blob: d9c3d2ae19af4cbd63e6853a3d3e6b4173a4b9dc [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
Dmitry Frenkelf3348f92020-07-15 13:05:58 -07001121 @staticmethod
1122 def null_postproc(resp, contents):
1123 return resp, contents
1124
John Asmuth864311d2014-04-24 15:46:08 -04001125
1126class BatchHttpRequest(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001127 """Batches multiple HttpRequest objects into a single HTTP request.
John Asmuth864311d2014-04-24 15:46:08 -04001128
1129 Example:
1130 from googleapiclient.http import BatchHttpRequest
1131
1132 def list_animals(request_id, response, exception):
1133 \"\"\"Do something with the animals list response.\"\"\"
1134 if exception is not None:
1135 # Do something with the exception.
1136 pass
1137 else:
1138 # Do something with the response.
1139 pass
1140
1141 def list_farmers(request_id, response, exception):
1142 \"\"\"Do something with the farmers list response.\"\"\"
1143 if exception is not None:
1144 # Do something with the exception.
1145 pass
1146 else:
1147 # Do something with the response.
1148 pass
1149
1150 service = build('farm', 'v2')
1151
1152 batch = BatchHttpRequest()
1153
1154 batch.add(service.animals().list(), list_animals)
1155 batch.add(service.farmers().list(), list_farmers)
1156 batch.execute(http=http)
1157 """
1158
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001159 @util.positional(1)
1160 def __init__(self, callback=None, batch_uri=None):
1161 """Constructor for a BatchHttpRequest.
John Asmuth864311d2014-04-24 15:46:08 -04001162
1163 Args:
1164 callback: callable, A callback to be called for each response, of the
1165 form callback(id, response, exception). The first parameter is the
1166 request id, and the second is the deserialized response object. The
1167 third is an googleapiclient.errors.HttpError exception object if an HTTP error
1168 occurred while processing the request, or None if no error occurred.
1169 batch_uri: string, URI to send batch requests to.
1170 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001171 if batch_uri is None:
1172 batch_uri = _LEGACY_BATCH_URI
Jon Wayne Parrottbae748a2018-03-28 10:21:12 -07001173
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001174 if batch_uri == _LEGACY_BATCH_URI:
Dmitry Frenkelf3348f92020-07-15 13:05:58 -07001175 LOGGER.warning(
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001176 "You have constructed a BatchHttpRequest using the legacy batch "
Brad Vogel6ddadd72020-05-15 10:02:04 -07001177 "endpoint %s. This endpoint will be turned down on August 12, 2020. "
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001178 "Please provide the API-specific endpoint or use "
1179 "service.new_batch_http_request(). For more details see "
1180 "https://developers.googleblog.com/2018/03/discontinuing-support-for-json-rpc-and.html"
1181 "and https://developers.google.com/api-client-library/python/guide/batch.",
1182 _LEGACY_BATCH_URI,
1183 )
1184 self._batch_uri = batch_uri
John Asmuth864311d2014-04-24 15:46:08 -04001185
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001186 # Global callback to be called for each individual response in the batch.
1187 self._callback = callback
John Asmuth864311d2014-04-24 15:46:08 -04001188
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001189 # A map from id to request.
1190 self._requests = {}
John Asmuth864311d2014-04-24 15:46:08 -04001191
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001192 # A map from id to callback.
1193 self._callbacks = {}
John Asmuth864311d2014-04-24 15:46:08 -04001194
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001195 # List of request ids, in the order in which they were added.
1196 self._order = []
John Asmuth864311d2014-04-24 15:46:08 -04001197
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001198 # The last auto generated id.
1199 self._last_auto_id = 0
John Asmuth864311d2014-04-24 15:46:08 -04001200
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001201 # Unique ID on which to base the Content-ID headers.
1202 self._base_id = None
John Asmuth864311d2014-04-24 15:46:08 -04001203
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001204 # A map from request id to (httplib2.Response, content) response pairs
1205 self._responses = {}
John Asmuth864311d2014-04-24 15:46:08 -04001206
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001207 # A map of id(Credentials) that have been refreshed.
1208 self._refreshed_credentials = {}
John Asmuth864311d2014-04-24 15:46:08 -04001209
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001210 def _refresh_and_apply_credentials(self, request, http):
1211 """Refresh the credentials and apply to the request.
