A simple class to score text with LabMT and pandas based on the code at http://neuro.imm.dtu.dk/wiki/LabMT

labmt_scorer.py

import pandas as pd
import re


class LabMTScorer():

    """ A simple class to score text with LabMT and pandas,
    based on the example provided by  Finn Årup Nielsen
    at http://neuro.imm.dtu.dk/wiki/LabMT  """

    def __init__(self, source=None):
        if source is None:
            source = ('http://www.plosone.org/article/'
                      'fetchSingleRepresentation.action?'
                      'uri=info:doi/10.1371/journal.pone.0026752.s001')
        labmt_csv = pd.read_csv(source, skiprows=2, sep='\t', index_col=0)
        self.avg_happiness = labmt_csv.happiness_average.mean()
        self.happiness = (
            labmt_csv.happiness_average - self.avg_happiness).to_dict()

    def score_tokens(self, tokens):
        score = sum([self.happiness.get(tok.lower(), 0.0)
                     for tok in tokens]) / len(tokens)
        return score

    def tokenize(self, text):
        tokens = re.split(r"[^0-9A-Za-z\-']+", text)
        return tokens

    def score_text(self, text):
        tokens = self.tokenize(text)
        score = self.score_tokens(tokens)

        return score


def test():
    test_texts = ["He was a bad, bad man. Utterly corrupt.",
                  "She's the best. So awesome and great."]
    scorer = LabMTScorer()
    for test_text in test_texts:
        result = scorer.score_text(test_text)
        print(test_text, result)

if __name__ == "__main__":
    test()