cnsenti

Why we included this project

Chinese text sentiment analysis often means either training a model or calling an API, but cnsenti takes a simpler route: it counts sentiment words from built-in lexicons. The Sentiment class uses the Hownet dictionary to tally positive and negative terms, and its sentiment_calculate method also handles degree adverbs and negation words that change the meaning. The Emotion class draws on the Dalian University of Technology emotion ontology to count seven emotion categories like joy, sadness, anger, and fear. You can supply your own positive and negative word lists as plain text files, which helps when the default dictionaries miss domain-specific terms. One thing to know: the project is no longer maintained; its functionality has moved into the cntext library, so treat cnsenti as a compact reference or a starting point rather than something to build on long-term.