John Asmuth864311d2014-04-24 15:46:08 -04001212
1213 Args:
1214 request: HttpRequest, the request.
1215 http: httplib2.Http, the global http object for the batch.
1216 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001217 # For the credentials to refresh, but only once per refresh_token
1218 # If there is no http per the request then refresh the http passed in
1219 # via execute()
1220 creds = None
1221 request_credentials = False
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001222
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001223 if request.http is not None:
1224 creds = _auth.get_credentials_from_http(request.http)
1225 request_credentials = True
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001226
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001227 if creds is None and http is not None:
1228 creds = _auth.get_credentials_from_http(http)
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001229
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001230 if creds is not None:
1231 if id(creds) not in self._refreshed_credentials:
1232 _auth.refresh_credentials(creds)
1233 self._refreshed_credentials[id(creds)] = 1
John Asmuth864311d2014-04-24 15:46:08 -04001234
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001235 # Only apply the credentials if we are using the http object passed in,
1236 # otherwise apply() will get called during _serialize_request().
1237 if request.http is None or not request_credentials:
1238 _auth.apply_credentials(creds, request.headers)
Jon Wayne Parrottd3a5cf42017-06-19 17:55:04 -07001239
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001240 def _id_to_header(self, id_):
1241 """Convert an id to a Content-ID header value.
John Asmuth864311d2014-04-24 15:46:08 -04001242
1243 Args:
1244 id_: string, identifier of individual request.
1245
1246 Returns:
1247 A Content-ID header with the id_ encoded into it. A UUID is prepended to
1248 the value because Content-ID headers are supposed to be universally
1249 unique.
1250 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001251 if self._base_id is None:
1252 self._base_id = uuid.uuid4()
John Asmuth864311d2014-04-24 15:46:08 -04001253
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001254 # NB: we intentionally leave whitespace between base/id and '+', so RFC2822
1255 # line folding works properly on Python 3; see
Marie J.I48f503f2020-05-15 13:32:11 -04001256 # https://github.com/googleapis/google-api-python-client/issues/164
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001257 return "<%s + %s>" % (self._base_id, quote(id_))
John Asmuth864311d2014-04-24 15:46:08 -04001258
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001259 def _header_to_id(self, header):
1260 """Convert a Content-ID header value to an id.
John Asmuth864311d2014-04-24 15:46:08 -04001261
1262 Presumes the Content-ID header conforms to the format that _id_to_header()
1263 returns.
1264
1265 Args:
1266 header: string, Content-ID header value.
1267
1268 Returns:
1269 The extracted id value.
1270
1271 Raises:
1272 BatchError if the header is not in the expected format.
1273 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001274 if header[0] != "<" or header[-1] != ">":
1275 raise BatchError("Invalid value for Content-ID: %s" % header)
1276 if "+" not in header:
1277 raise BatchError("Invalid value for Content-ID: %s" % header)
1278 base, id_ = header[1:-1].split(" + ", 1)
John Asmuth864311d2014-04-24 15:46:08 -04001279
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001280 return unquote(id_)
John Asmuth864311d2014-04-24 15:46:08 -04001281
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001282 def _serialize_request(self, request):
1283 """Convert an HttpRequest object into a string.
John Asmuth864311d2014-04-24 15:46:08 -04001284
1285 Args:
1286 request: HttpRequest, the request to serialize.
1287
1288 Returns:
1289 The request as a string in application/http format.
1290 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001291 # Construct status line
1292 parsed = urlparse(request.uri)
1293 request_line = urlunparse(
1294 ("", "", parsed.path, parsed.params, parsed.query, "")
John Asmuth864311d2014-04-24 15:46:08 -04001295 )
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001296 status_line = request.method + " " + request_line + " HTTP/1.1\n"
1297 major, minor = request.headers.get("content-type", "application/json").split(
1298 "/"
1299 )
1300 msg = MIMENonMultipart(major, minor)
1301 headers = request.headers.copy()
John Asmuth864311d2014-04-24 15:46:08 -04001302
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001303 if request.http is not None:
1304 credentials = _auth.get_credentials_from_http(request.http)
1305 if credentials is not None:
1306 _auth.apply_credentials(credentials, headers)
John Asmuth864311d2014-04-24 15:46:08 -04001307
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001308 # MIMENonMultipart adds its own Content-Type header.
