file_utils.py 8.16 KB
Newer Older
thomwolf's avatar
thomwolf committed
1
2
3
4
5
"""
Utilities for working with the local dataset cache.
This file is adapted from the AllenNLP library at https://github.com/allenai/allennlp
Copyright by the AllenNLP authors.
"""
thomwolf's avatar
thomwolf committed
6
from __future__ import (absolute_import, division, print_function, unicode_literals)
thomwolf's avatar
thomwolf committed
7

thomwolf's avatar
thomwolf committed
8
import json
thomwolf's avatar
thomwolf committed
9
import logging
thomwolf's avatar
thomwolf committed
10
import os
thomwolf's avatar
thomwolf committed
11
12
13
import shutil
import tempfile
from functools import wraps
thomwolf's avatar
thomwolf committed
14
from hashlib import sha256
15
import sys
thomwolf's avatar
thomwolf committed
16
from io import open
thomwolf's avatar
thomwolf committed
17
18
19

import boto3
import requests
thomwolf's avatar
thomwolf committed
20
21
from botocore.exceptions import ClientError
from tqdm import tqdm
thomwolf's avatar
thomwolf committed
22

thomwolf's avatar
thomwolf committed
23
24
25
26
27
28
29
30
31
try:
    from urllib.parse import urlparse
except ImportError:
    from urlparse import urlparse

try:
    from pathlib import Path
    PYTORCH_PRETRAINED_BERT_CACHE = Path(os.getenv('PYTORCH_PRETRAINED_BERT_CACHE',
                                                   Path.home() / '.pytorch_pretrained_bert'))
32
except (AttributeError, ImportError):
thomwolf's avatar
thomwolf committed
33
34
    PYTORCH_PRETRAINED_BERT_CACHE = os.getenv('PYTORCH_PRETRAINED_BERT_CACHE',
                                              os.path.join(os.path.expanduser("~"), '.pytorch_pretrained_bert'))
thomwolf's avatar
thomwolf committed
35

36
37
38
CONFIG_NAME = "config.json"
WEIGHTS_NAME = "pytorch_model.bin"

thomwolf's avatar
thomwolf committed
39
logger = logging.getLogger(__name__)  # pylint: disable=invalid-name
thomwolf's avatar
thomwolf committed
40
41


thomwolf's avatar
thomwolf committed
42
def url_to_filename(url, etag=None):
thomwolf's avatar
thomwolf committed
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
    """
    Convert `url` into a hashed filename in a repeatable way.
    If `etag` is specified, append its hash to the url's, delimited
    by a period.
    """
    url_bytes = url.encode('utf-8')
    url_hash = sha256(url_bytes)
    filename = url_hash.hexdigest()

    if etag:
        etag_bytes = etag.encode('utf-8')
        etag_hash = sha256(etag_bytes)
        filename += '.' + etag_hash.hexdigest()

    return filename


thomwolf's avatar
thomwolf committed
60
def filename_to_url(filename, cache_dir=None):
thomwolf's avatar
thomwolf committed
61
62
    """
    Return the url and etag (which may be ``None``) stored for `filename`.
thomwolf's avatar
thomwolf committed
63
    Raise ``EnvironmentError`` if `filename` or its stored metadata do not exist.
thomwolf's avatar
thomwolf committed
64
65
66
    """
    if cache_dir is None:
        cache_dir = PYTORCH_PRETRAINED_BERT_CACHE
67
68
    if sys.version_info[0] == 3 and isinstance(cache_dir, Path):
        cache_dir = str(cache_dir)
thomwolf's avatar
thomwolf committed
69
70
71

    cache_path = os.path.join(cache_dir, filename)
    if not os.path.exists(cache_path):
thomwolf's avatar
thomwolf committed
72
        raise EnvironmentError("file {} not found".format(cache_path))
thomwolf's avatar
thomwolf committed
73
74
75

    meta_path = cache_path + '.json'
    if not os.path.exists(meta_path):
thomwolf's avatar
thomwolf committed
76
        raise EnvironmentError("file {} not found".format(meta_path))
thomwolf's avatar
thomwolf committed
77

