blob: 8aec1d51b5a174153871e074636a900af1296a4b [file] [log] [blame]
Craig Citro751b7fb2014-09-23 11:20:38 -07001# Copyright 2014 Google Inc. All Rights Reserved.
John Asmuth864311d2014-04-24 15:46:08 -04002#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7# http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14
15"""Classes to encapsulate a single HTTP request.
16
17The classes implement a command pattern, with every
18object supporting an execute() method that does the
cspeidelfbaf9d72018-05-10 12:50:12 -060019actual HTTP request.
John Asmuth864311d2014-04-24 15:46:08 -040020"""
INADA Naoki0bceb332014-08-20 15:27:52 +090021from __future__ import absolute_import
INADA Naokie4ea1a92015-03-04 03:45:42 +090022import six
eesheeshc6425a02016-02-12 15:07:06 +000023from six.moves import http_client
INADA Naokie4ea1a92015-03-04 03:45:42 +090024from six.moves import range
John Asmuth864311d2014-04-24 15:46:08 -040025
Bu Sun Kim66bb32c2019-10-30 10:11:58 -070026__author__ = "jcgregorio@google.com (Joe Gregorio)"
John Asmuth864311d2014-04-24 15:46:08 -040027
Pat Ferateed9affd2015-03-03 16:03:15 -080028from six import BytesIO, StringIO
Pat Ferated5b61bd2015-03-03 16:04:11 -080029from six.moves.urllib.parse import urlparse, urlunparse, quote, unquote
Pat Ferateed9affd2015-03-03 16:03:15 -080030
John Asmuth864311d2014-04-24 15:46:08 -040031import base64
32import copy
33import gzip
34import httplib2
Craig Citro6ae34d72014-08-18 23:10:09 -070035import json
John Asmuth864311d2014-04-24 15:46:08 -040036import logging
John Asmuth864311d2014-04-24 15:46:08 -040037import mimetypes
38import os
39import random
eesheeshc6425a02016-02-12 15:07:06 +000040import socket
John Asmuth864311d2014-04-24 15:46:08 -040041import sys
42import time
John Asmuth864311d2014-04-24 15:46:08 -040043import uuid
44
Tay Ray Chuan3146c922016-04-20 16:38:19 +000045# TODO(issue 221): Remove this conditional import jibbajabba.
46try:
Bu Sun Kim66bb32c2019-10-30 10:11:58 -070047 import ssl
Tay Ray Chuan3146c922016-04-20 16:38:19 +000048except ImportError:
Bu Sun Kim66bb32c2019-10-30 10:11:58 -070049 _ssl_SSLError = object()
Tay Ray Chuan3146c922016-04-20 16:38:19 +000050else:
Bu Sun Kim66bb32c2019-10-30 10:11:58 -070051 _ssl_SSLError = ssl.SSLError
Tay Ray Chuan3146c922016-04-20 16:38:19 +000052
John Asmuth864311d2014-04-24 15:46:08 -040053from email.generator import Generator
54from email.mime.multipart import MIMEMultipart
55from email.mime.nonmultipart import MIMENonMultipart
56from email.parser import FeedParser
Pat Ferateb240c172015-03-03 16:23:51 -080057
Helen Koikede13e3b2018-04-26 16:05:16 -030058from googleapiclient import _helpers as util
Jon Wayne Parrott6755f612016-08-15 10:52:26 -070059
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -070060from googleapiclient import _auth
Pat Ferateb240c172015-03-03 16:23:51 -080061from googleapiclient.errors import BatchError
62from googleapiclient.errors import HttpError
63from googleapiclient.errors import InvalidChunkSizeError
64from googleapiclient.errors import ResumableUploadError
65from googleapiclient.errors import UnexpectedBodyError
66from googleapiclient.errors import UnexpectedMethodError
67from googleapiclient.model import JsonModel
John Asmuth864311d2014-04-24 15:46:08 -040068
69
Emmett Butler09699152016-02-08 14:26:00 -080070LOGGER = logging.getLogger(__name__)
71
Bu Sun Kim66bb32c2019-10-30 10:11:58 -070072DEFAULT_CHUNK_SIZE = 100 * 1024 * 1024
John Asmuth864311d2014-04-24 15:46:08 -040073
74MAX_URI_LENGTH = 2048
75
Xinan Line2dccec2018-12-07 05:28:33 +090076MAX_BATCH_LIMIT = 1000
77
eesheeshc6425a02016-02-12 15:07:06 +000078_TOO_MANY_REQUESTS = 429
79
Igor Maravić22435292017-01-19 22:28:22 +010080DEFAULT_HTTP_TIMEOUT_SEC = 60
81
Bu Sun Kim66bb32c2019-10-30 10:11:58 -070082_LEGACY_BATCH_URI = "https://www.googleapis.com/batch"
Jon Wayne Parrottbae748a2018-03-28 10:21:12 -070083
Damian Gadomskic7516a22020-03-23 20:39:21 +010084if six.PY2:
85 # That's a builtin python3 exception, nonexistent in python2.
86 # Defined to None to avoid NameError while trying to catch it
87 ConnectionError = None
88
eesheeshc6425a02016-02-12 15:07:06 +000089
90def _should_retry_response(resp_status, content):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -070091 """Determines whether a response should be retried.
eesheeshc6425a02016-02-12 15:07:06 +000092
93 Args:
94 resp_status: The response status received.
Nilayan Bhattacharya90ffb852017-12-05 15:30:32 -080095 content: The response content body.
eesheeshc6425a02016-02-12 15:07:06 +000096
97 Returns:
98 True if the response should be retried, otherwise False.
99 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700100 # Retry on 5xx errors.
101 if resp_status >= 500:
102 return True
eesheeshc6425a02016-02-12 15:07:06 +0000103
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700104 # Retry on 429 errors.
105 if resp_status == _TOO_MANY_REQUESTS:
106 return True
eesheeshc6425a02016-02-12 15:07:06 +0000107
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700108 # For 403 errors, we have to check for the `reason` in the response to
109 # determine if we should retry.
110 if resp_status == six.moves.http_client.FORBIDDEN:
111 # If there's no details about the 403 type, don't retry.
112 if not content:
113 return False
eesheeshc6425a02016-02-12 15:07:06 +0000114
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700115 # Content is in JSON format.
116 try:
117 data = json.loads(content.decode("utf-8"))
118 if isinstance(data, dict):
119 reason = data["error"]["errors"][0]["reason"]
120 else:
121 reason = data[0]["error"]["errors"]["reason"]
122 except (UnicodeDecodeError, ValueError, KeyError):
123 LOGGER.warning("Invalid JSON content from response: %s", content)
124 return False
eesheeshc6425a02016-02-12 15:07:06 +0000125
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700126 LOGGER.warning('Encountered 403 Forbidden with reason "%s"', reason)
eesheeshc6425a02016-02-12 15:07:06 +0000127
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700128 # Only retry on rate limit related failures.
129 if reason in ("userRateLimitExceeded", "rateLimitExceeded"):
130 return True
eesheeshc6425a02016-02-12 15:07:06 +0000131
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700132 # Everything else is a success or non-retriable so break.
133 return False
eesheeshc6425a02016-02-12 15:07:06 +0000134
John Asmuth864311d2014-04-24 15:46:08 -0400135
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700136def _retry_request(
137 http, num_retries, req_type, sleep, rand, uri, method, *args, **kwargs
138):
139 """Retries an HTTP request multiple times while handling errors.
Sergiy Byelozyorov703c92c2015-12-21 23:27:48 +0100140
141 If after all retries the request still fails, last error is either returned as
142 return value (for HTTP 5xx errors) or thrown (for ssl.SSLError).
143
144 Args:
145 http: Http object to be used to execute request.
146 num_retries: Maximum number of retries.
147 req_type: Type of the request (used for logging retries).
148 sleep, rand: Functions to sleep for random time between retries.
149 uri: URI to be requested.
150 method: HTTP method to be used.
151 args, kwargs: Additional arguments passed to http.request.
152
153 Returns:
154 resp, content - Response from the http request (may be HTTP 5xx).
155 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700156 resp = None
157 content = None
158 exception = None
159 for retry_num in range(num_retries + 1):
160 if retry_num > 0:
161 # Sleep before retrying.
162 sleep_time = rand() * 2 ** retry_num
163 LOGGER.warning(
164 "Sleeping %.2f seconds before retry %d of %d for %s: %s %s, after %s",
165 sleep_time,
166 retry_num,
167 num_retries,
168 req_type,
169 method,
170 uri,
171 resp.status if resp else exception,
172 )
173 sleep(sleep_time)
Sergiy Byelozyorov703c92c2015-12-21 23:27:48 +0100174
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700175 try:
176 exception = None
177 resp, content = http.request(uri, method, *args, **kwargs)
178 # Retry on SSL errors and socket timeout errors.
179 except _ssl_SSLError as ssl_error:
180 exception = ssl_error
181 except socket.timeout as socket_timeout:
182 # It's important that this be before socket.error as it's a subclass
183 # socket.timeout has no errorcode
184 exception = socket_timeout
Damian Gadomskic7516a22020-03-23 20:39:21 +0100185 except ConnectionError as connection_error:
186 # Needs to be before socket.error as it's a subclass of
187 # OSError (socket.error)
188 exception = connection_error
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700189 except socket.error as socket_error:
190 # errno's contents differ by platform, so we have to match by name.
191 if socket.errno.errorcode.get(socket_error.errno) not in {
192 "WSAETIMEDOUT",
193 "ETIMEDOUT",
194 "EPIPE",
195 "ECONNABORTED",
196 }:
197 raise
198 exception = socket_error
199 except httplib2.ServerNotFoundError as server_not_found_error:
200 exception = server_not_found_error
eesheeshc6425a02016-02-12 15:07:06 +0000201
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700202 if exception:
203 if retry_num == num_retries:
204 raise exception
205 else:
206 continue
eesheeshc6425a02016-02-12 15:07:06 +0000207
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700208 if not _should_retry_response(resp.status, content):
209 break
Sergiy Byelozyorov703c92c2015-12-21 23:27:48 +0100210
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700211 return resp, content
Sergiy Byelozyorov703c92c2015-12-21 23:27:48 +0100212
213
John Asmuth864311d2014-04-24 15:46:08 -0400214class MediaUploadProgress(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700215 """Status of a resumable upload."""