1309 if "content-type" in headers:
1310 del headers["content-type"]
John Asmuth864311d2014-04-24 15:46:08 -04001311
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001312 for key, value in six.iteritems(headers):
1313 msg[key] = value
1314 msg["Host"] = parsed.netloc
1315 msg.set_unixfrom(None)
John Asmuth864311d2014-04-24 15:46:08 -04001316
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001317 if request.body is not None:
1318 msg.set_payload(request.body)
1319 msg["content-length"] = str(len(request.body))
John Asmuth864311d2014-04-24 15:46:08 -04001320
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001321 # Serialize the mime message.
1322 fp = StringIO()
1323 # maxheaderlen=0 means don't line wrap headers.
1324 g = Generator(fp, maxheaderlen=0)
1325 g.flatten(msg, unixfrom=False)
1326 body = fp.getvalue()
John Asmuth864311d2014-04-24 15:46:08 -04001327
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001328 return status_line + body
John Asmuth864311d2014-04-24 15:46:08 -04001329
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001330 def _deserialize_response(self, payload):
1331 """Convert string into httplib2 response and content.
John Asmuth864311d2014-04-24 15:46:08 -04001332
1333 Args:
1334 payload: string, headers and body as a string.
1335
1336 Returns:
1337 A pair (resp, content), such as would be returned from httplib2.request.
1338 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001339 # Strip off the status line
1340 status_line, payload = payload.split("\n", 1)
1341 protocol, status, reason = status_line.split(" ", 2)
John Asmuth864311d2014-04-24 15:46:08 -04001342
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001343 # Parse the rest of the response
1344 parser = FeedParser()
1345 parser.feed(payload)
1346 msg = parser.close()
1347 msg["status"] = status
John Asmuth864311d2014-04-24 15:46:08 -04001348
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001349 # Create httplib2.Response from the parsed headers.
1350 resp = httplib2.Response(msg)
1351 resp.reason = reason
1352 resp.version = int(protocol.split("/", 1)[1].replace(".", ""))
John Asmuth864311d2014-04-24 15:46:08 -04001353
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001354 content = payload.split("\r\n\r\n", 1)[1]
John Asmuth864311d2014-04-24 15:46:08 -04001355
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001356 return resp, content
John Asmuth864311d2014-04-24 15:46:08 -04001357
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001358 def _new_id(self):
1359 """Create a new id.
John Asmuth864311d2014-04-24 15:46:08 -04001360
1361 Auto incrementing number that avoids conflicts with ids already used.
1362
1363 Returns:
1364 string, a new unique id.
1365 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001366 self._last_auto_id += 1
1367 while str(self._last_auto_id) in self._requests:
1368 self._last_auto_id += 1
1369 return str(self._last_auto_id)
John Asmuth864311d2014-04-24 15:46:08 -04001370
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001371 @util.positional(2)
1372 def add(self, request, callback=None, request_id=None):
1373 """Add a new request.
John Asmuth864311d2014-04-24 15:46:08 -04001374
1375 Every callback added will be paired with a unique id, the request_id. That
1376 unique id will be passed back to the callback when the response comes back
1377 from the server. The default behavior is to have the library generate it's
1378 own unique id. If the caller passes in a request_id then they must ensure
1379 uniqueness for each request_id, and if they are not an exception is
cspeidelfbaf9d72018-05-10 12:50:12 -06001380 raised. Callers should either supply all request_ids or never supply a
John Asmuth864311d2014-04-24 15:46:08 -04001381 request id, to avoid such an error.
1382
1383 Args:
1384 request: HttpRequest, Request to add to the batch.
1385 callback: callable, A callback to be called for this response, of the
1386 form callback(id, response, exception). The first parameter is the
1387 request id, and the second is the deserialized response object. The
1388 third is an googleapiclient.errors.HttpError exception object if an HTTP error
1389 occurred while processing the request, or None if no errors occurred.
Chris McDonough3cf5e602018-07-18 16:18:38 -04001390 request_id: string, A unique id for the request. The id will be passed
1391 to the callback with the response.
John Asmuth864311d2014-04-24 15:46:08 -04001392
1393 Returns:
1394 None
1395
1396 Raises:
1397 BatchError if a media request is added to a batch.
1398 KeyError is the request_id is not unique.
1399 """
Xinan Line2dccec2018-12-07 05:28:33 +09001400
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001401 if len(self._order) >= MAX_BATCH_LIMIT:
1402 raise BatchError(
1403 "Exceeded the maximum calls(%d) in a single batch request."
