file_utils.py 8.89 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

8
import sys
thomwolf's avatar
thomwolf committed
9
import json
thomwolf's avatar
thomwolf committed
10
import logging
thomwolf's avatar
thomwolf committed
11
import os
thomwolf's avatar
thomwolf committed
12
13
import shutil
import tempfile
14
import fnmatch
thomwolf's avatar
thomwolf committed
15
from functools import wraps
thomwolf's avatar
thomwolf committed
16
17
from hashlib import sha256
from io import open
thomwolf's avatar
thomwolf committed
18
19
20

import boto3
import requests
thomwolf's avatar
thomwolf committed
21
22
from botocore.exceptions import ClientError
from tqdm import tqdm
thomwolf's avatar
thomwolf committed
23

24
25
26
27
28
29
30
try:
    from torch.hub import _get_torch_home
    torch_cache_home = _get_torch_home()
except ImportError:
    torch_cache_home = os.path.expanduser(
        os.getenv('TORCH_HOME', os.path.join(
            os.getenv('XDG_CACHE_HOME', '~/.cache'), 'torch')))
thomwolf's avatar
thomwolf committed
31
default_cache_path = os.path.join(torch_cache_home, 'pytorch_transformers')
32

thomwolf's avatar
thomwolf committed
33
34
35
36
37
38
39
try:
    from urllib.parse import urlparse
except ImportError:
    from urlparse import urlparse

try:
    from pathlib import Path
40
    PYTORCH_PRETRAINED_BERT_CACHE = Path(
41
        os.getenv('PYTORCH_TRANSFORMERS_CACHE', os.getenv('PYTORCH_PRETRAINED_BERT_CACHE', default_cache_path)))
42
except (AttributeError, ImportError):
43
44
45
46
47
    PYTORCH_PRETRAINED_BERT_CACHE = os.getenv('PYTORCH_TRANSFORMERS_CACHE',
                                              os.getenv('PYTORCH_PRETRAINED_BERT_CACHE',
                                                        default_cache_path))

PYTORCH_TRANSFORMERS_CACHE = PYTORCH_PRETRAINED_BERT_CACHE  # Kept for backward compatibility
thomwolf's avatar
thomwolf committed
48

thomwolf's avatar
thomwolf committed
49
logger = logging.getLogger(__name__)  # pylint: disable=invalid-name
thomwolf's avatar
thomwolf committed
50
51


thomwolf's avatar
thomwolf committed
52
def url_to_filename(url, etag=None):
thomwolf's avatar
thomwolf committed
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
    """
    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
70
def filename_to_url(filename, cache_dir=None):
thomwolf's avatar
thomwolf committed
71
72
    """
    Return the url and etag (which may be ``None``) stored for `filename`.
thomwolf's avatar
thomwolf committed
73
    Raise ``EnvironmentError`` if `filename` or its stored metadata do not exist.
thomwolf's avatar
thomwolf committed
74
75
    """
    if cache_dir is None:
76
        cache_dir = PYTORCH_TRANSFORMERS_CACHE
77
78
    if sys.version_info[0] == 3 and isinstance(cache_dir, Path):
        cache_dir = str(cache_dir)
thomwolf's avatar
thomwolf committed
79
80
81

    cache_path = os.path.join(cache_dir, filename)
    if not os.path.exists(cache_path):
thomwolf's avatar
thomwolf committed
82
        raise EnvironmentError("file {} not found".format(cache_path))
thomwolf's avatar
thomwolf committed
83
84
85

    meta_path = cache_path + '.json'
    if not os.path.exists(meta_path):
thomwolf's avatar
thomwolf committed
86
        raise EnvironmentError("file {} not found".format(meta_path))
thomwolf's avatar
thomwolf committed
87

thomwolf's avatar
thomwolf committed
88
    with open(meta_path, encoding="utf-8") as meta_file:
thomwolf's avatar
thomwolf committed
89
90
91
92
93
94
95
        metadata = json.load(meta_file)
    url = metadata['url']
    etag = metadata['etag']

    return url, etag


thomwolf's avatar
thomwolf committed
96
def cached_path(url_or_filename, cache_dir=None):
thomwolf's avatar
thomwolf committed
97
98
99
100
101
102
103
    """
    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:
104
        cache_dir = PYTORCH_TRANSFORMERS_CACHE
105
106
107
108
    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
109
110
111
112
113
114
115
116
117
118
119