John Asmuth864311d2014-04-24 15:46:08 -0400216
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700217 def __init__(self, resumable_progress, total_size):
218 """Constructor.
John Asmuth864311d2014-04-24 15:46:08 -0400219
220 Args:
221 resumable_progress: int, bytes sent so far.
222 total_size: int, total bytes in complete upload, or None if the total
223 upload size isn't known ahead of time.
224 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700225 self.resumable_progress = resumable_progress
226 self.total_size = total_size
John Asmuth864311d2014-04-24 15:46:08 -0400227
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700228 def progress(self):
229 """Percent of upload completed, as a float.
John Asmuth864311d2014-04-24 15:46:08 -0400230
231 Returns:
232 the percentage complete as a float, returning 0.0 if the total size of
233 the upload is unknown.
234 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700235 if self.total_size is not None and self.total_size != 0:
236 return float(self.resumable_progress) / float(self.total_size)
237 else:
238 return 0.0
John Asmuth864311d2014-04-24 15:46:08 -0400239
240
241class MediaDownloadProgress(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700242 """Status of a resumable download."""
John Asmuth864311d2014-04-24 15:46:08 -0400243
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700244 def __init__(self, resumable_progress, total_size):
245 """Constructor.
John Asmuth864311d2014-04-24 15:46:08 -0400246
247 Args:
248 resumable_progress: int, bytes received so far.
249 total_size: int, total bytes in complete download.
250 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700251 self.resumable_progress = resumable_progress
252 self.total_size = total_size
John Asmuth864311d2014-04-24 15:46:08 -0400253
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700254 def progress(self):
255 """Percent of download completed, as a float.
John Asmuth864311d2014-04-24 15:46:08 -0400256
257 Returns:
258 the percentage complete as a float, returning 0.0 if the total size of
259 the download is unknown.
260 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700261 if self.total_size is not None and self.total_size != 0:
262 return float(self.resumable_progress) / float(self.total_size)
263 else:
264 return 0.0
John Asmuth864311d2014-04-24 15:46:08 -0400265
266
267class MediaUpload(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700268 """Describes a media object to upload.
John Asmuth864311d2014-04-24 15:46:08 -0400269
270 Base class that defines the interface of MediaUpload subclasses.
271
272 Note that subclasses of MediaUpload may allow you to control the chunksize
273 when uploading a media object. It is important to keep the size of the chunk
274 as large as possible to keep the upload efficient. Other factors may influence
275 the size of the chunk you use, particularly if you are working in an
276 environment where individual HTTP requests may have a hardcoded time limit,
277 such as under certain classes of requests under Google App Engine.
278
279 Streams are io.Base compatible objects that support seek(). Some MediaUpload
280 subclasses support using streams directly to upload data. Support for
281 streaming may be indicated by a MediaUpload sub-class and if appropriate for a
282 platform that stream will be used for uploading the media object. The support
283 for streaming is indicated by has_stream() returning True. The stream() method
284 should return an io.Base object that supports seek(). On platforms where the
285 underlying httplib module supports streaming, for example Python 2.6 and
286 later, the stream will be passed into the http library which will result in
287 less memory being used and possibly faster uploads.
288
289 If you need to upload media that can't be uploaded using any of the existing
290 MediaUpload sub-class then you can sub-class MediaUpload for your particular
291 needs.
292 """
293
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700294 def chunksize(self):
295 """Chunk size for resumable uploads.
John Asmuth864311d2014-04-24 15:46:08 -0400296
297 Returns:
298 Chunk size in bytes.
299 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700300 raise NotImplementedError()
John Asmuth864311d2014-04-24 15:46:08 -0400301
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700302 def mimetype(self):
303 """Mime type of the body.
John Asmuth864311d2014-04-24 15:46:08 -0400304
305 Returns:
306 Mime type.
307 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700308 return "application/octet-stream"
John Asmuth864311d2014-04-24 15:46:08 -0400309
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700310 def size(self):
311 """Size of upload.
John Asmuth864311d2014-04-24 15:46:08 -0400312
313 Returns:
314 Size of the body, or None of the size is unknown.
315 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700316 return None
John Asmuth864311d2014-04-24 15:46:08 -0400317
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700318 def resumable(self):
319 """Whether this upload is resumable.
John Asmuth864311d2014-04-24 15:46:08 -0400320
321 Returns:
322 True if resumable upload or False.
323 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700324 return False
John Asmuth864311d2014-04-24 15:46:08 -0400325
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700326 def getbytes(self, begin, end):
327 """Get bytes from the media.
John Asmuth864311d2014-04-24 15:46:08 -0400328
329 Args:
330 begin: int, offset from beginning of file.
331 length: int, number of bytes to read, starting at begin.
332
333 Returns:
334 A string of bytes read. May be shorter than length if EOF was reached
335 first.
336 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700337 raise NotImplementedError()
John Asmuth864311d2014-04-24 15:46:08 -0400338
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700339 def has_stream(self):
340 """Does the underlying upload support a streaming interface.
John Asmuth864311d2014-04-24 15:46:08 -0400341
342 Streaming means it is an io.IOBase subclass that supports seek, i.e.
343 seekable() returns True.
344
345 Returns:
346 True if the call to stream() will return an instance of a seekable io.Base
347 subclass.
348 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700349 return False
John Asmuth864311d2014-04-24 15:46:08 -0400350
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700351 def stream(self):
352 """A stream interface to the data being uploaded.
John Asmuth864311d2014-04-24 15:46:08 -0400353
354 Returns:
355 The returned value is an io.IOBase subclass that supports seek, i.e.
356 seekable() returns True.
357 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700358 raise NotImplementedError()
John Asmuth864311d2014-04-24 15:46:08 -0400359
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700360 @util.positional(1)
361 def _to_json(self, strip=None):
362 """Utility function for creating a JSON representation of a MediaUpload.
John Asmuth864311d2014-04-24 15:46:08 -0400363
364 Args:
365 strip: array, An array of names of members to not include in the JSON.
366
367 Returns:
368 string, a JSON representation of this instance, suitable to pass to
369 from_json().
370 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700371 t = type(self)
372 d = copy.copy(self.__dict__)
373 if strip is not None:
374 for member in strip:
375 del d[member]
376 d["_class"] = t.__name__
377 d["_module"] = t.__module__
378 return json.dumps(d)
John Asmuth864311d2014-04-24 15:46:08 -0400379
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700380 def to_json(self):
381 """Create a JSON representation of an instance of MediaUpload.
John Asmuth864311d2014-04-24 15:46:08 -0400382
383 Returns:
384 string, a JSON representation of this instance, suitable to pass to
385 from_json().
386 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700387 return self._to_json()
John Asmuth864311d2014-04-24 15:46:08 -0400388
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700389 @classmethod
390 def new_from_json(cls, s):
391 """Utility class method to instantiate a MediaUpload subclass from a JSON
John Asmuth864311d2014-04-24 15:46:08 -0400392 representation produced by to_json().
393
394 Args:
395 s: string, JSON from to_json().
396
397 Returns:
398 An instance of the subclass of MediaUpload that was serialized with
399 to_json().
400 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700401 data = json.loads(s)
402 # Find and call the right classmethod from_json() to restore the object.
403 module = data["_module"]
404 m = __import__(module, fromlist=module.split(".")[:-1])
405 kls = getattr(m, data["_class"])
406 from_json = getattr(kls, "from_json")
407 return from_json(s)
John Asmuth864311d2014-04-24 15:46:08 -0400408
409
410class MediaIoBaseUpload(MediaUpload):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700411 """A MediaUpload for a io.Base objects.
John Asmuth864311d2014-04-24 15:46:08 -0400412
413 Note that the Python file object is compatible with io.Base and can be used
414 with this class also.
415
Pat Ferateed9affd2015-03-03 16:03:15 -0800416 fh = BytesIO('...Some data to upload...')
John Asmuth864311d2014-04-24 15:46:08 -0400417 media = MediaIoBaseUpload(fh, mimetype='image/png',
418 chunksize=1024*1024, resumable=True)
419 farm.animals().insert(
420 id='cow',
421 name='cow.png',
422 media_body=media).execute()
423
424 Depending on the platform you are working on, you may pass -1 as the
425 chunksize, which indicates that the entire file should be uploaded in a single
426 request. If the underlying platform supports streams, such as Python 2.6 or
427 later, then this can be very efficient as it avoids multiple connections, and
428 also avoids loading the entire file into memory before sending it. Note that
429 Google App Engine has a 5MB limit on request size, so you should never set
430 your chunksize larger than 5MB, or to -1.
431 """
432
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700433 @util.positional(3)
434 def __init__(self, fd, mimetype, chunksize=DEFAULT_CHUNK_SIZE, resumable=False):
435 """Constructor.
John Asmuth864311d2014-04-24 15:46:08 -0400436
437 Args:
438 fd: io.Base or file object, The source of the bytes to upload. MUST be
439 opened in blocking mode, do not use streams opened in non-blocking mode.
440 The given stream must be seekable, that is, it must be able to call
441 seek() on fd.
442 mimetype: string, Mime-type of the file.
443 chunksize: int, File will be uploaded in chunks of this many bytes. Only
444 used if resumable=True. Pass in a value of -1 if the file is to be
445 uploaded as a single chunk. Note that Google App Engine has a 5MB limit
446 on request size, so you should never set your chunksize larger than 5MB,
447 or to -1.
448 resumable: bool, True if this is a resumable upload. False means upload
449 in a single request.
