algorithm.py

# Algorithm for discerning high level information about folks in the civic technology scene
# 1. Categorize sentences into one of the categories under consideration:
#     * Funding
#     * Data
#     * Employment
#     * Collaboration
#     * Location - from investigator
# 2. Parse individual sentences once high level labels have been defined to do deep semantic analysis automatically
#    * We do this by pulling out the named entities and using the correct action words to imply causality
# 3. Save to database

import spacy
import nltk
from spacy.parts_of_speech import *
from sklearn.feature_extraction.stop_words import ENGLISH_STOP_WORDS
from nltk.corpus import stopwords
from sklearn.feature_extraction.text import CountVectorizer
import string
from sklearn.svm import LinearSVC
from sklearn.feature_extraction.text import TfidfTransformer #consider moving from CountVectorizer to this
from nltk.classify.scikitlearn import SklearnClassifier
import itertools
# from text_classify.algorithms import *
import pickle

def preprocess(sentence,label=None):
    tokens = nltk.word_tokenize(sentence)
    tokens = [w for w in tokens if w not in stopwords.words("english")]
    features = {}
    for token in tokens:
        features[token]=tokens.count(token)
    if label:
      return (features,label)
    else:
        return features

def svm(train_data,preprocessing=True):
    training_data = []
    for data in train_data: 
        training_data.append(preprocess(data[0],label=data[1]))
    cl = SklearnClassifier(LinearSVC())
    cl.train(training_data)
    return cl

def tokenizeText(sample):
  tokens = sample.split(" ")
  # stoplist the tokens
  tokens = [tok for tok in tokens if tok not in STOPLIST]

  # stoplist symbols
  tokens = [tok for tok in tokens if tok not in SYMBOLS]

  # remove large strings of whitespace
  while "" in tokens:
      tokens.remove("")
  while " " in tokens:
      tokens.remove(" ")
  while "\n" in tokens:
      tokens.remove("\n")
  while "\n\n" in tokens:
      tokens.remove("\n\n")
  return tokens


def accuracy(classifier_name,classifier,test_data):
    """
    example usage: 
    from text_classify.algorithms import *
    import pickle
    training_data = pickle.load(open("training_data","rb"))
    test_data = pickle.load(open("test_data,"rb"))
    cl = svm(training_data)
    accuracy("svm",cl,test_data)
    
    the possible choices for classifiers at present are:
       svm, decision_tree, naive_bayes
    """
    if classifier_name in ["svm","decision_tree"]:
        testing = [preprocess(data[0],label=data[1]) for data in test_data]
        counter = 0.0
        # print test_data
        for ind,data in enumerate(test_data):
            if classifier.classify(testing[ind][0]) == data[1]:
                counter += 1
        return counter/len(test_data)
        # print counter
        # print len(test_data)
        # print counter/len(test_data)
    if classifier_name == "naive_bayes":
        return classifier.accuracy(test_data)
    else:
        return "no idea!"


def readfile_tokenizedata(file):
  # read in data file (test dataset???)
  with open(file,"r") as f:
    data = f.read()

  #sentence tokenization using spaCy
  nlp = spacy.load("en")
  d = data.decode('utf-8')
  doc = nlp(d, tag=True) #tokenized doc
  sentences = [sent.string.strip() for sent in doc.sents]
  return sentences


STOPLIST = set(stopwords.words('english') + ["n't", "'s", "'m", "ca"] + list(ENGLISH_STOP_WORDS))
SYMBOLS = " ".join(string.punctuation).split(" ") + ["-----", "---", "..."]

# TODO: Fill in the labels for each training set

# funding (directional)
labels = ['Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Funding', 'Funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding', 'Not funding']
# data (directional)
labels2 = ['Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data', 'Data', 'Not data', 'Not data', 'Not data', 'Not data', 'Not data']
# employment (bidirectional)
labels3 = ['Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Not employment', 'Employment', 'Not employment', 'Employment', 'Not employment', 'Not employment', 'Employment', 'Employment', 'Employment', 'Not employment', 'Not employment']
# collaboration (bidirectional)
labels4 = ['Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Not collaboration', 'Collaboration', 'Collaboration']
# location (bidirectional)
labels5 = ['Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location', 'Not location']


# read in and tokenize each training set

# Option 1: word tokenization (removes stopwords, symbols, whitespace chunks)
# with open("funding_training.txt", 'r') as f:
#   data1 = f.read()
#   for d in tokenizeText(data1):
#     print d

# Option 2: sentence tokenization using spaCy
data1 = readfile_tokenizedata("funding_training.txt")
data2 = readfile_tokenizedata("data_training.txt")
data3 = readfile_tokenizedata("employment_training.txt")
data4 = readfile_tokenizedata("collaboration_training.txt")
data5 = readfile_tokenizedata("location_training.txt")

for d in data1:
  preprocess(data)


#Generate training set for each set of labels
training1 = zip(data1,labels)
training2 = zip(data2, labels2)
training3 = zip(data3, labels3)
training4 = zip(data4, labels4)
training5 = zip(data5, labels5)


# Create classifier for each training set
cl1 = svm(training1)
cl2 = svm(training2)
cl3 = svm(training3)
cl4 = svm(training4)
cl5 = svm(training5)

#test set - this will eventually be all of the Civicist/TP content?
test_data = preprocess("sample.txt")

accuracy("svm",cl1, test_data)
accuracy("svm",cl2, test_data)
accuracy("svm",cl3, test_data)
accuracy("svm",cl4, test_data)
accuracy("svm",cl5, test_data)