1404 % MAX_BATCH_LIMIT
1405 )
1406 if request_id is None:
1407 request_id = self._new_id()
1408 if request.resumable is not None:
1409 raise BatchError("Media requests cannot be used in a batch request.")
1410 if request_id in self._requests:
1411 raise KeyError("A request with this ID already exists: %s" % request_id)
1412 self._requests[request_id] = request
1413 self._callbacks[request_id] = callback
1414 self._order.append(request_id)
John Asmuth864311d2014-04-24 15:46:08 -04001415
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001416 def _execute(self, http, order, requests):
1417 """Serialize batch request, send to server, process response.
John Asmuth864311d2014-04-24 15:46:08 -04001418
1419 Args:
1420 http: httplib2.Http, an http object to be used to make the request with.
1421 order: list, list of request ids in the order they were added to the
1422 batch.
Dmitry Frenkelf3348f92020-07-15 13:05:58 -07001423 requests: list, list of request objects to send.
John Asmuth864311d2014-04-24 15:46:08 -04001424
1425 Raises:
Tim Gates43fc0cf2020-04-21 08:03:25 +10001426 httplib2.HttpLib2Error if a transport error has occurred.
John Asmuth864311d2014-04-24 15:46:08 -04001427 googleapiclient.errors.BatchError if the response is the wrong format.
1428 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001429 message = MIMEMultipart("mixed")
1430 # Message should not write out it's own headers.
1431 setattr(message, "_write_headers", lambda self: None)
John Asmuth864311d2014-04-24 15:46:08 -04001432
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001433 # Add all the individual requests.
1434 for request_id in order:
1435 request = requests[request_id]
John Asmuth864311d2014-04-24 15:46:08 -04001436
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001437 msg = MIMENonMultipart("application", "http")
1438 msg["Content-Transfer-Encoding"] = "binary"
1439 msg["Content-ID"] = self._id_to_header(request_id)
John Asmuth864311d2014-04-24 15:46:08 -04001440
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001441 body = self._serialize_request(request)
1442 msg.set_payload(body)
1443 message.attach(msg)
John Asmuth864311d2014-04-24 15:46:08 -04001444
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001445 # encode the body: note that we can't use `as_string`, because
1446 # it plays games with `From ` lines.
1447 fp = StringIO()
1448 g = Generator(fp, mangle_from_=False)
1449 g.flatten(message, unixfrom=False)
1450 body = fp.getvalue()
John Asmuth864311d2014-04-24 15:46:08 -04001451
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001452 headers = {}
1453 headers["content-type"] = (
1454 "multipart/mixed; " 'boundary="%s"'
1455 ) % message.get_boundary()
John Asmuth864311d2014-04-24 15:46:08 -04001456
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001457 resp, content = http.request(
1458 self._batch_uri, method="POST", body=body, headers=headers
1459 )
John Asmuth864311d2014-04-24 15:46:08 -04001460
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001461 if resp.status >= 300:
1462 raise HttpError(resp, content, uri=self._batch_uri)
John Asmuth864311d2014-04-24 15:46:08 -04001463
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001464 # Prepend with a content-type header so FeedParser can handle it.
1465 header = "content-type: %s\r\n\r\n" % resp["content-type"]
1466 # PY3's FeedParser only accepts unicode. So we should decode content
1467 # here, and encode each payload again.
1468 if six.PY3:
1469 content = content.decode("utf-8")
1470 for_parser = header + content
John Asmuth864311d2014-04-24 15:46:08 -04001471
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001472 parser = FeedParser()
1473 parser.feed(for_parser)
1474 mime_response = parser.close()
John Asmuth864311d2014-04-24 15:46:08 -04001475
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001476 if not mime_response.is_multipart():
1477 raise BatchError(
1478 "Response not in multipart/mixed format.", resp=resp, content=content
1479 )
John Asmuth864311d2014-04-24 15:46:08 -04001480
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001481 for part in mime_response.get_payload():
1482 request_id = self._header_to_id(part["Content-ID"])
1483 response, content = self._deserialize_response(part.get_payload())
1484 # We encode content here to emulate normal http response.