thomwolf's avatar
thomwolf committed
78
    with open(meta_path, encoding="utf-8") as meta_file:
thomwolf's avatar
thomwolf committed
79
80
81
82
83
84
85
        metadata = json.load(meta_file)
    url = metadata['url']
    etag = metadata['etag']

    return url, etag


thomwolf's avatar
thomwolf committed
86
def cached_path(url_or_filename, cache_dir=None):
thomwolf's avatar
thomwolf committed
87
88
89
90
91
92
93
94
    """
    Given something that might be a URL (or might be a local path),
    determine which. If it's a URL, download the file and cache it, and
    return the path to the cached file. If it's already a local path,
    make sure the file exists and then return the path.
    """
    if cache_dir is None:
        cache_dir = PYTORCH_PRETRAINED_BERT_CACHE
95
96
97
98
    if sys.version_info[0] == 3 and isinstance(url_or_filename, Path):
        url_or_filename = str(url_or_filename)
    if sys.version_info[0] == 3 and isinstance(cache_dir, Path):
        cache_dir = str(cache_dir)
thomwolf's avatar
thomwolf committed
99
100
101
102
103
104
105
106
107
108
109

    parsed = urlparse(url_or_filename)

    if parsed.scheme in ('http', 'https', 's3'):
        # URL, so get it from the cache (downloading if necessary)
        return get_from_cache(url_or_filename, cache_dir)
    elif os.path.exists(url_or_filename):
        # File, and it exists.
        return url_or_filename
    elif parsed.scheme == '':
        # File, but it doesn't exist.
thomwolf's avatar
thomwolf committed
110
        raise EnvironmentError("file {} not found".format(url_or_filename))
thomwolf's avatar
thomwolf committed
111
112
113
114
115
    else:
        # Something unknown
        raise ValueError("unable to parse {} as a URL or as a local path".format(url_or_filename))


thomwolf's avatar
thomwolf committed
116
def split_s3_path(url):
thomwolf's avatar
thomwolf committed
117
118
119
120
121
122
123
124
125
126
127
128
    """Split a full s3 path into the bucket name and path."""
    parsed = urlparse(url)
    if not parsed.netloc or not parsed.path:
        raise ValueError("bad s3 path {}".format(url))
    bucket_name = parsed.netloc
    s3_path = parsed.path
    # Remove '/' at beginning of path.
    if s3_path.startswith("/"):
        s3_path = s3_path[1:]
    return bucket_name, s3_path


thomwolf's avatar
thomwolf committed
129
def s3_request(func):
thomwolf's avatar
thomwolf committed
130
131
132
133
134
135
    """
    Wrapper function for s3 requests in order to create more helpful error
    messages.
    """

    @wraps(func)
thomwolf's avatar
thomwolf committed
136
    def wrapper(url, *args, **kwargs):
thomwolf's avatar
thomwolf committed
137
138
139
140
        try:
            return func(url, *args, **kwargs)
        except ClientError as exc:
            if int(exc.response["Error"]["Code"]) == 404:
thomwolf's avatar
thomwolf committed
141
                raise EnvironmentError("file {} not found".format(url))
thomwolf's avatar
thomwolf committed
142
143
144
145
146
147
148
            else:
                raise

    return wrapper


@s3_request
thomwolf's avatar
thomwolf committed
149
def s3_etag(url):
thomwolf's avatar
thomwolf committed
150
151
152
153
154
155
156
157
    """Check ETag on S3 object."""
    s3_resource = boto3.resource("s3")
    bucket_name, s3_path = split_s3_path(url)
    s3_object = s3_resource.Object(bucket_name, s3_path)
    return s3_object.e_tag


