#!/usr/bin/env python

# -*- coding: utf-8 -*-

import os

import logging

from flask import Flask, request, jsonify

from flask_cors import CORS

import nltk.data

from train import load_classifier

logger = logging.getLogger('marlowe.redobject.server')

app = Flask(__name__)

CORS(app)

@app.route('/', methods=['POST'])

def index():

text = request.form.get('text')

if text is not None:

cl = load_classifier()

# split into sentences, evaluate each sentence, then average the scores

# the classifier and dataset is by sentence so best to only give

# it one sentence at a time

text = str(request.form.get('text').encode("utf8")).strip()

tokenizer = nltk.data.load('tokenizers/punkt/english.pickle')

sentences = tokenizer.tokenize(text)

scores = []

for s in sentences:

prob_dist = cl.prob_classify(s)

objectivity = float(round(prob_dist.prob("objective"), 2))

scores.append(objectivity)

average = sum(scores) / len(scores)

data = {'objectivity': round(average, 2)}

else:

data = {'objectivity': 1.0}

return jsonify(data)

if __name__ == '__main__':

debug = os.environ.get('DEBUG')

if debug is not None:

debug = True

else:

debug = False

app.run(debug=debug, host='0.0.0.0', port=int(os.environ.get('PORT', '5000')))