1485 if isinstance(content, six.text_type):
1486 content = content.encode("utf-8")
1487 self._responses[request_id] = (response, content)
John Asmuth864311d2014-04-24 15:46:08 -04001488
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001489 @util.positional(1)
1490 def execute(self, http=None):
1491 """Execute all the requests as a single batched HTTP request.
John Asmuth864311d2014-04-24 15:46:08 -04001492
1493 Args:
1494 http: httplib2.Http, an http object to be used in place of the one the
1495 HttpRequest request object was constructed with. If one isn't supplied
1496 then use a http object from the requests in this batch.
1497
1498 Returns:
1499 None
1500
1501 Raises:
Tim Gates43fc0cf2020-04-21 08:03:25 +10001502 httplib2.HttpLib2Error if a transport error has occurred.
John Asmuth864311d2014-04-24 15:46:08 -04001503 googleapiclient.errors.BatchError if the response is the wrong format.
1504 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001505 # If we have no requests return
1506 if len(self._order) == 0:
1507 return None
John Asmuth864311d2014-04-24 15:46:08 -04001508
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001509 # If http is not supplied use the first valid one given in the requests.
1510 if http is None:
1511 for request_id in self._order:
1512 request = self._requests[request_id]
1513 if request is not None:
1514 http = request.http
1515 break
John Asmuth864311d2014-04-24 15:46:08 -04001516
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001517 if http is None:
1518 raise ValueError("Missing a valid http object.")
John Asmuth864311d2014-04-24 15:46:08 -04001519
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001520 # Special case for OAuth2Credentials-style objects which have not yet been
1521 # refreshed with an initial access_token.
1522 creds = _auth.get_credentials_from_http(http)
1523 if creds is not None:
1524 if not _auth.is_valid(creds):
1525 LOGGER.info("Attempting refresh to obtain initial access_token")
1526 _auth.refresh_credentials(creds)
Gabriel Garcia23174be2016-05-25 17:28:07 +02001527
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001528 self._execute(http, self._order, self._requests)
John Asmuth864311d2014-04-24 15:46:08 -04001529
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001530 # Loop over all the requests and check for 401s. For each 401 request the
1531 # credentials should be refreshed and then sent again in a separate batch.
1532 redo_requests = {}
1533 redo_order = []
John Asmuth864311d2014-04-24 15:46:08 -04001534
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001535 for request_id in self._order:
1536 resp, content = self._responses[request_id]
1537 if resp["status"] == "401":
1538 redo_order.append(request_id)
1539 request = self._requests[request_id]
1540 self._refresh_and_apply_credentials(request, http)
1541 redo_requests[request_id] = request
John Asmuth864311d2014-04-24 15:46:08 -04001542
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001543 if redo_requests:
1544 self._execute(http, redo_order, redo_requests)
John Asmuth864311d2014-04-24 15:46:08 -04001545
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001546 # Now process all callbacks that are erroring, and raise an exception for
1547 # ones that return a non-2xx response? Or add extra parameter to callback
1548 # that contains an HttpError?
John Asmuth864311d2014-04-24 15:46:08 -04001549
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001550 for request_id in self._order:
1551 resp, content = self._responses[request_id]
John Asmuth864311d2014-04-24 15:46:08 -04001552
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001553 request = self._requests[request_id]
1554 callback = self._callbacks[request_id]
John Asmuth864311d2014-04-24 15:46:08 -04001555
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001556 response = None
1557 exception = None
1558 try:
1559 if resp.status >= 300:
1560 raise HttpError(resp, content, uri=request.uri)
1561 response = request.postproc(resp, content)
1562 except HttpError as e:
1563 exception = e
John Asmuth864311d2014-04-24 15:46:08 -04001564
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001565 if callback is not None:
1566 callback(request_id, response, exception)
1567 if self._callback is not None:
1568 self._callback(request_id, response, exception)
John Asmuth864311d2014-04-24 15:46:08 -04001569
1570
1571class HttpRequestMock(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001572 """Mock of HttpRequest.
John Asmuth864311d2014-04-24 15:46:08 -04001573
1574 Do not construct directly, instead use RequestMockBuilder.
1575 """
1576
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001577 def __init__(self, resp, content, postproc):
1578 """Constructor for HttpRequestMock
John Asmuth864311d2014-04-24 15:46:08 -04001579
1580 Args:
1581 resp: httplib2.Response, the response to emulate coming from the request
1582 content: string, the response body
1583 postproc: callable, the post processing function usually supplied by
1584 the model class. See model.JsonModel.response() as an example.
