deluge/preprocess.py at master · jbragg/deluge · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
"""preprocess.py"""

from __future__ import division
import csv
import collections
import itertools
import numpy as np


def load_data(fname):
    """Reads a csv file and loads votes and labels into convenient format.

    Input is csv file with the following fields:
    - item (str): item name
    - label (str): label name
    - selected (int): vote

    Returns:
    - items ([str]): item names
    - labels ([str]): label names
    - votes (dict): (item name, label name) -> votes (list: bool)

    """
    with open(fname, 'r') as f:
        reader = csv.DictReader(f)
        items = set()
        labels = set()
        votes = collections.defaultdict(list)
        for row in reader:
            items.add(row['item'])
            labels.add(row['label'])
            votes[row['item'], row['label']].append(bool(int(row['selected'])))

        return list(items), list(labels), votes


class Data:

    def __init__(self, fin):
        items, labels, votes = load_data(fin)
        self.items = items
        self.labels = labels
        self.votes = votes

    def make_posneg(self):
        """Convert votes to matrix format.

        Returns: item name -> matrix (2 x |labels|) with the number of
        positive votes (first row) and negative votes (second row)

        """
        posneg = dict()
        for i in self.items:
            pos = np.array([sum(self.votes[i, l]) for
                            l in self.labels])
            neg = np.array([len(self.votes[i, l]) for
                            l in self.labels]) - pos
            posneg[i] = np.vstack((pos, neg))

        return posneg

    def __repr__(self):
        return '%s, %s\nitem1: %s' % (self.items, self.labels, self.votes.items()[0])