450 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700451 super(MediaIoBaseUpload, self).__init__()
452 self._fd = fd
453 self._mimetype = mimetype
454 if not (chunksize == -1 or chunksize > 0):
455 raise InvalidChunkSizeError()
456 self._chunksize = chunksize
457 self._resumable = resumable
John Asmuth864311d2014-04-24 15:46:08 -0400458
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700459 self._fd.seek(0, os.SEEK_END)
460 self._size = self._fd.tell()
John Asmuth864311d2014-04-24 15:46:08 -0400461
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700462 def chunksize(self):
463 """Chunk size for resumable uploads.
John Asmuth864311d2014-04-24 15:46:08 -0400464
465 Returns:
466 Chunk size in bytes.
467 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700468 return self._chunksize
John Asmuth864311d2014-04-24 15:46:08 -0400469
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700470 def mimetype(self):
471 """Mime type of the body.
John Asmuth864311d2014-04-24 15:46:08 -0400472
473 Returns:
474 Mime type.
475 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700476 return self._mimetype
John Asmuth864311d2014-04-24 15:46:08 -0400477
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700478 def size(self):
479 """Size of upload.
John Asmuth864311d2014-04-24 15:46:08 -0400480
481 Returns:
482 Size of the body, or None of the size is unknown.
483 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700484 return self._size
John Asmuth864311d2014-04-24 15:46:08 -0400485
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700486 def resumable(self):
487 """Whether this upload is resumable.
John Asmuth864311d2014-04-24 15:46:08 -0400488
489 Returns:
490 True if resumable upload or False.
491 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700492 return self._resumable
John Asmuth864311d2014-04-24 15:46:08 -0400493
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700494 def getbytes(self, begin, length):
495 """Get bytes from the media.
John Asmuth864311d2014-04-24 15:46:08 -0400496
497 Args:
498 begin: int, offset from beginning of file.
499 length: int, number of bytes to read, starting at begin.
500
501 Returns:
502 A string of bytes read. May be shorted than length if EOF was reached
503 first.
504 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700505 self._fd.seek(begin)
506 return self._fd.read(length)
John Asmuth864311d2014-04-24 15:46:08 -0400507
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700508 def has_stream(self):
509 """Does the underlying upload support a streaming interface.
John Asmuth864311d2014-04-24 15:46:08 -0400510
511 Streaming means it is an io.IOBase subclass that supports seek, i.e.
512 seekable() returns True.
513
514 Returns:
515 True if the call to stream() will return an instance of a seekable io.Base
516 subclass.
517 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700518 return True
John Asmuth864311d2014-04-24 15:46:08 -0400519
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700520 def stream(self):
521 """A stream interface to the data being uploaded.
John Asmuth864311d2014-04-24 15:46:08 -0400522
523 Returns:
524 The returned value is an io.IOBase subclass that supports seek, i.e.
525 seekable() returns True.
526 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700527 return self._fd
John Asmuth864311d2014-04-24 15:46:08 -0400528
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700529 def to_json(self):
530 """This upload type is not serializable."""
531 raise NotImplementedError("MediaIoBaseUpload is not serializable.")
John Asmuth864311d2014-04-24 15:46:08 -0400532
533
534class MediaFileUpload(MediaIoBaseUpload):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700535 """A MediaUpload for a file.
John Asmuth864311d2014-04-24 15:46:08 -0400536
537 Construct a MediaFileUpload and pass as the media_body parameter of the
538 method. For example, if we had a service that allowed uploading images:
539
John Asmuth864311d2014-04-24 15:46:08 -0400540 media = MediaFileUpload('cow.png', mimetype='image/png',
541 chunksize=1024*1024, resumable=True)
542 farm.animals().insert(
543 id='cow',
544 name='cow.png',
545 media_body=media).execute()
546
547 Depending on the platform you are working on, you may pass -1 as the
548 chunksize, which indicates that the entire file should be uploaded in a single
549 request. If the underlying platform supports streams, such as Python 2.6 or
550 later, then this can be very efficient as it avoids multiple connections, and
551 also avoids loading the entire file into memory before sending it. Note that
552 Google App Engine has a 5MB limit on request size, so you should never set
553 your chunksize larger than 5MB, or to -1.
554 """
555
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700556 @util.positional(2)
557 def __init__(
558 self, filename, mimetype=None, chunksize=DEFAULT_CHUNK_SIZE, resumable=False
559 ):
560 """Constructor.
John Asmuth864311d2014-04-24 15:46:08 -0400561
562 Args:
563 filename: string, Name of the file.
564 mimetype: string, Mime-type of the file. If None then a mime-type will be
565 guessed from the file extension.
566 chunksize: int, File will be uploaded in chunks of this many bytes. Only
567 used if resumable=True. Pass in a value of -1 if the file is to be
568 uploaded in a single chunk. Note that Google App Engine has a 5MB limit
569 on request size, so you should never set your chunksize larger than 5MB,
570 or to -1.
571 resumable: bool, True if this is a resumable upload. False means upload
572 in a single request.
573 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700574 self._filename = filename
575 fd = open(self._filename, "rb")
576 if mimetype is None:
577 # No mimetype provided, make a guess.
578 mimetype, _ = mimetypes.guess_type(filename)
579 if mimetype is None:
580 # Guess failed, use octet-stream.
581 mimetype = "application/octet-stream"
582 super(MediaFileUpload, self).__init__(
583 fd, mimetype, chunksize=chunksize, resumable=resumable
584 )
John Asmuth864311d2014-04-24 15:46:08 -0400585
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700586 def __del__(self):
587 self._fd.close()
Xiaofei Wang20b67582019-07-17 11:16:53 -0700588
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700589 def to_json(self):
590 """Creating a JSON representation of an instance of MediaFileUpload.
John Asmuth864311d2014-04-24 15:46:08 -0400591
592 Returns:
593 string, a JSON representation of this instance, suitable to pass to
594 from_json().
595 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700596 return self._to_json(strip=["_fd"])
John Asmuth864311d2014-04-24 15:46:08 -0400597
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700598 @staticmethod
599 def from_json(s):
600 d = json.loads(s)
601 return MediaFileUpload(
602 d["_filename"],
603 mimetype=d["_mimetype"],
604 chunksize=d["_chunksize"],
605 resumable=d["_resumable"],
606 )
John Asmuth864311d2014-04-24 15:46:08 -0400607
608
609class MediaInMemoryUpload(MediaIoBaseUpload):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700610 """MediaUpload for a chunk of bytes.
John Asmuth864311d2014-04-24 15:46:08 -0400611
612 DEPRECATED: Use MediaIoBaseUpload with either io.TextIOBase or StringIO for
613 the stream.
614 """
615
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700616 @util.positional(2)
617 def __init__(
618 self,
619 body,
620 mimetype="application/octet-stream",
621 chunksize=DEFAULT_CHUNK_SIZE,
622 resumable=False,
623 ):
624 """Create a new MediaInMemoryUpload.
John Asmuth864311d2014-04-24 15:46:08 -0400625
626 DEPRECATED: Use MediaIoBaseUpload with either io.TextIOBase or StringIO for
627 the stream.
628
629 Args:
630 body: string, Bytes of body content.
631 mimetype: string, Mime-type of the file or default of
632 'application/octet-stream'.
633 chunksize: int, File will be uploaded in chunks of this many bytes. Only
634 used if resumable=True.
635 resumable: bool, True if this is a resumable upload. False means upload
636 in a single request.
637 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700638 fd = BytesIO(body)
639 super(MediaInMemoryUpload, self).__init__(
640 fd, mimetype, chunksize=chunksize, resumable=resumable
641 )
John Asmuth864311d2014-04-24 15:46:08 -0400642
643
644class MediaIoBaseDownload(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700645 """"Download media resources.
John Asmuth864311d2014-04-24 15:46:08 -0400646
647 Note that the Python file object is compatible with io.Base and can be used
648 with this class also.
649
650
651 Example:
652 request = farms.animals().get_media(id='cow')
653 fh = io.FileIO('cow.png', mode='wb')
654 downloader = MediaIoBaseDownload(fh, request, chunksize=1024*1024)
655
656 done = False
657 while done is False:
658 status, done = downloader.next_chunk()
659 if status:
660 print "Download %d%%." % int(status.progress() * 100)
661 print "Download Complete!"
662 """
663
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700664 @util.positional(3)
665 def __init__(self, fd, request, chunksize=DEFAULT_CHUNK_SIZE):
666 """Constructor.
John Asmuth864311d2014-04-24 15:46:08 -0400667
668 Args:
669 fd: io.Base or file object, The stream in which to write the downloaded
670 bytes.
671 request: googleapiclient.http.HttpRequest, the media request to perform in
672 chunks.
673 chunksize: int, File will be downloaded in chunks of this many bytes.
674 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700675 self._fd = fd
676 self._request = request
677 self._uri = request.uri
678 self._chunksize = chunksize
679 self._progress = 0
680 self._total_size = None
681 self._done = False
John Asmuth864311d2014-04-24 15:46:08 -0400682
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700683 # Stubs for testing.
684 self._sleep = time.sleep
685 self._rand = random.random
John Asmuth864311d2014-04-24 15:46:08 -0400686
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700687 self._headers = {}
688 for k, v in six.iteritems(request.headers):
689 # allow users to supply custom headers by setting them on the request
690 # but strip out the ones that are set by default on requests generated by
691 # API methods like Drive's files().get(fileId=...)
692 if not k.lower() in ("accept", "accept-encoding", "user-agent"):
693 self._headers[k] = v
Chris McDonough0dc81bf2018-07-19 11:19:58 -0400694
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700695 @util.positional(1)
696 def next_chunk(self, num_retries=0):
697 """Get the next chunk of the download.
John Asmuth864311d2014-04-24 15:46:08 -0400698
699 Args:
Zhihao Yuancc6d3982016-07-27 11:40:45 -0500700 num_retries: Integer, number of times to retry with randomized
John Asmuth864311d2014-04-24 15:46:08 -0400701 exponential backoff. If all retries fail, the raised HttpError
702 represents the last request. If zero (default), we attempt the
703 request only once.