1585 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001586 self.resp = resp
1587 self.content = content
1588 self.postproc = postproc
1589 if resp is None:
1590 self.resp = httplib2.Response({"status": 200, "reason": "OK"})
1591 if "reason" in self.resp:
1592 self.resp.reason = self.resp["reason"]
John Asmuth864311d2014-04-24 15:46:08 -04001593
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001594 def execute(self, http=None):
1595 """Execute the request.
John Asmuth864311d2014-04-24 15:46:08 -04001596
1597 Same behavior as HttpRequest.execute(), but the response is
1598 mocked and not really from an HTTP request/response.
1599 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001600 return self.postproc(self.resp, self.content)
John Asmuth864311d2014-04-24 15:46:08 -04001601
1602
1603class RequestMockBuilder(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001604 """A simple mock of HttpRequest
John Asmuth864311d2014-04-24 15:46:08 -04001605
1606 Pass in a dictionary to the constructor that maps request methodIds to
1607 tuples of (httplib2.Response, content, opt_expected_body) that should be
1608 returned when that method is called. None may also be passed in for the
1609 httplib2.Response, in which case a 200 OK response will be generated.
1610 If an opt_expected_body (str or dict) is provided, it will be compared to
1611 the body and UnexpectedBodyError will be raised on inequality.
1612
1613 Example:
1614 response = '{"data": {"id": "tag:google.c...'
1615 requestBuilder = RequestMockBuilder(
1616 {
1617 'plus.activities.get': (None, response),
1618 }
1619 )
1620 googleapiclient.discovery.build("plus", "v1", requestBuilder=requestBuilder)
1621
1622 Methods that you do not supply a response for will return a
1623 200 OK with an empty string as the response content or raise an excpetion
1624 if check_unexpected is set to True. The methodId is taken from the rpcName
1625 in the discovery document.
1626
1627 For more details see the project wiki.
1628 """
1629
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001630 def __init__(self, responses, check_unexpected=False):
1631 """Constructor for RequestMockBuilder
John Asmuth864311d2014-04-24 15:46:08 -04001632
1633 The constructed object should be a callable object
1634 that can replace the class HttpResponse.
1635
1636 responses - A dictionary that maps methodIds into tuples
1637 of (httplib2.Response, content). The methodId
1638 comes from the 'rpcName' field in the discovery
1639 document.
1640 check_unexpected - A boolean setting whether or not UnexpectedMethodError
1641 should be raised on unsupplied method.
1642 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001643 self.responses = responses
1644 self.check_unexpected = check_unexpected
John Asmuth864311d2014-04-24 15:46:08 -04001645
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001646 def __call__(
1647 self,
1648 http,
1649 postproc,
1650 uri,
1651 method="GET",
1652 body=None,
1653 headers=None,
1654 methodId=None,
1655 resumable=None,
1656 ):
1657 """Implements the callable interface that discovery.build() expects
John Asmuth864311d2014-04-24 15:46:08 -04001658 of requestBuilder, which is to build an object compatible with
1659 HttpRequest.execute(). See that method for the description of the
1660 parameters and the expected response.
1661 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001662 if methodId in self.responses:
1663 response = self.responses[methodId]
1664 resp, content = response[:2]
1665 if len(response) > 2:
1666 # Test the body against the supplied expected_body.
1667 expected_body = response[2]
1668 if bool(expected_body) != bool(body):
1669 # Not expecting a body and provided one
1670 # or expecting a body and not provided one.
