rec_metric.py 2.38 KB
Newer Older
WenmuZhou's avatar
WenmuZhou committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import Levenshtein
tink2123's avatar
tink2123 committed
16
import string
WenmuZhou's avatar
WenmuZhou committed
17
18
19


class RecMetric(object):
tink2123's avatar
tink2123 committed
20
    def __init__(self, main_indicator='acc', is_filter=False, **kwargs):
WenmuZhou's avatar
WenmuZhou committed
21
        self.main_indicator = main_indicator
tink2123's avatar
tink2123 committed
22
        self.is_filter = is_filter
WenmuZhou's avatar
add eps  
WenmuZhou committed
23
        self.eps = 1e-5
WenmuZhou's avatar
WenmuZhou committed
24
25
        self.reset()

tink2123's avatar
tink2123 committed
26
27
28
29
30
    def _normalize_text(self, text):
        text = ''.join(
            filter(lambda x: x in (string.digits + string.ascii_letters), text))
        return text.lower()

WenmuZhou's avatar
WenmuZhou committed
31
32
33
34
35
36
    def __call__(self, pred_label, *args, **kwargs):
        preds, labels = pred_label
        correct_num = 0
        all_num = 0
        norm_edit_dis = 0.0
        for (pred, pred_conf), (target, _) in zip(preds, labels):
littletomatodonkey's avatar
littletomatodonkey committed
37
38
            pred = pred.replace(" ", "")
            target = target.replace(" ", "")
tink2123's avatar
tink2123 committed
39
40
41
            if self.is_filter:
                pred = self._normalize_text(pred)
                target = self._normalize_text(target)
WenmuZhou's avatar
WenmuZhou committed
42
            norm_edit_dis += Levenshtein.distance(pred, target) / max(
WenmuZhou's avatar
WenmuZhou committed
43
                len(pred), len(target), 1)
WenmuZhou's avatar
WenmuZhou committed
44
45
46
47
48
49
50
            if pred == target:
                correct_num += 1
            all_num += 1
        self.correct_num += correct_num
        self.all_num += all_num
        self.norm_edit_dis += norm_edit_dis
        return {
WenmuZhou's avatar
add eps  
WenmuZhou committed
51
52
            'acc': correct_num / (all_num + self.eps),
            'norm_edit_dis': 1 - norm_edit_dis / (all_num + self.eps)
WenmuZhou's avatar
WenmuZhou committed
53
54
55
56
        }

    def get_metric(self):
        """
zhoujun's avatar
zhoujun committed
57
        return metrics {
WenmuZhou's avatar
WenmuZhou committed
58
59
60
61
                 'acc': 0,
                 'norm_edit_dis': 0,
            }
        """
WenmuZhou's avatar
add eps  
WenmuZhou committed
62
63
        acc = 1.0 * self.correct_num / (self.all_num + self.eps)
        norm_edit_dis = 1 - self.norm_edit_dis / (self.all_num + self.eps)
WenmuZhou's avatar
WenmuZhou committed
64
65
66
67
68
69
70
        self.reset()
        return {'acc': acc, 'norm_edit_dis': norm_edit_dis}

    def reset(self):
        self.correct_num = 0
        self.all_num = 0
        self.norm_edit_dis = 0