704
705 Returns:
Nilayan Bhattacharya89906ac2017-10-27 13:47:23 -0700706 (status, done): (MediaDownloadProgress, boolean)
John Asmuth864311d2014-04-24 15:46:08 -0400707 The value of 'done' will be True when the media has been fully
Daniel44067782018-01-16 23:17:56 +0100708 downloaded or the total size of the media is unknown.
John Asmuth864311d2014-04-24 15:46:08 -0400709
710 Raises:
711 googleapiclient.errors.HttpError if the response was not a 2xx.
Tim Gates43fc0cf2020-04-21 08:03:25 +1000712 httplib2.HttpLib2Error if a transport error has occurred.
John Asmuth864311d2014-04-24 15:46:08 -0400713 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700714 headers = self._headers.copy()
715 headers["range"] = "bytes=%d-%d" % (
716 self._progress,
717 self._progress + self._chunksize,
718 )
719 http = self._request.http
John Asmuth864311d2014-04-24 15:46:08 -0400720
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700721 resp, content = _retry_request(
722 http,
723 num_retries,
724 "media download",
725 self._sleep,
726 self._rand,
727 self._uri,
728 "GET",
729 headers=headers,
730 )
John Asmuth864311d2014-04-24 15:46:08 -0400731
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700732 if resp.status in [200, 206]:
733 if "content-location" in resp and resp["content-location"] != self._uri:
734 self._uri = resp["content-location"]
735 self._progress += len(content)
736 self._fd.write(content)
John Asmuth864311d2014-04-24 15:46:08 -0400737
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700738 if "content-range" in resp:
739 content_range = resp["content-range"]
740 length = content_range.rsplit("/", 1)[1]
741 self._total_size = int(length)
742 elif "content-length" in resp:
743 self._total_size = int(resp["content-length"])
John Asmuth864311d2014-04-24 15:46:08 -0400744
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700745 if self._total_size is None or self._progress == self._total_size:
746 self._done = True
747 return MediaDownloadProgress(self._progress, self._total_size), self._done
748 else:
749 raise HttpError(resp, content, uri=self._uri)
John Asmuth864311d2014-04-24 15:46:08 -0400750
751
752class _StreamSlice(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700753 """Truncated stream.
John Asmuth864311d2014-04-24 15:46:08 -0400754
755 Takes a stream and presents a stream that is a slice of the original stream.
756 This is used when uploading media in chunks. In later versions of Python a
757 stream can be passed to httplib in place of the string of data to send. The
758 problem is that httplib just blindly reads to the end of the stream. This
759 wrapper presents a virtual stream that only reads to the end of the chunk.
760 """
761
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700762 def __init__(self, stream, begin, chunksize):
763 """Constructor.
John Asmuth864311d2014-04-24 15:46:08 -0400764
765 Args:
766 stream: (io.Base, file object), the stream to wrap.
767 begin: int, the seek position the chunk begins at.
768 chunksize: int, the size of the chunk.
769 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700770 self._stream = stream
771 self._begin = begin
772 self._chunksize = chunksize
773 self._stream.seek(begin)
John Asmuth864311d2014-04-24 15:46:08 -0400774
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700775 def read(self, n=-1):
776 """Read n bytes.
John Asmuth864311d2014-04-24 15:46:08 -0400777
778 Args:
779 n, int, the number of bytes to read.
780
781 Returns:
782 A string of length 'n', or less if EOF is reached.
783 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700784 # The data left available to read sits in [cur, end)
785 cur = self._stream.tell()
786 end = self._begin + self._chunksize
787 if n == -1 or cur + n > end:
788 n = end - cur
789 return self._stream.read(n)
John Asmuth864311d2014-04-24 15:46:08 -0400790
791
792class HttpRequest(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700793 """Encapsulates a single HTTP request."""
John Asmuth864311d2014-04-24 15:46:08 -0400794
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700795 @util.positional(4)
796 def __init__(
797 self,
798 http,
799 postproc,
800 uri,
801 method="GET",
802 body=None,
803 headers=None,
804 methodId=None,
805 resumable=None,
806 ):
807 """Constructor for an HttpRequest.
John Asmuth864311d2014-04-24 15:46:08 -0400808
809 Args:
810 http: httplib2.Http, the transport object to use to make a request
811 postproc: callable, called on the HTTP response and content to transform
812 it into a data object before returning, or raising an exception
813 on an error.
814 uri: string, the absolute URI to send the request to
815 method: string, the HTTP method to use
816 body: string, the request body of the HTTP request,
817 headers: dict, the HTTP request headers
818 methodId: string, a unique identifier for the API method being called.
819 resumable: MediaUpload, None if this is not a resumbale request.
820 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700821 self.uri = uri
822 self.method = method
823 self.body = body
824 self.headers = headers or {}
825 self.methodId = methodId
826 self.http = http
827 self.postproc = postproc
828 self.resumable = resumable
829 self.response_callbacks = []
830 self._in_error_state = False
John Asmuth864311d2014-04-24 15:46:08 -0400831
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700832 # The size of the non-media part of the request.
833 self.body_size = len(self.body or "")
John Asmuth864311d2014-04-24 15:46:08 -0400834
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700835 # The resumable URI to send chunks to.
836 self.resumable_uri = None
John Asmuth864311d2014-04-24 15:46:08 -0400837
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700838 # The bytes that have been uploaded.
839 self.resumable_progress = 0
John Asmuth864311d2014-04-24 15:46:08 -0400840
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700841 # Stubs for testing.
842 self._rand = random.random
843 self._sleep = time.sleep
John Asmuth864311d2014-04-24 15:46:08 -0400844
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700845 @util.positional(1)
846 def execute(self, http=None, num_retries=0):
847 """Execute the request.
John Asmuth864311d2014-04-24 15:46:08 -0400848
849 Args:
850 http: httplib2.Http, an http object to be used in place of the
851 one the HttpRequest request object was constructed with.
Zhihao Yuancc6d3982016-07-27 11:40:45 -0500852 num_retries: Integer, number of times to retry with randomized
John Asmuth864311d2014-04-24 15:46:08 -0400853 exponential backoff. If all retries fail, the raised HttpError
854 represents the last request. If zero (default), we attempt the
855 request only once.
856
857 Returns:
858 A deserialized object model of the response body as determined
859 by the postproc.
860
861 Raises:
862 googleapiclient.errors.HttpError if the response was not a 2xx.
Tim Gates43fc0cf2020-04-21 08:03:25 +1000863 httplib2.HttpLib2Error if a transport error has occurred.
John Asmuth864311d2014-04-24 15:46:08 -0400864 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700865 if http is None:
866 http = self.http
John Asmuth864311d2014-04-24 15:46:08 -0400867
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700868 if self.resumable:
869 body = None
870 while body is None:
871 _, body = self.next_chunk(http=http, num_retries=num_retries)
872 return body
John Asmuth864311d2014-04-24 15:46:08 -0400873
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700874 # Non-resumable case.
John Asmuth864311d2014-04-24 15:46:08 -0400875
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700876 if "content-length" not in self.headers:
877 self.headers["content-length"] = str(self.body_size)
878 # If the request URI is too long then turn it into a POST request.
879 # Assume that a GET request never contains a request body.
880 if len(self.uri) > MAX_URI_LENGTH and self.method == "GET":
881 self.method = "POST"
882 self.headers["x-http-method-override"] = "GET"
883 self.headers["content-type"] = "application/x-www-form-urlencoded"
884 parsed = urlparse(self.uri)
885 self.uri = urlunparse(
886 (parsed.scheme, parsed.netloc, parsed.path, parsed.params, None, None)
887 )
888 self.body = parsed.query
889 self.headers["content-length"] = str(len(self.body))
John Asmuth864311d2014-04-24 15:46:08 -0400890
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700891 # Handle retries for server-side errors.
892 resp, content = _retry_request(
893 http,
894 num_retries,
895 "request",
896 self._sleep,
897 self._rand,
898 str(self.uri),
899 method=str(self.method),
900 body=self.body,
901 headers=self.headers,
902 )
John Asmuth864311d2014-04-24 15:46:08 -0400903
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700904 for callback in self.response_callbacks:
905 callback(resp)
906 if resp.status >= 300:
907 raise HttpError(resp, content, uri=self.uri)
908 return self.postproc(resp, content)
John Asmuth864311d2014-04-24 15:46:08 -0400909
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700910 @util.positional(2)
911 def add_response_callback(self, cb):
912 """add_response_headers_callback
John Asmuth864311d2014-04-24 15:46:08 -0400913
914 Args:
915 cb: Callback to be called on receiving the response headers, of signature:
916
917 def cb(resp):
918 # Where resp is an instance of httplib2.Response
919 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700920 self.response_callbacks.append(cb)
John Asmuth864311d2014-04-24 15:46:08 -0400921
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700922 @util.positional(1)
923 def next_chunk(self, http=None, num_retries=0):
924 """Execute the next step of a resumable upload.
John Asmuth864311d2014-04-24 15:46:08 -0400925
926 Can only be used if the method being executed supports media uploads and
927 the MediaUpload object passed in was flagged as using resumable upload.
928
929 Example:
930
931 media = MediaFileUpload('cow.png', mimetype='image/png',
932 chunksize=1000, resumable=True)
933 request = farm.animals().insert(
934 id='cow',
935 name='cow.png',
936 media_body=media)
937
938 response = None
939 while response is None:
940 status, response = request.next_chunk()
941 if status:
942 print "Upload %d%% complete." % int(status.progress() * 100)
943
944
945 Args:
946 http: httplib2.Http, an http object to be used in place of the
947 one the HttpRequest request object was constructed with.
Zhihao Yuancc6d3982016-07-27 11:40:45 -0500948 num_retries: Integer, number of times to retry with randomized
John Asmuth864311d2014-04-24 15:46:08 -0400949 exponential backoff. If all retries fail, the raised HttpError
950 represents the last request. If zero (default), we attempt the
951 request only once.
952
953 Returns:
954 (status, body): (ResumableMediaStatus, object)
955 The body will be None until the resumable media is fully uploaded.
956
957 Raises:
958 googleapiclient.errors.HttpError if the response was not a 2xx.