1671 raise UnexpectedBodyError(expected_body, body)
1672 if isinstance(expected_body, str):
1673 expected_body = json.loads(expected_body)
1674 body = json.loads(body)
1675 if body != expected_body:
1676 raise UnexpectedBodyError(expected_body, body)
1677 return HttpRequestMock(resp, content, postproc)
1678 elif self.check_unexpected:
1679 raise UnexpectedMethodError(methodId=methodId)
1680 else:
1681 model = JsonModel(False)
1682 return HttpRequestMock(None, "{}", model.response)
John Asmuth864311d2014-04-24 15:46:08 -04001683
1684
1685class HttpMock(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001686 """Mock of httplib2.Http"""
John Asmuth864311d2014-04-24 15:46:08 -04001687
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001688 def __init__(self, filename=None, headers=None):
1689 """
John Asmuth864311d2014-04-24 15:46:08 -04001690 Args:
1691 filename: string, absolute filename to read response from
1692 headers: dict, header to return with response
1693 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001694 if headers is None:
1695 headers = {"status": "200"}
1696 if filename:
Dmitry Frenkelf3348f92020-07-15 13:05:58 -07001697 with open(filename, "rb") as f:
1698 self.data = f.read()
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001699 else:
1700 self.data = None
1701 self.response_headers = headers
1702 self.headers = None
1703 self.uri = None
1704 self.method = None
1705 self.body = None
1706 self.headers = None
John Asmuth864311d2014-04-24 15:46:08 -04001707
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001708 def request(
1709 self,
1710 uri,
1711 method="GET",
1712 body=None,
1713 headers=None,
1714 redirections=1,
1715 connection_type=None,
1716 ):
1717 self.uri = uri
1718 self.method = method
1719 self.body = body
1720 self.headers = headers
1721 return httplib2.Response(self.response_headers), self.data
John Asmuth864311d2014-04-24 15:46:08 -04001722
1723
1724class HttpMockSequence(object):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001725 """Mock of httplib2.Http
John Asmuth864311d2014-04-24 15:46:08 -04001726
1727 Mocks a sequence of calls to request returning different responses for each
1728 call. Create an instance initialized with the desired response headers
1729 and content and then use as if an httplib2.Http instance.
1730
1731 http = HttpMockSequence([
1732 ({'status': '401'}, ''),
1733 ({'status': '200'}, '{"access_token":"1/3w","expires_in":3600}'),
1734 ({'status': '200'}, 'echo_request_headers'),
1735 ])
1736 resp, content = http.request("http://examples.com")
1737
1738 There are special values you can pass in for content to trigger
1739 behavours that are helpful in testing.
1740
1741 'echo_request_headers' means return the request headers in the response body
1742 'echo_request_headers_as_json' means return the request headers in
1743 the response body
1744 'echo_request_body' means return the request body in the response body
1745 'echo_request_uri' means return the request uri in the response body
1746 """
1747
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001748 def __init__(self, iterable):
1749 """
John Asmuth864311d2014-04-24 15:46:08 -04001750 Args:
1751 iterable: iterable, a sequence of pairs of (headers, body)
1752 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001753 self._iterable = iterable
1754 self.follow_redirects = True
Dmitry Frenkelf3348f92020-07-15 13:05:58 -07001755 self.request_sequence = list()
John Asmuth864311d2014-04-24 15:46:08 -04001756
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001757 def request(
1758 self,
1759 uri,
1760 method="GET",
1761 body=None,
1762 headers=None,
1763 redirections=1,
1764 connection_type=None,
1765 ):
Dmitry Frenkelf3348f92020-07-15 13:05:58 -07001766 # Remember the request so after the fact this mock can be examined
1767 self.request_sequence.append((uri, method, body, headers))
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001768 resp, content = self._iterable.pop(0)
Matt McDonaldef6420a2020-04-14 16:28:13 -04001769 content = six.ensure_binary(content)
1770
1771 if content == b"echo_request_headers":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001772 content = headers
Matt McDonaldef6420a2020-04-14 16:28:13 -04001773 elif content == b"echo_request_headers_as_json":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001774 content = json.dumps(headers)
Matt McDonaldef6420a2020-04-14 16:28:13 -04001775 elif content == b"echo_request_body":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001776 if hasattr(body, "read"):
1777 content = body.read()
1778 else:
1779 content = body
Matt McDonaldef6420a2020-04-14 16:28:13 -04001780 elif content == b"echo_request_uri":
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001781 content = uri
1782 if isinstance(content, six.text_type):
1783 content = content.encode("utf-8")
1784 return httplib2.Response(resp), content
John Asmuth864311d2014-04-24 15:46:08 -04001785
1786
1787def set_user_agent(http, user_agent):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001788 """Set the user-agent on every request.
John Asmuth864311d2014-04-24 15:46:08 -04001789
1790 Args:
1791 http - An instance of httplib2.Http
1792 or something that acts like it.
