and more!)

TARDIS uses simple rules and dictionaries to analyze text. By default it uses built-in dictionaries to measure sentiment, i.e. how happy or sad text is. It handles negations, so it knows “not happy” means “sad”, and it handles modifiers, so it knows that “very happy” is more happy than “happy”. TARDIS also supports unicode emojis and multi-word tokens (so you can tell it that “supreme court” is neutral, instead of a combination of “supreme” (positive) and “court” (neutral). TARDIS also supports user-defined dictionaries and can be used to analyze other constructs beyond sentiment.

Features

• Handles ASCII and UTF-8 emojis :)