Tim Gates43fc0cf2020-04-21 08:03:25 +1000959 httplib2.HttpLib2Error if a transport error has occurred.
John Asmuth864311d2014-04-24 15:46:08 -0400960 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700961 if http is None:
962 http = self.http
John Asmuth864311d2014-04-24 15:46:08 -0400963
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700964 if self.resumable.size() is None:
965 size = "*"
966 else:
967 size = str(self.resumable.size())
John Asmuth864311d2014-04-24 15:46:08 -0400968
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700969 if self.resumable_uri is None:
970 start_headers = copy.copy(self.headers)
971 start_headers["X-Upload-Content-Type"] = self.resumable.mimetype()
972 if size != "*":
973 start_headers["X-Upload-Content-Length"] = size
974 start_headers["content-length"] = str(self.body_size)
John Asmuth864311d2014-04-24 15:46:08 -0400975
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700976 resp, content = _retry_request(
977 http,
978 num_retries,
979 "resumable URI request",
980 self._sleep,
981 self._rand,
982 self.uri,
983 method=self.method,
984 body=self.body,
985 headers=start_headers,
986 )
John Asmuth864311d2014-04-24 15:46:08 -0400987
Bu Sun Kim66bb32c2019-10-30 10:11:58 -0700988 if resp.status == 200 and "location" in resp:
989 self.resumable_uri = resp["location"]
990 else:
991 raise ResumableUploadError(resp, content)
992 elif self._in_error_state:
993 # If we are in an error state then query the server for current state of
994 # the upload by sending an empty PUT and reading the 'range' header in
995 # the response.
996 headers = {"Content-Range": "bytes */%s" % size, "content-length": "0"}
997 resp, content = http.request(self.resumable_uri, "PUT", headers=headers)
998 status, body = self._process_response(resp, content)
999 if body:
1000 # The upload was complete.
1001 return (status, body)
John Asmuth864311d2014-04-24 15:46:08 -04001002
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001003 if self.resumable.has_stream():
1004 data = self.resumable.stream()
1005 if self.resumable.chunksize() == -1:
1006 data.seek(self.resumable_progress)
1007 chunk_end = self.resumable.size() - self.resumable_progress - 1
1008 else:
1009 # Doing chunking with a stream, so wrap a slice of the stream.
1010 data = _StreamSlice(
1011 data, self.resumable_progress, self.resumable.chunksize()
1012 )
1013 chunk_end = min(
1014 self.resumable_progress + self.resumable.chunksize() - 1,
1015 self.resumable.size() - 1,
1016 )
1017 else:
1018 data = self.resumable.getbytes(
1019 self.resumable_progress, self.resumable.chunksize()
1020 )
John Asmuth864311d2014-04-24 15:46:08 -04001021
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001022 # A short read implies that we are at EOF, so finish the upload.
1023 if len(data) < self.resumable.chunksize():
1024 size = str(self.resumable_progress + len(data))
John Asmuth864311d2014-04-24 15:46:08 -04001025
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001026 chunk_end = self.resumable_progress + len(data) - 1
John Asmuth864311d2014-04-24 15:46:08 -04001027
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001028 headers = {
1029 "Content-Range": "bytes %d-%d/%s"
1030 % (self.resumable_progress, chunk_end, size),
1031 # Must set the content-length header here because httplib can't
1032 # calculate the size when working with _StreamSlice.
1033 "Content-Length": str(chunk_end - self.resumable_progress + 1),
John Asmuth864311d2014-04-24 15:46:08 -04001034 }
1035
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001036 for retry_num in range(num_retries + 1):
1037 if retry_num > 0:
1038 self._sleep(self._rand() * 2 ** retry_num)
1039 LOGGER.warning(
1040 "Retry #%d for media upload: %s %s, following status: %d"
1041 % (retry_num, self.method, self.uri, resp.status)
1042 )
John Asmuth864311d2014-04-24 15:46:08 -04001043
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001044 try:
1045 resp, content = http.request(
1046 self.resumable_uri, method="PUT", body=data, headers=headers
1047 )
1048 except:
1049 self._in_error_state = True
1050 raise
1051 if not _should_retry_response(resp.status, content):
1052 break
John Asmuth864311d2014-04-24 15:46:08 -04001053
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001054 return self._process_response(resp, content)
John Asmuth864311d2014-04-24 15:46:08 -04001055
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001056 def _process_response(self, resp, content):
1057 """Process the response from a single chunk upload.
John Asmuth864311d2014-04-24 15:46:08 -04001058
1059 Args:
1060 resp: httplib2.Response, the response object.
1061 content: string, the content of the response.
1062
1063 Returns:
1064 (status, body): (ResumableMediaStatus, object)
1065 The body will be None until the resumable media is fully uploaded.
1066
1067 Raises:
1068 googleapiclient.errors.HttpError if the response was not a 2xx or a 308.
1069 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001070 if resp.status in [200, 201]:
1071 self._in_error_state = False
1072 return None, self.postproc(resp, content)
1073 elif resp.status == 308:
1074 self._in_error_state = False
1075 # A "308 Resume Incomplete" indicates we are not done.
1076 try:
1077 self.resumable_progress = int(resp["range"].split("-")[1]) + 1
1078 except KeyError:
1079 # If resp doesn't contain range header, resumable progress is 0
1080 self.resumable_progress = 0
1081 if "location" in resp:
1082 self.resumable_uri = resp["location"]
1083 else:
1084 self._in_error_state = True
1085 raise HttpError(resp, content, uri=self.uri)
John Asmuth864311d2014-04-24 15:46:08 -04001086
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001087 return (
1088 MediaUploadProgress(self.resumable_progress, self.resumable.size()),
1089 None,
1090 )
John Asmuth864311d2014-04-24 15:46:08 -04001091
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001092 def to_json(self):
1093 """Returns a JSON representation of the HttpRequest."""
1094 d = copy.copy(self.__dict__)
1095 if d["resumable"] is not None:
1096 d["resumable"] = self.resumable.to_json()
1097 del d["http"]
1098 del d["postproc"]
1099 del d["_sleep"]
1100 del d["_rand"]
John Asmuth864311d2014-04-24 15:46:08 -04001101
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001102 return json.dumps(d)
John Asmuth864311d2014-04-24 15:46:08 -04001103
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001104 @staticmethod
1105 def from_json(s, http, postproc):
1106 """Returns an HttpRequest populated with info from a JSON object."""
1107 d = json.loads(s)
1108 if d["resumable"] is not None:
1109 d["resumable"] = MediaUpload.new_from_json(d["resumable"])
1110 return HttpRequest(
1111 http,
1112 postproc,
1113 uri=d["uri"],
1114 method=d["method"],
1115 body=d["body"],
1116 headers=d["headers"],
1117 methodId=d["methodId"],
1118 resumable=d["resumable"],
1119 )
John Asmuth864311d2014-04-24 15:46:08 -04001120
1121
1122class BatchHttpRequest(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001123 """Batches multiple HttpRequest objects into a single HTTP request.
John Asmuth864311d2014-04-24 15:46:08 -04001124
1125 Example:
1126 from googleapiclient.http import BatchHttpRequest
1127
1128 def list_animals(request_id, response, exception):
1129 \"\"\"Do something with the animals list response.\"\"\"
1130 if exception is not None:
1131 # Do something with the exception.
1132 pass
1133 else:
1134 # Do something with the response.
1135 pass
1136
1137 def list_farmers(request_id, response, exception):
1138 \"\"\"Do something with the farmers list response.\"\"\"
1139 if exception is not None:
1140 # Do something with the exception.
1141 pass
1142 else:
1143 # Do something with the response.
1144 pass
1145
1146 service = build('farm', 'v2')
1147
1148 batch = BatchHttpRequest()
1149
1150 batch.add(service.animals().list(), list_animals)
1151 batch.add(service.farmers().list(), list_farmers)
1152 batch.execute(http=http)
1153 """
1154
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001155 @util.positional(1)
1156 def __init__(self, callback=None, batch_uri=None):
1157 """Constructor for a BatchHttpRequest.
John Asmuth864311d2014-04-24 15:46:08 -04001158
1159 Args:
1160 callback: callable, A callback to be called for each response, of the
1161 form callback(id, response, exception). The first parameter is the
1162 request id, and the second is the deserialized response object. The
1163 third is an googleapiclient.errors.HttpError exception object if an HTTP error
1164 occurred while processing the request, or None if no error occurred.
1165 batch_uri: string, URI to send batch requests to.
1166 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001167 if batch_uri is None:
1168 batch_uri = _LEGACY_BATCH_URI
Jon Wayne Parrottbae748a2018-03-28 10:21:12 -07001169
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001170 if batch_uri == _LEGACY_BATCH_URI:
1171 LOGGER.warn(
1172 "You have constructed a BatchHttpRequest using the legacy batch "
1173 "endpoint %s. This endpoint will be turned down on March 25, 2019. "
1174 "Please provide the API-specific endpoint or use "
1175 "service.new_batch_http_request(). For more details see "
1176 "https://developers.googleblog.com/2018/03/discontinuing-support-for-json-rpc-and.html"
1177 "and https://developers.google.com/api-client-library/python/guide/batch.",
1178 _LEGACY_BATCH_URI,
1179 )
1180 self._batch_uri = batch_uri
John Asmuth864311d2014-04-24 15:46:08 -04001181
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001182 # Global callback to be called for each individual response in the batch.
1183 self._callback = callback
John Asmuth864311d2014-04-24 15:46:08 -04001184
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001185 # A map from id to request.
1186 self._requests = {}
John Asmuth864311d2014-04-24 15:46:08 -04001187
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001188 # A map from id to callback.
1189 self._callbacks = {}
John Asmuth864311d2014-04-24 15:46:08 -04001190
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001191 # List of request ids, in the order in which they were added.
1192 self._order = []
John Asmuth864311d2014-04-24 15:46:08 -04001193
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001194 # The last auto generated id.
