Sentiment analysis with Python NLTK and IBM Tone Analyzer

sentiment.py

import json

from watson_developer_cloud import ToneAnalyzerV3

from nltk.sentiment.vader import SentimentIntensityAnalyzer


class SentimentService():
    def __init__(self):
        self.service = SentimentIntensityAnalyzer()

    def analyze(self, text):
        return self.format_polarity(self.service.polarity_scores(text))

    def format_polarity(self, polarity):
        return polarity


class ToneWatsonService():
    DEFAULT_VERSION = '2016-02-11'

    def __init__(self, access):
        '''
          access is a dict with username and password keys
        '''
        self.access = access

    def process(self, text):
        '''
        Response:
            {
            "document_tone": {
                "tone_categories": [
                {
                    "tones": [
                    {
                        "score": 0.000264,
                        "tone_id": "anger",
                        "tone_name": "Anger"
                    },
                    {
                        "score": 0.00016,
                        "tone_id": "disgust",
                        "tone_name": "Disgust"
                    },
                    {
                        "score": 0.000376,
                        "tone_id": "fear",
                        "tone_name": "Fear"
                    },
                    {
                        "score": 0.996063,
                        "tone_id": "joy",
                        "tone_name": "Joy"
                    },
                    {
                        "score": 0.001101,
                        "tone_id": "sadness",
                        "tone_name": "Sadness"
                    }
                    ],
                    "category_id": "emotion_tone",
                    "category_name": "Emotion Tone"
                },
                {
                    "tones": [
                    {
                        "score": 0.0,
                        "tone_id": "analytical",
                        "tone_name": "Analytical"
                    },
                    {
                        "score": 0.97759,
                        "tone_id": "confident",
                        "tone_name": "Confident"
                    },
                    {
                        "score": 0.0,
                        "tone_id": "tentative",
                        "tone_name": "Tentative"
                    }
                    ],
                    "category_id": "writing_tone",
                    "category_name": "Writing Tone"
                },
                {
                    "tones": [
                    {
                        "score": 0.096859,
                        "tone_id": "openness_big5",
                        "tone_name": "Openness"
                    },
                    {
                        "score": 0.264058,
                        "tone_id": "conscientiousness_big5",
                        "tone_name": "Conscientiousness"
                    },
                    {
                        "score": 0.472657,
                        "tone_id": "extraversion_big5",
                        "tone_name": "Extraversion"
                    },
                    {
                        "score": 0.61522,
                        "tone_id": "agreeableness_big5",
                        "tone_name": "Agreeableness"
                    },
                    {
                        "score": 0.104851,
                        "tone_id": "neuroticism_big5",
                        "tone_name": "Emotional Range"
                    }
                    ],
                    "category_id": "social_tone",
                    "category_name": "Social Tone"
                }
                ]
            }
            }
        '''
        tone_analyzer = ToneAnalyzerV3(
            username=self.access['username'],
            password=self.access['password'],
            version=self.DEFAULT_VERSION)

        return tone_analyzer.tone(text)