-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathutils.py
231 lines (202 loc) · 10.3 KB
/
utils.py
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
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
# the three columns we use: input, label, and id
Column_to_check = {
'winogrande': {'input': 'sentence', 'label': lambda x: x[f'option{x["answer"]}'], 'id': 'id'},
'ceval': {'input': 'question', 'label': lambda x: x[x['answer']], 'id': 'id'},
'ceval_train': {'input': 'question', 'label': lambda x: x[x['answer']], 'id': 'id'},
'mmlu': {'input': 'question', 'label': lambda x: x[x['answer']], 'id': 'id'},
'mmlu_train': {'input': 'question', 'label': lambda x: x['choices'][x['answer']], 'id': 'id'},
'hellaswag': {'input': 'ctx', 'label': lambda x: x['endings'][int(x['label'])], 'id': 'ind'},
'ARC': {'input': 'question', 'label': lambda x: x['choices']['text'][x['choices']['label'].index(x['answerKey'])], 'id': 'id'},
'commonsense_qa': {'input': 'question', 'label': lambda x: x['choices']['text'][x['choices']['label'].index(x['answerKey'])], 'id': 'id'},
'squad_v2': {'passage': 'context', 'question': 'question', 'label': lambda x:x['answers']['text'], 'id': 'id'},
'squad_v2_train': {'passage': 'context', 'question': 'question', 'label': lambda x:x['answers']['text'], 'id': 'id'},
'quac': {'passage': 'background', 'id': 'dialogue_id'},
'boolq': {'passage': 'passage', 'id': 'question'},
'wiki_clean': {'passage': 'text', 'id': 'title'},
'bbc_clean': {'passage': 'content', 'id': 'link'},
'wiki_all': {'passage': 'text', 'id': 'title'},
'bbc_all': {'passage': 'content', 'id': 'link'},
}
# The name of benchmarks on the Huggingface Hub, and the split to be used
Hf_Name_and_Split = {
'winogrande': {'hf_name': 'liyucheng/winogrande_val', 'split': 'validation'},
'ceval': {'hf_name': 'liyucheng/ceval_all', 'split': 'val'},
'ceval_train': {'hf_name': 'liyucheng/ceval_all_dev', 'split': 'dev'},
'mmlu': {'hf_name': 'liyucheng/mmlu_test', 'split': 'train'},
'mmlu_train': {'hf_name': 'liyucheng/mmlu_train', 'split': 'train'},
'hellaswag': {'hf_name': 'Rowan/hellaswag', 'split': 'validation'},
'ARC': {'hf_name': 'liyucheng/arc_test', 'split': 'test'},
'commonsense_qa': {'hf_name': 'commonsense_qa', 'split': 'validation'},
'quac': {'hf_name': 'quac', 'split': 'validation'},
'boolq': {'hf_name': 'boolq', 'split': 'validation'},
'squad_v2': {'hf_name': 'squad_v2', 'split': 'validation'},
'squad_v2_train': {'hf_name': 'squad_v2', 'split': 'train'},
'wiki_clean': {'hf_name': 'RealTimeData/wikitext_latest', 'split': 'train'},
'bbc_clean': {'hf_name': 'RealTimeData/bbc_latest', 'split': 'train'},
'wiki_all': {'hf_name': 'RealTimeData/wikitext_alltime', 'split': 'train'},
'bbc_all': {'hf_name': 'RealTimeData/bbc_alltime', 'split': 'train'},
}
# This is used to choose the right Bing market, based on the language of the dataset
Dataset_lang = {
'winogrande': 'en-US',
'ceval': 'zh-CN',
'mmlu': 'en-US',
'hellaswag': 'en-US',
'ARC': 'en-US',
'commonsense_qa': 'en-US',
}
Recall_threshold_for_dataset = {
'winogrande': 0.7,
'ceval': 0.8,
'mmlu': 0.7,
'hellaswag': 0.7,
'ARC': 0.7,
'commonsense_qa': 0.7,
}
from typing import Any
import numpy as np
from datasets import load_dataset
import re
import json
np.random.seed(42)
from nltk.tokenize import word_tokenize
en_processor = lambda x: [token.lower() for token in word_tokenize(x) if token not in ['.', ',', '?', '!', ';', ':', '"', "'", '']]
zh_processor = lambda x: [token for token in ' '.join(x).split(' ') if token not in [ '，', '。', '？', '！', '；', '：', '“', '”', '‘', '’', '（', '）', '《', '》', '、',]]
def random_sample_ds(ds, n = 100):
# ds is a dataset object
# n is the number of samples to be randomly selected
if len(ds) <= n:
return ds
return ds.select(np.random.choice(len(ds), n))
def prepare_query(dataset_name, row):
"""
Here we verbalize the test instances to form a textual query so that we can send them to Bing Search or calculating ppl.
For multi-choice benchmarks, we use 'input' and 'label' to form the query.
For reading comprehension benchmarks, we use 'passage', 'question', and 'answer' to form the query.
For test samples which have blanks in the input, we replace the blank with the label.
Otherwise, we directly append the answer to the input.