@s3_request
thomwolf's avatar
thomwolf committed
158
def s3_get(url, temp_file):
thomwolf's avatar
thomwolf committed
159
160
161
162
163
164
    """Pull a file directly from S3."""
    s3_resource = boto3.resource("s3")
    bucket_name, s3_path = split_s3_path(url)
    s3_resource.Bucket(bucket_name).download_fileobj(s3_path, temp_file)


thomwolf's avatar
thomwolf committed
165
def http_get(url, temp_file):
thomwolf's avatar
thomwolf committed
166
167
168
169
170
171
172
173
174
175
176
    req = requests.get(url, stream=True)
    content_length = req.headers.get('Content-Length')
    total = int(content_length) if content_length is not None else None
    progress = tqdm(unit="B", total=total)
    for chunk in req.iter_content(chunk_size=1024):
        if chunk: # filter out keep-alive new chunks
            progress.update(len(chunk))
            temp_file.write(chunk)
    progress.close()


thomwolf's avatar
thomwolf committed
177
def get_from_cache(url, cache_dir=None):
thomwolf's avatar
thomwolf committed
178
179
180
181
182
183
    """
    Given a URL, look for the corresponding dataset in the local cache.
    If it's not there, download it. Then return the path to the cached file.
    """
    if cache_dir is None:
        cache_dir = PYTORCH_PRETRAINED_BERT_CACHE
184
185
    if sys.version_info[0] == 3 and isinstance(cache_dir, Path):
        cache_dir = str(cache_dir)
thomwolf's avatar
thomwolf committed
186

thomwolf's avatar
thomwolf committed
187
188
    if not os.path.exists(cache_dir):
        os.makedirs(cache_dir)
thomwolf's avatar
thomwolf committed
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

    # Get eTag to add to filename, if it exists.
    if url.startswith("s3://"):
        etag = s3_etag(url)
    else:
        response = requests.head(url, allow_redirects=True)
        if response.status_code != 200:
            raise IOError("HEAD request failed for url {} with status code {}"
                          .format(url, response.status_code))
        etag = response.headers.get("ETag")

    filename = url_to_filename(url, etag)

    # get cache path to put the file
    cache_path = os.path.join(cache_dir, filename)

    if not os.path.exists(cache_path):
        # Download to temporary file, then copy to cache dir once finished.
        # Otherwise you get corrupt cache entries if the download gets interrupted.
        with tempfile.NamedTemporaryFile() as temp_file:
            logger.info("%s not found in cache, downloading to %s", url, temp_file.name)

            # GET file object
            if url.startswith("s3://"):
                s3_get(url, temp_file)
            else:
                http_get(url, temp_file)

            # we are copying the file before closing it, so flush to avoid truncation
            temp_file.flush()
            # shutil.copyfileobj() starts at the current position, so go to the start
            temp_file.seek(0)

            logger.info("copying %s to cache at %s", temp_file.name, cache_path)
            with open(cache_path, 'wb') as cache_file:
                shutil.copyfileobj(temp_file, cache_file)

            logger.info("creating metadata file for %s", cache_path)
            meta = {'url': url, 'etag': etag}
            meta_path = cache_path + '.json'
thomwolf's avatar
thomwolf committed
229
            with open(meta_path, 'w', encoding="utf-8") as meta_file:
thomwolf's avatar
thomwolf committed
230
231
232
233
234
235
236
                json.dump(meta, meta_file)

            logger.info("removing temp file %s", temp_file.name)

    return cache_path


thomwolf's avatar
thomwolf committed
237
def read_set_from_file(filename):
thomwolf's avatar
thomwolf committed
238
239
240
241
242
    '''
    Extract a de-duped collection (set) of text from a file.
    Expected file format is one item per line.
    '''
    collection = set()
243
    with open(filename, 'r', encoding='utf-8') as file_:
thomwolf's avatar
thomwolf committed
244
245
246
247
248
        for line in file_:
            collection.add(line.rstrip())
    return collection


thomwolf's avatar
thomwolf committed
249
def get_file_extension(path, dot=True, lower=True):
thomwolf's avatar
thomwolf committed
250
251
252
    ext = os.path.splitext(path)[1]
    ext = ext if dot else ext[1:]
    return ext.lower() if lower else ext