1195 self._last_auto_id = 0
John Asmuth864311d2014-04-24 15:46:08 -04001196
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001197 # Unique ID on which to base the Content-ID headers.
1198 self._base_id = None
John Asmuth864311d2014-04-24 15:46:08 -04001199
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001200 # A map from request id to (httplib2.Response, content) response pairs
1201 self._responses = {}
John Asmuth864311d2014-04-24 15:46:08 -04001202
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001203 # A map of id(Credentials) that have been refreshed.
1204 self._refreshed_credentials = {}
John Asmuth864311d2014-04-24 15:46:08 -04001205
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001206 def _refresh_and_apply_credentials(self, request, http):
1207 """Refresh the credentials and apply to the request.
John Asmuth864311d2014-04-24 15:46:08 -04001208
1209 Args:
1210 request: HttpRequest, the request.
1211 http: httplib2.Http, the global http object for the batch.
1212 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001213 # For the credentials to refresh, but only once per refresh_token
1214 # If there is no http per the request then refresh the http passed in
1215 # via execute()
1216 creds = None
1217 request_credentials = False
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001218
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001219 if request.http is not None:
1220 creds = _auth.get_credentials_from_http(request.http)
1221 request_credentials = True
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001222
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001223 if creds is None and http is not None:
1224 creds = _auth.get_credentials_from_http(http)
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001225
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001226 if creds is not None:
1227 if id(creds) not in self._refreshed_credentials:
1228 _auth.refresh_credentials(creds)
1229 self._refreshed_credentials[id(creds)] = 1
John Asmuth864311d2014-04-24 15:46:08 -04001230
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001231 # Only apply the credentials if we are using the http object passed in,
1232 # otherwise apply() will get called during _serialize_request().
1233 if request.http is None or not request_credentials:
1234 _auth.apply_credentials(creds, request.headers)
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001235
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001236 def _id_to_header(self, id_):
1237 """Convert an id to a Content-ID header value.
John Asmuth864311d2014-04-24 15:46:08 -04001238
1239 Args:
1240 id_: string, identifier of individual request.
1241
1242 Returns:
1243 A Content-ID header with the id_ encoded into it. A UUID is prepended to
1244 the value because Content-ID headers are supposed to be universally
1245 unique.
1246 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001247 if self._base_id is None:
1248 self._base_id = uuid.uuid4()
John Asmuth864311d2014-04-24 15:46:08 -04001249
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001250 # NB: we intentionally leave whitespace between base/id and '+', so RFC2822
1251 # line folding works properly on Python 3; see
1252 # https://github.com/google/google-api-python-client/issues/164
1253 return "<%s + %s>" % (self._base_id, quote(id_))
John Asmuth864311d2014-04-24 15:46:08 -04001254
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001255 def _header_to_id(self, header):
1256 """Convert a Content-ID header value to an id.
John Asmuth864311d2014-04-24 15:46:08 -04001257
1258 Presumes the Content-ID header conforms to the format that _id_to_header()
1259 returns.
1260
1261 Args:
1262 header: string, Content-ID header value.
1263
1264 Returns:
1265 The extracted id value.
1266
1267 Raises:
1268 BatchError if the header is not in the expected format.
1269 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001270 if header[0] != "<" or header[-1] != ">":
1271 raise BatchError("Invalid value for Content-ID: %s" % header)
1272 if "+" not in header:
1273 raise BatchError("Invalid value for Content-ID: %s" % header)
1274 base, id_ = header[1:-1].split(" + ", 1)
John Asmuth864311d2014-04-24 15:46:08 -04001275
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001276 return unquote(id_)
John Asmuth864311d2014-04-24 15:46:08 -04001277
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001278 def _serialize_request(self, request):
1279 """Convert an HttpRequest object into a string.
John Asmuth864311d2014-04-24 15:46:08 -04001280
1281 Args:
1282 request: HttpRequest, the request to serialize.
1283
1284 Returns:
1285 The request as a string in application/http format.
1286 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001287 # Construct status line
1288 parsed = urlparse(request.uri)
1289 request_line = urlunparse(
1290 ("", "", parsed.path, parsed.params, parsed.query, "")
John Asmuth864311d2014-04-24 15:46:08 -04001291 )
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001292 status_line = request.method + " " + request_line + " HTTP/1.1\n"
1293 major, minor = request.headers.get("content-type", "application/json").split(
1294 "/"
1295 )
1296 msg = MIMENonMultipart(major, minor)
1297 headers = request.headers.copy()
John Asmuth864311d2014-04-24 15:46:08 -04001298
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001299 if request.http is not None:
1300 credentials = _auth.get_credentials_from_http(request.http)
1301 if credentials is not None:
1302 _auth.apply_credentials(credentials, headers)
John Asmuth864311d2014-04-24 15:46:08 -04001303
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001304 # MIMENonMultipart adds its own Content-Type header.
1305 if "content-type" in headers:
1306 del headers["content-type"]
John Asmuth864311d2014-04-24 15:46:08 -04001307
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001308 for key, value in six.iteritems(headers):
1309 msg[key] = value
1310 msg["Host"] = parsed.netloc
1311 msg.set_unixfrom(None)
John Asmuth864311d2014-04-24 15:46:08 -04001312
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001313 if request.body is not None:
1314 msg.set_payload(request.body)
1315 msg["content-length"] = str(len(request.body))
John Asmuth864311d2014-04-24 15:46:08 -04001316
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001317 # Serialize the mime message.
1318 fp = StringIO()
1319 # maxheaderlen=0 means don't line wrap headers.
1320 g = Generator(fp, maxheaderlen=0)
1321 g.flatten(msg, unixfrom=False)
1322 body = fp.getvalue()
John Asmuth864311d2014-04-24 15:46:08 -04001323
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001324 return status_line + body
John Asmuth864311d2014-04-24 15:46:08 -04001325
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001326 def _deserialize_response(self, payload):
1327 """Convert string into httplib2 response and content.
John Asmuth864311d2014-04-24 15:46:08 -04001328
1329 Args:
1330 payload: string, headers and body as a string.
1331
1332 Returns:
1333 A pair (resp, content), such as would be returned from httplib2.request.
1334 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001335 # Strip off the status line
1336 status_line, payload = payload.split("\n", 1)
1337 protocol, status, reason = status_line.split(" ", 2)
John Asmuth864311d2014-04-24 15:46:08 -04001338
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001339 # Parse the rest of the response
1340 parser = FeedParser()
1341 parser.feed(payload)
1342 msg = parser.close()
1343 msg["status"] = status
John Asmuth864311d2014-04-24 15:46:08 -04001344
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001345 # Create httplib2.Response from the parsed headers.
1346 resp = httplib2.Response(msg)
1347 resp.reason = reason
1348 resp.version = int(protocol.split("/", 1)[1].replace(".", ""))
John Asmuth864311d2014-04-24 15:46:08 -04001349
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001350 content = payload.split("\r\n\r\n", 1)[1]
John Asmuth864311d2014-04-24 15:46:08 -04001351
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001352 return resp, content
John Asmuth864311d2014-04-24 15:46:08 -04001353
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001354 def _new_id(self):
1355 """Create a new id.
John Asmuth864311d2014-04-24 15:46:08 -04001356
1357 Auto incrementing number that avoids conflicts with ids already used.
1358
1359 Returns:
1360 string, a new unique id.
1361 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001362 self._last_auto_id += 1
1363 while str(self._last_auto_id) in self._requests:
1364 self._last_auto_id += 1
1365 return str(self._last_auto_id)
John Asmuth864311d2014-04-24 15:46:08 -04001366
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001367 @util.positional(2)
1368 def add(self, request, callback=None, request_id=None):
1369 """Add a new request.
John Asmuth864311d2014-04-24 15:46:08 -04001370
1371 Every callback added will be paired with a unique id, the request_id. That
1372 unique id will be passed back to the callback when the response comes back
1373 from the server. The default behavior is to have the library generate it's
1374 own unique id. If the caller passes in a request_id then they must ensure
1375 uniqueness for each request_id, and if they are not an exception is
cspeidelfbaf9d72018-05-10 12:50:12 -06001376 raised. Callers should either supply all request_ids or never supply a
John Asmuth864311d2014-04-24 15:46:08 -04001377 request id, to avoid such an error.
1378
1379 Args:
1380 request: HttpRequest, Request to add to the batch.
1381 callback: callable, A callback to be called for this response, of the
1382 form callback(id, response, exception). The first parameter is the
1383 request id, and the second is the deserialized response object. The
1384 third is an googleapiclient.errors.HttpError exception object if an HTTP error
1385 occurred while processing the request, or None if no errors occurred.
Chris McDonough3cf5e602018-07-18 16:18:38 -04001386 request_id: string, A unique id for the request. The id will be passed
1387 to the callback with the response.
John Asmuth864311d2014-04-24 15:46:08 -04001388
1389 Returns:
1390 None
1391
1392 Raises:
1393 BatchError if a media request is added to a batch.
1394 KeyError is the request_id is not unique.
1395 """
Xinan Line2dccec2018-12-07 05:28:33 +09001396
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001397 if len(self._order) >= MAX_BATCH_LIMIT:
1398 raise BatchError(
1399 "Exceeded the maximum calls(%d) in a single batch request."
1400 % MAX_BATCH_LIMIT
1401 )
1402 if request_id is None:
1403 request_id = self._new_id()
1404 if request.resumable is not None:
1405 raise BatchError("Media requests cannot be used in a batch request.")
1406 if request_id in self._requests:
1407 raise KeyError("A request with this ID already exists: %s" % request_id)
1408 self._requests[request_id] = request
1409 self._callbacks[request_id] = callback
1410 self._order.append(request_id)
John Asmuth864311d2014-04-24 15:46:08 -04001411
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001412 def _execute(self, http, order, requests):
1413 """Serialize batch request, send to server, process response.
John Asmuth864311d2014-04-24 15:46:08 -04001414
1415 Args:
1416 http: httplib2.Http, an http object to be used to make the request with.
1417 order: list, list of request ids in the order they were added to the
1418 batch.
1419 request: list, list of request objects to send.