    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
120
        raise EnvironmentError("file {} not found".format(url_or_filename))
thomwolf's avatar
thomwolf committed
121
122
123
124
125
    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
126
def split_s3_path(url):
thomwolf's avatar
thomwolf committed
127
128
129
130
131
132
133
134
135
136
137
138
    """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
139
def s3_request(func):
thomwolf's avatar
thomwolf committed
140
141
142
143
144
145
    """
    Wrapper function for s3 requests in order to create more helpful error
    messages.
    """

    @wraps(func)
thomwolf's avatar
thomwolf committed
146
    def wrapper(url, *args, **kwargs):
thomwolf's avatar
thomwolf committed
147
148
149
150
        try:
            return func(url, *args, **kwargs)
        except ClientError as exc:
            if int(exc.response["Error"]["Code"]) == 404:
thomwolf's avatar
thomwolf committed
151
                raise EnvironmentError("file {} not found".format(url))
thomwolf's avatar
thomwolf committed
152
153
154
155
156
157
158
            else:
                raise

    return wrapper


@s3_request
thomwolf's avatar
thomwolf committed
159
def s3_etag(url):
thomwolf's avatar
thomwolf committed
160
161
162
163
164
165
166
167
    """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
168
def s3_get(url, temp_file):
thomwolf's avatar
thomwolf committed
169
170
171
172
173
174
    """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
175
def http_get(url, temp_file):
thomwolf's avatar
thomwolf committed
176
177
178
179
180
181
182
183
184
185
186
    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
187
def get_from_cache(url, cache_dir=None):
thomwolf's avatar
thomwolf committed
188
189
190
191
192
    """
    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:
193
        cache_dir = PYTORCH_TRANSFORMERS_CACHE
194
195
    if sys.version_info[0] == 3 and isinstance(cache_dir, Path):
        cache_dir = str(cache_dir)
196
197
    if sys.version_info[0] == 2 and not isinstance(cache_dir, str):
        cache_dir = str(cache_dir)
thomwolf's avatar
thomwolf committed
198

thomwolf's avatar
thomwolf committed
199
200
    if not os.path.exists(cache_dir):
        os.makedirs(cache_dir)
thomwolf's avatar
thomwolf committed
201
202
203
204
205

    # Get eTag to add to filename, if it exists.
    if url.startswith("s3://"):
        etag = s3_etag(url)
    else:
206
207
208
209
210
211
212
213
        try:
            response = requests.head(url, allow_redirects=True)
            if response.status_code != 200:
                etag = None
            else:
                etag = response.headers.get("ETag")
        except EnvironmentError:
            etag = None
thomwolf's avatar
thomwolf committed
214

215
216
    if sys.version_info[0] == 2 and etag is not None:
        etag = etag.decode('utf-8')
thomwolf's avatar
thomwolf committed
217
218
219
220
221
    filename = url_to_filename(url, etag)

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

222
223
224
225
226
227
228
229
    # If we don't have a connection (etag is None) and can't identify the file
    # try to get the last downloaded one
    if not os.path.exists(cache_path) and etag is None:
        matching_files = fnmatch.filter(os.listdir(cache_dir), filename + '.*')
        matching_files = list(filter(lambda s: not s.endswith('.json'), matching_files))
        if matching_files:
            cache_path = os.path.join(cache_dir, matching_files[-1])

thomwolf's avatar
thomwolf committed
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
    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'
254
            with open(meta_path, 'w') as meta_file:
thomwolf's avatar
thomwolf committed
255
256
257
258
                output_string = json.dumps(meta)
                if sys.version_info[0] == 2 and isinstance(output_string, str):
                    output_string = unicode(output_string, 'utf-8')  # The beauty of python 2
                meta_file.write(output_string)
thomwolf's avatar
thomwolf committed
259
260
261
262

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

    return cache_path