vader-sentiment-cpp

Quick Start

vader-sentiment-cpp is a fast, C++ port of the Python NLTK Vader Sentiment Analyzer, optimized for speed. This version retains key functionality while being lighter and quicker, ideal for performance-sensitive applications.

Approach

This implementation mimics the core sentiment analysis features of the original model using a sentiment lexicon to classify text as positive, negative, or neutral.

🔧 Example Usage

1. Loading the Lexicon

Load the lexicon.txt file into an unordered map to store words and their sentiment scores.

std::ifstream lexicon("resources/lexicon.txt");
std::unordered_map<std::string, float> lexicals;
if (lexicon) {
    std::string line;
    while (std::getline(lexicon, line)) {
        uint16_t pos = 0, last = 0;
        pos = line.find("\t", last);
        std::string word = line.substr(last, pos - last);
        last = pos + 1;
        pos = line.find("\t", last);
        float val = std::stof(line.substr(last, pos - last));
        lexicals[word] = val;
    }
    lexicon.close();
} else throw std::runtime_error("[Vader](ERROR) Cannot open lexicon.txt");

2. Storing Review Data

Define a Review struct to store review details: ID, rating, and text.

struct Review {
    uint16_t m_rating;
    uint64_t m_id;
    std::string m_corpus;

    Review(uint16_t rating, uint64_t id, std::string corpus) :
        m_id(id), m_rating(rating), m_corpus(corpus) {};
};

3. Storing Review Data

Load data from the reviews.csv file into a vector of Review objects.

std::vector<Review> revs;
std::ifstream file("resources/reviews.csv");
if (file) {
    std::string line;
    std::getline(file, line); // Skip header
    while (std::getline(file, line)) {
        uint16_t pos = 0, last = 0;
        pos = line.find(',', last);
        uint32_t id = std::stoi(line.substr(last, pos - last));
        last = pos + 1;
        pos = line.find(',', last);
        uint16_t rating = std::stoi(line.substr(last, pos - last));
        last = pos + 1;
        revs.emplace_back(rating, id, line.substr(last, pos - last));
    }
} else {
    throw std::runtime_error("[Vader](ERROR) Cannot open reviews.csv");
}

4. Evaluating Sentiment

Evaluate the sentiment of the first review using vader::eval.

vader::eval(revs[0].m_corpus, lexicals);

Citation

Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text. Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.