1420
1421 Raises:
Tim Gates43fc0cf2020-04-21 08:03:25 +10001422 httplib2.HttpLib2Error if a transport error has occurred.
John Asmuth864311d2014-04-24 15:46:08 -04001423 googleapiclient.errors.BatchError if the response is the wrong format.
1424 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001425 message = MIMEMultipart("mixed")
1426 # Message should not write out it's own headers.
1427 setattr(message, "_write_headers", lambda self: None)
John Asmuth864311d2014-04-24 15:46:08 -04001428
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001429 # Add all the individual requests.
1430 for request_id in order:
1431 request = requests[request_id]
John Asmuth864311d2014-04-24 15:46:08 -04001432
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001433 msg = MIMENonMultipart("application", "http")
1434 msg["Content-Transfer-Encoding"] = "binary"
1435 msg["Content-ID"] = self._id_to_header(request_id)
John Asmuth864311d2014-04-24 15:46:08 -04001436
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001437 body = self._serialize_request(request)
1438 msg.set_payload(body)
1439 message.attach(msg)
John Asmuth864311d2014-04-24 15:46:08 -04001440
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001441 # encode the body: note that we can't use `as_string`, because
1442 # it plays games with `From ` lines.
1443 fp = StringIO()
1444 g = Generator(fp, mangle_from_=False)
1445 g.flatten(message, unixfrom=False)
1446 body = fp.getvalue()
John Asmuth864311d2014-04-24 15:46:08 -04001447
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001448 headers = {}
1449 headers["content-type"] = (
1450 "multipart/mixed; " 'boundary="%s"'
1451 ) % message.get_boundary()
John Asmuth864311d2014-04-24 15:46:08 -04001452
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001453 resp, content = http.request(
1454 self._batch_uri, method="POST", body=body, headers=headers
1455 )
John Asmuth864311d2014-04-24 15:46:08 -04001456
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001457 if resp.status >= 300:
1458 raise HttpError(resp, content, uri=self._batch_uri)
John Asmuth864311d2014-04-24 15:46:08 -04001459
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001460 # Prepend with a content-type header so FeedParser can handle it.
1461 header = "content-type: %s\r\n\r\n" % resp["content-type"]
1462 # PY3's FeedParser only accepts unicode. So we should decode content
1463 # here, and encode each payload again.
1464 if six.PY3:
1465 content = content.decode("utf-8")
1466 for_parser = header + content
John Asmuth864311d2014-04-24 15:46:08 -04001467
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001468 parser = FeedParser()
1469 parser.feed(for_parser)
1470 mime_response = parser.close()
John Asmuth864311d2014-04-24 15:46:08 -04001471
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001472 if not mime_response.is_multipart():
1473 raise BatchError(
1474 "Response not in multipart/mixed format.", resp=resp, content=content
1475 )
John Asmuth864311d2014-04-24 15:46:08 -04001476
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001477 for part in mime_response.get_payload():
1478 request_id = self._header_to_id(part["Content-ID"])
1479 response, content = self._deserialize_response(part.get_payload())
1480 # We encode content here to emulate normal http response.
1481 if isinstance(content, six.text_type):
1482 content = content.encode("utf-8")
1483 self._responses[request_id] = (response, content)
John Asmuth864311d2014-04-24 15:46:08 -04001484
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001485 @util.positional(1)
1486 def execute(self, http=None):
1487 """Execute all the requests as a single batched HTTP request.
John Asmuth864311d2014-04-24 15:46:08 -04001488
1489 Args:
1490 http: httplib2.Http, an http object to be used in place of the one the
1491 HttpRequest request object was constructed with. If one isn't supplied
1492 then use a http object from the requests in this batch.
1493
1494 Returns:
1495 None
1496
1497 Raises:
Tim Gates43fc0cf2020-04-21 08:03:25 +10001498 httplib2.HttpLib2Error if a transport error has occurred.
John Asmuth864311d2014-04-24 15:46:08 -04001499 googleapiclient.errors.BatchError if the response is the wrong format.
1500 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001501 # If we have no requests return
1502 if len(self._order) == 0:
1503 return None
John Asmuth864311d2014-04-24 15:46:08 -04001504
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001505 # If http is not supplied use the first valid one given in the requests.
1506 if http is None:
1507 for request_id in self._order:
1508 request = self._requests[request_id]
1509 if request is not None:
1510 http = request.http
1511 break
John Asmuth864311d2014-04-24 15:46:08 -04001512
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001513 if http is None:
1514 raise ValueError("Missing a valid http object.")
John Asmuth864311d2014-04-24 15:46:08 -04001515
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001516 # Special case for OAuth2Credentials-style objects which have not yet been
1517 # refreshed with an initial access_token.
1518 creds = _auth.get_credentials_from_http(http)
1519 if creds is not None:
1520 if not _auth.is_valid(creds):
1521 LOGGER.info("Attempting refresh to obtain initial access_token")
1522 _auth.refresh_credentials(creds)
Gabriel Garcia23174be2016-05-25 17:28:07 +02001523
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001524 self._execute(http, self._order, self._requests)
John Asmuth864311d2014-04-24 15:46:08 -04001525
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001526 # Loop over all the requests and check for 401s. For each 401 request the
1527 # credentials should be refreshed and then sent again in a separate batch.
1528 redo_requests = {}
1529 redo_order = []
John Asmuth864311d2014-04-24 15:46:08 -04001530
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001531 for request_id in self._order:
1532 resp, content = self._responses[request_id]
1533 if resp["status"] == "401":
1534 redo_order.append(request_id)
1535 request = self._requests[request_id]
1536 self._refresh_and_apply_credentials(request, http)
1537 redo_requests[request_id] = request
John Asmuth864311d2014-04-24 15:46:08 -04001538
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001539 if redo_requests:
1540 self._execute(http, redo_order, redo_requests)
John Asmuth864311d2014-04-24 15:46:08 -04001541
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001542 # Now process all callbacks that are erroring, and raise an exception for
1543 # ones that return a non-2xx response? Or add extra parameter to callback
1544 # that contains an HttpError?
John Asmuth864311d2014-04-24 15:46:08 -04001545
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001546 for request_id in self._order:
1547 resp, content = self._responses[request_id]
John Asmuth864311d2014-04-24 15:46:08 -04001548
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001549 request = self._requests[request_id]
1550 callback = self._callbacks[request_id]
John Asmuth864311d2014-04-24 15:46:08 -04001551
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001552 response = None
1553 exception = None
1554 try:
1555 if resp.status >= 300:
1556 raise HttpError(resp, content, uri=request.uri)
1557 response = request.postproc(resp, content)
1558 except HttpError as e:
1559 exception = e
John Asmuth864311d2014-04-24 15:46:08 -04001560
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001561 if callback is not None:
1562 callback(request_id, response, exception)
1563 if self._callback is not None:
1564 self._callback(request_id, response, exception)
John Asmuth864311d2014-04-24 15:46:08 -04001565
1566
1567class HttpRequestMock(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001568 """Mock of HttpRequest.
John Asmuth864311d2014-04-24 15:46:08 -04001569
1570 Do not construct directly, instead use RequestMockBuilder.
1571 """
1572
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001573 def __init__(self, resp, content, postproc):
1574 """Constructor for HttpRequestMock
John Asmuth864311d2014-04-24 15:46:08 -04001575
1576 Args:
1577 resp: httplib2.Response, the response to emulate coming from the request
1578 content: string, the response body
1579 postproc: callable, the post processing function usually supplied by
1580 the model class. See model.JsonModel.response() as an example.
1581 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001582 self.resp = resp
1583 self.content = content
1584 self.postproc = postproc
1585 if resp is None:
1586 self.resp = httplib2.Response({"status": 200, "reason": "OK"})
1587 if "reason" in self.resp:
1588 self.resp.reason = self.resp["reason"]
John Asmuth864311d2014-04-24 15:46:08 -04001589
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001590 def execute(self, http=None):
1591 """Execute the request.
John Asmuth864311d2014-04-24 15:46:08 -04001592
1593 Same behavior as HttpRequest.execute(), but the response is
1594 mocked and not really from an HTTP request/response.
1595 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001596 return self.postproc(self.resp, self.content)
John Asmuth864311d2014-04-24 15:46:08 -04001597
1598
1599class RequestMockBuilder(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001600 """A simple mock of HttpRequest
John Asmuth864311d2014-04-24 15:46:08 -04001601
1602 Pass in a dictionary to the constructor that maps request methodIds to
1603 tuples of (httplib2.Response, content, opt_expected_body) that should be
1604 returned when that method is called. None may also be passed in for the
1605 httplib2.Response, in which case a 200 OK response will be generated.
1606 If an opt_expected_body (str or dict) is provided, it will be compared to
1607 the body and UnexpectedBodyError will be raised on inequality.
1608
1609 Example:
1610 response = '{"data": {"id": "tag:google.c...'
1611 requestBuilder = RequestMockBuilder(
1612 {
1613 'plus.activities.get': (None, response),
1614 }
1615 )
1616 googleapiclient.discovery.build("plus", "v1", requestBuilder=requestBuilder)
1617
1618 Methods that you do not supply a response for will return a
1619 200 OK with an empty string as the response content or raise an excpetion
1620 if check_unexpected is set to True. The methodId is taken from the rpcName
1621 in the discovery document.
1622
1623 For more details see the project wiki.
1624 """
1625
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001626 def __init__(self, responses, check_unexpected=False):
1627 """Constructor for RequestMockBuilder
John Asmuth864311d2014-04-24 15:46:08 -04001628
1629 The constructed object should be a callable object
1630 that can replace the class HttpResponse.
1631
1632 responses - A dictionary that maps methodIds into tuples
1633 of (httplib2.Response, content). The methodId
1634 comes from the 'rpcName' field in the discovery
1635 document.
1636 check_unexpected - A boolean setting whether or not UnexpectedMethodError
1637 should be raised on unsupplied method.