"""
assert dataset_name in Column_to_check.keys(), \
f'Column_to_check for {dataset_name} is not configed in utils.py'
id_ = row[Column_to_check[dataset_name]['id']]
if Column_to_check[dataset_name].get('passage', None) is not None:
# rendering reading comprehension benchmarks
passage = row[Column_to_check[dataset_name]['passage']]
query = f'{passage}'
return {
'id': id_,
'query': query,
}
# multi-choice benchmarks are more complicated, sometimes need to fill-blanks
assert Column_to_check[dataset_name].get('input', None) is not None
input_ = row[Column_to_check[dataset_name]['input']]
label = Column_to_check[dataset_name]['label'](row)
def fill_blanks(question, answers, placeholder = '____'):
global label
if placeholder not in question:
return f'{question} {answers}'
if placeholder == '_':
return re.sub(r'_', answers, question)
num_blanks = re.findall(r'_{2,}', question)
if len(num_blanks) > 1:
if '，' in answers:
answers = [ans.strip() for ans in answers.split('，')]
elif ',' in answers:
answers =[ans.strip() for ans in answers.split(',')]
elif '；' in answers:
answers = [ans.strip() for ans in answers.split('；')]
else:
return None
raise ValueError(f'Cannot split {answers} into multiple answers. The question is {question}.')
if not len(num_blanks) == len(answers):
return None
raise ValueError(f'Number of blanks in {question} does not match the number of answers {answers}.')
else:
answers = [answers]
label = ' '.join(answers)
return re.sub(r'_{2,}', lambda match: answers.pop(0), question)
verbalize = {
'winogrande': lambda input_, label: fill_blanks(input_, label, '_'),
'ceval': lambda input_, label: fill_blanks(input_, label, '____'),
'ceval_train': lambda input_, label: fill_blanks(input_, label, '____'),
'mmlu': lambda input_, label: fill_blanks(input_, label),
'mmlu_train': lambda input_, label: fill_blanks(input_, label),
'hellaswag': lambda input_, label: f'{input_} {label}',
'ARC': lambda input_, label: f'{input_} {label}',
'commonsense_qa': lambda input_, label: f'{input_} {label}',
}
return {
'id': id_,
'input': input_,
'label': label,
'query': verbalize[dataset_name](input_, label),
}
def prepare_dataset(dataset_names, n = 500, config = None):
"""Load the datasets from Huggingface Hub, and randomly sample n samples from each dataset.
if n == 'all', then load all samples.
"""
dses = {}
for dataset_name in dataset_names:
assert dataset_name in Hf_Name_and_Split.keys(), \
f'Hf_Name_and_Split for {dataset_name} is not defined in utils.py'
if dataset_name in ['wiki_clean', 'bbc_clean', 'wiki_all', 'bbc_all']:
assert config is not None, 'The config is used to set a time range, if you use wiki_clean, bbc_clean, wiki_all, bbc_all you need to set a time with config = time'
ds = load_dataset(Hf_Name_and_Split[dataset_name]['hf_name'], config, split = Hf_Name_and_Split[dataset_name]['split'])
elif Hf_Name_and_Split[dataset_name].get('config', None) is not None:
ds = load_dataset(Hf_Name_and_Split[dataset_name]['hf_name'], Hf_Name_and_Split[dataset_name]['config'], split = Hf_Name_and_Split[dataset_name]['split'])
else:
ds = load_dataset(Hf_Name_and_Split[dataset_name]['hf_name'], split = Hf_Name_and_Split[dataset_name]['split'])
print(dataset_name)
print(ds)
print('=====================')
if n != 'all':
ds = random_sample_ds(ds, n = n)
dses[dataset_name] = ds
return dses
class CommonCrawlPresenceChecker:
"""Check whether a URL is present in a range of Common Crawl dumps.
But the Common Crawl API is almost broken recently, so we wont use it.
"""
def __init__(self, time_range = ('2017-01-01', '2021-01-01')) -> None:
import requests
self.requests = requests
self.all_dumps(time_range = time_range)
def all_dumps(self, time_range):
from datetime import datetime
start_date, end_date = datetime.strptime(time_range[0], '%Y-%m-%d'), datetime.strptime(time_range[1], '%Y-%m-%d')
response = self.requests.get('http://index.commoncrawl.org/collinfo.json')
dumps = response.json()
self.dumps_in_range = []
for dump in dumps:
year, week = dump['id'].split('-')[2], dump['id'].split('-')[3]
dump_date = datetime.fromisocalendar(year, week, 1)
if start_date <= dump_date <= end_date:
self.dumps_in_range.append(dump)
return self.dumps_in_range
def check_presence(self, url):
presences_in_dumps = []
for dump in self.dumps_in_range:
api = f"https://index.commoncrawl.org/{dump['id']}-index?url={url}&output=json"
response = self.requests.get(api)
try:
pages = [json.loads(x) for x in response.content.strip().split('\n')]
except:
presences_in_dumps.append(False)
continue
if pages:
presences_in_dumps.append(True)
else:
presences_in_dumps.append(False)
return any(presences_in_dumps)
def check_presence_by_wayback(self, url):
api_url = f"https://web.archive.org/cdx/search/cdx?url={url}"
response = self.requests.get(api_url)
results = response.text
if results:
return True
else:
return False
def __call__(self, url):
# The Common Crawl API is almost broken, so we use Wayback Machine instead
# return self.check_presence(url)
return self.check_presence_by_wayback(url)