cascade-reconstruction-by-tree-samples/eval_helpers.py at master · xiaohan2012/cascade-reconstruction-by-tree-samples · 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
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
import numpy as np
import os
import pickle as pkl
from tqdm import tqdm

from sklearn.metrics import average_precision_score
from joblib import Parallel, delayed
from glob import glob

def eval_node_parallel_task(input_path, output_dir, score_func):
    basename = os.path.basename(input_path)
    output_path = os.path.join(output_dir, basename)
    try:
        obs, c = pkl.load(open(input_path, 'rb'))
        n = len(c)
        inf_probas = pkl.load(open(output_path, 'rb'))['inf_probas']
    except IOError:
        return np.nan

    hidden = list(set(np.arange(n)) - set(obs))
    y_true = np.array((c >= 0), dtype=np.double)

    return score_func(y_true[hidden], inf_probas[hidden])


def evaluate_score(input_dir, output_dir, score_func):

    scores = Parallel(n_jobs=-1)(

        delayed(eval_node_parallel_task)(
            input_path, output_dir, score_func)

        for input_path in tqdm(glob(input_dir + '*.pkl')))
    return scores


def eval_map(*args):
    return evaluate_score(*args, score_func=average_precision_score)


def evaluate_edge_prediction(g, true_edges, pred_edge_freq, eval_func):
    edge_true_vect = g.new_edge_property('float')
    edge_true_vect.set_value(0)
    for u, v in true_edges:
        edge_true_vect[g.edge(u, v)] = 1

    edge_pred_vect = g.new_edge_property('float')
    edge_pred_vect.set_value(0)
    for (u, v), f in pred_edge_freq.items():
        edge_pred_vect[g.edge(u, v)] = f
    return eval_func(edge_true_vect.a, edge_pred_vect.a)


def eval_edge_parallel_task(g, input_path, output_dir, score_func):
    """one task to send to joblib.Parallel
    """
    basename = os.path.basename(input_path)
    output_path = os.path.join(output_dir, basename)

    _, _, true_edges = pkl.load(open(input_path, 'rb'))
    edge_freq = pkl.load(open(output_path, 'rb'))['edge_freq']
    return evaluate_edge_prediction(g, true_edges, edge_freq, score_func)


def evaluate_edge_in_batch(g, input_dir, output_dir, score_func):
    scores = Parallel(n_jobs=-1)(
        delayed(eval_edge_parallel_task)(g, input_path, output_dir, score_func)
        for input_path in tqdm(glob(input_dir + '*.pkl'))
    )

    # for input_path in tqdm(glob(input_dir + '*.pkl')):
    #     basename = os.path.basename(input_path)
    #     output_path = os.path.join(output_dir, basename)

    #     _, _, true_edges = pkl.load(open(input_path, 'rb'))
    #     edge_freq = pkl.load(open(output_path, 'rb'))['edge_freq']
    #     scores.append(evaluate_edge_prediction(g, true_edges, edge_freq, score_func))

    return scores


def eval_edge_map(*args):
    return evaluate_edge_in_batch(*args, score_func=average_precision_score)