1638 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001639 self.responses = responses
1640 self.check_unexpected = check_unexpected
John Asmuth864311d2014-04-24 15:46:08 -04001641
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001642 def __call__(
1643 self,
1644 http,
1645 postproc,
1646 uri,
1647 method="GET",
1648 body=None,
1649 headers=None,
1650 methodId=None,
1651 resumable=None,
1652 ):
1653 """Implements the callable interface that discovery.build() expects
John Asmuth864311d2014-04-24 15:46:08 -04001654 of requestBuilder, which is to build an object compatible with
1655 HttpRequest.execute(). See that method for the description of the
1656 parameters and the expected response.
1657 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001658 if methodId in self.responses:
1659 response = self.responses[methodId]
1660 resp, content = response[:2]
1661 if len(response) > 2:
1662 # Test the body against the supplied expected_body.
1663 expected_body = response[2]
1664 if bool(expected_body) != bool(body):
1665 # Not expecting a body and provided one
1666 # or expecting a body and not provided one.
1667 raise UnexpectedBodyError(expected_body, body)
1668 if isinstance(expected_body, str):
1669 expected_body = json.loads(expected_body)
1670 body = json.loads(body)
1671 if body != expected_body:
1672 raise UnexpectedBodyError(expected_body, body)
1673 return HttpRequestMock(resp, content, postproc)
1674 elif self.check_unexpected:
1675 raise UnexpectedMethodError(methodId=methodId)
1676 else:
1677 model = JsonModel(False)
1678 return HttpRequestMock(None, "{}", model.response)
John Asmuth864311d2014-04-24 15:46:08 -04001679
1680
1681class HttpMock(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001682 """Mock of httplib2.Http"""
John Asmuth864311d2014-04-24 15:46:08 -04001683
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001684 def __init__(self, filename=None, headers=None):
1685 """
John Asmuth864311d2014-04-24 15:46:08 -04001686 Args:
1687 filename: string, absolute filename to read response from
1688 headers: dict, header to return with response
1689 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001690 if headers is None:
1691 headers = {"status": "200"}
1692 if filename:
1693 f = open(filename, "rb")
1694 self.data = f.read()
1695 f.close()
1696 else:
1697 self.data = None
1698 self.response_headers = headers
1699 self.headers = None
1700 self.uri = None
1701 self.method = None
1702 self.body = None
1703 self.headers = None
John Asmuth864311d2014-04-24 15:46:08 -04001704
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001705 def request(
1706 self,
1707 uri,
1708 method="GET",
1709 body=None,
1710 headers=None,
1711 redirections=1,
1712 connection_type=None,
1713 ):
1714 self.uri = uri
1715 self.method = method
1716 self.body = body
1717 self.headers = headers
1718 return httplib2.Response(self.response_headers), self.data
John Asmuth864311d2014-04-24 15:46:08 -04001719
1720
1721class HttpMockSequence(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001722 """Mock of httplib2.Http
John Asmuth864311d2014-04-24 15:46:08 -04001723
1724 Mocks a sequence of calls to request returning different responses for each
1725 call. Create an instance initialized with the desired response headers
1726 and content and then use as if an httplib2.Http instance.
1727
1728 http = HttpMockSequence([
1729 ({'status': '401'}, ''),
1730 ({'status': '200'}, '{"access_token":"1/3w","expires_in":3600}'),
1731 ({'status': '200'}, 'echo_request_headers'),
1732 ])
1733 resp, content = http.request("http://examples.com")
1734
1735 There are special values you can pass in for content to trigger
1736 behavours that are helpful in testing.
1737
1738 'echo_request_headers' means return the request headers in the response body
1739 'echo_request_headers_as_json' means return the request headers in
1740 the response body
1741 'echo_request_body' means return the request body in the response body
1742 'echo_request_uri' means return the request uri in the response body
1743 """
1744
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001745 def __init__(self, iterable):
1746 """
John Asmuth864311d2014-04-24 15:46:08 -04001747 Args:
1748 iterable: iterable, a sequence of pairs of (headers, body)
1749 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001750 self._iterable = iterable
1751 self.follow_redirects = True
John Asmuth864311d2014-04-24 15:46:08 -04001752
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001753 def request(
1754 self,
1755 uri,
1756 method="GET",
1757 body=None,
1758 headers=None,
1759 redirections=1,
1760 connection_type=None,
1761 ):
1762 resp, content = self._iterable.pop(0)
Matt McDonaldef6420a2020-04-14 16:28:13 -04001763 content = six.ensure_binary(content)
1764
1765 if content == b"echo_request_headers":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001766 content = headers
Matt McDonaldef6420a2020-04-14 16:28:13 -04001767 elif content == b"echo_request_headers_as_json":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001768 content = json.dumps(headers)
Matt McDonaldef6420a2020-04-14 16:28:13 -04001769 elif content == b"echo_request_body":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001770 if hasattr(body, "read"):
1771 content = body.read()
1772 else:
1773 content = body
Matt McDonaldef6420a2020-04-14 16:28:13 -04001774 elif content == b"echo_request_uri":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001775 content = uri
1776 if isinstance(content, six.text_type):
1777 content = content.encode("utf-8")
1778 return httplib2.Response(resp), content
John Asmuth864311d2014-04-24 15:46:08 -04001779
1780
1781def set_user_agent(http, user_agent):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001782 """Set the user-agent on every request.
John Asmuth864311d2014-04-24 15:46:08 -04001783
1784 Args:
1785 http - An instance of httplib2.Http
1786 or something that acts like it.
1787 user_agent: string, the value for the user-agent header.
1788
1789 Returns:
1790 A modified instance of http that was passed in.
1791
1792 Example:
1793
1794 h = httplib2.Http()
1795 h = set_user_agent(h, "my-app-name/6.0")
1796
1797 Most of the time the user-agent will be set doing auth, this is for the rare
1798 cases where you are accessing an unauthenticated endpoint.
1799 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001800 request_orig = http.request
John Asmuth864311d2014-04-24 15:46:08 -04001801
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001802 # The closure that will replace 'httplib2.Http.request'.
1803 def new_request(
1804 uri,
1805 method="GET",
1806 body=None,
1807 headers=None,
1808 redirections=httplib2.DEFAULT_MAX_REDIRECTS,
1809 connection_type=None,
1810 ):
1811 """Modify the request headers to add the user-agent."""
1812 if headers is None:
1813 headers = {}
1814 if "user-agent" in headers:
1815 headers["user-agent"] = user_agent + " " + headers["user-agent"]
1816 else:
1817 headers["user-agent"] = user_agent
1818 resp, content = request_orig(
1819 uri,
1820 method=method,
1821 body=body,
1822 headers=headers,
1823 redirections=redirections,
1824 connection_type=connection_type,
1825 )
1826 return resp, content
John Asmuth864311d2014-04-24 15:46:08 -04001827
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001828 http.request = new_request
1829 return http
John Asmuth864311d2014-04-24 15:46:08 -04001830
1831
1832def tunnel_patch(http):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001833 """Tunnel PATCH requests over POST.
John Asmuth864311d2014-04-24 15:46:08 -04001834 Args:
1835 http - An instance of httplib2.Http
1836 or something that acts like it.
1837
1838 Returns:
1839 A modified instance of http that was passed in.
1840
1841 Example:
1842
1843 h = httplib2.Http()
1844 h = tunnel_patch(h, "my-app-name/6.0")
1845
1846 Useful if you are running on a platform that doesn't support PATCH.
1847 Apply this last if you are using OAuth 1.0, as changing the method
1848 will result in a different signature.
1849 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001850 request_orig = http.request
John Asmuth864311d2014-04-24 15:46:08 -04001851
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001852 # The closure that will replace 'httplib2.Http.request'.
1853 def new_request(
1854 uri,
1855 method="GET",
1856 body=None,
1857 headers=None,
1858 redirections=httplib2.DEFAULT_MAX_REDIRECTS,
1859 connection_type=None,
1860 ):
1861 """Modify the request headers to add the user-agent."""
1862 if headers is None:
1863 headers = {}
1864 if method == "PATCH":
1865 if "oauth_token" in headers.get("authorization", ""):
1866 LOGGER.warning(
1867 "OAuth 1.0 request made with Credentials after tunnel_patch."
1868 )
1869 headers["x-http-method-override"] = "PATCH"
1870 method = "POST"
1871 resp, content = request_orig(
1872 uri,
1873 method=method,
1874 body=body,
1875 headers=headers,
1876 redirections=redirections,
1877 connection_type=connection_type,
1878 )
1879 return resp, content
John Asmuth864311d2014-04-24 15:46:08 -04001880
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001881 http.request = new_request
1882 return http
Igor Maravić22435292017-01-19 22:28:22 +01001883
1884
1885def build_http():
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001886 """Builds httplib2.Http object
Igor Maravić22435292017-01-19 22:28:22 +01001887
1888 Returns:
1889 A httplib2.Http object, which is used to make http requests, and which has timeout set by default.
1890 To override default timeout call
1891
1892 socket.setdefaulttimeout(timeout_in_sec)
1893
1894 before interacting with this method.
1895 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001896 if socket.getdefaulttimeout() is not None:
1897 http_timeout = socket.getdefaulttimeout()
1898 else:
1899 http_timeout = DEFAULT_HTTP_TIMEOUT_SEC
Bu Sun Kimb3b773f2020-03-11 12:58:16 -07001900 http = httplib2.Http(timeout=http_timeout)
1901 # 308's are used by several Google APIs (Drive, YouTube)
1902 # for Resumable Uploads rather than Permanent Redirects.
1903 # This asks httplib2 to exclude 308s from the status codes
1904 # it treats as redirects
Bu Sun Kima480d532020-03-13 12:52:22 -07001905 try:
1906 http.redirect_codes = http.redirect_codes - {308}
1907 except AttributeError:
1908 # Apache Beam tests depend on this library and cannot
1909 # currently upgrade their httplib2 version
1910 # http.redirect_codes does not exist in previous versions
1911 # of httplib2, so pass
1912 pass
Bu Sun Kimb3b773f2020-03-11 12:58:16 -07001913
1914 return http