1793 user_agent: string, the value for the user-agent header.
1794
1795 Returns:
1796 A modified instance of http that was passed in.
1797
1798 Example:
1799
1800 h = httplib2.Http()
1801 h = set_user_agent(h, "my-app-name/6.0")
1802
1803 Most of the time the user-agent will be set doing auth, this is for the rare
1804 cases where you are accessing an unauthenticated endpoint.
1805 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001806 request_orig = http.request
John Asmuth864311d2014-04-24 15:46:08 -04001807
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001808 # The closure that will replace 'httplib2.Http.request'.
1809 def new_request(
1810 uri,
1811 method="GET",
1812 body=None,
1813 headers=None,
1814 redirections=httplib2.DEFAULT_MAX_REDIRECTS,
1815 connection_type=None,
1816 ):
1817 """Modify the request headers to add the user-agent."""
1818 if headers is None:
1819 headers = {}
1820 if "user-agent" in headers:
1821 headers["user-agent"] = user_agent + " " + headers["user-agent"]
1822 else:
1823 headers["user-agent"] = user_agent
1824 resp, content = request_orig(
1825 uri,
1826 method=method,
1827 body=body,
1828 headers=headers,
1829 redirections=redirections,
1830 connection_type=connection_type,
1831 )
1832 return resp, content
John Asmuth864311d2014-04-24 15:46:08 -04001833
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001834 http.request = new_request
1835 return http
John Asmuth864311d2014-04-24 15:46:08 -04001836
1837
1838def tunnel_patch(http):
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001839 """Tunnel PATCH requests over POST.
John Asmuth864311d2014-04-24 15:46:08 -04001840 Args:
1841 http - An instance of httplib2.Http
1842 or something that acts like it.
1843
1844 Returns:
1845 A modified instance of http that was passed in.
1846
1847 Example:
1848
1849 h = httplib2.Http()
1850 h = tunnel_patch(h, "my-app-name/6.0")
1851
1852 Useful if you are running on a platform that doesn't support PATCH.
1853 Apply this last if you are using OAuth 1.0, as changing the method
1854 will result in a different signature.
1855 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001856 request_orig = http.request
John Asmuth864311d2014-04-24 15:46:08 -04001857
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001858 # The closure that will replace 'httplib2.Http.request'.
1859 def new_request(
1860 uri,
1861 method="GET",
1862 body=None,
1863 headers=None,
1864 redirections=httplib2.DEFAULT_MAX_REDIRECTS,
1865 connection_type=None,
1866 ):
1867 """Modify the request headers to add the user-agent."""
1868 if headers is None:
1869 headers = {}
1870 if method == "PATCH":
1871 if "oauth_token" in headers.get("authorization", ""):
1872 LOGGER.warning(
1873 "OAuth 1.0 request made with Credentials after tunnel_patch."
1874 )
1875 headers["x-http-method-override"] = "PATCH"
1876 method = "POST"
1877 resp, content = request_orig(
1878 uri,
1879 method=method,
1880 body=body,
1881 headers=headers,
1882 redirections=redirections,
1883 connection_type=connection_type,
1884 )
1885 return resp, content
John Asmuth864311d2014-04-24 15:46:08 -04001886
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001887 http.request = new_request
1888 return http
Igor Maravić22435292017-01-19 22:28:22 +01001889
1890
1891def build_http():
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001892 """Builds httplib2.Http object
Igor Maravić22435292017-01-19 22:28:22 +01001893
1894 Returns:
1895 A httplib2.Http object, which is used to make http requests, and which has timeout set by default.
1896 To override default timeout call
1897
1898 socket.setdefaulttimeout(timeout_in_sec)
1899
1900 before interacting with this method.
1901 """
Bu Sun Kim66bb32c2019-10-30 10:11:58 -07001902 if socket.getdefaulttimeout() is not None:
1903 http_timeout = socket.getdefaulttimeout()
1904 else:
1905 http_timeout = DEFAULT_HTTP_TIMEOUT_SEC
Bu Sun Kimb3b773f2020-03-11 12:58:16 -07001906 http = httplib2.Http(timeout=http_timeout)
1907 # 308's are used by several Google APIs (Drive, YouTube)
1908 # for Resumable Uploads rather than Permanent Redirects.
1909 # This asks httplib2 to exclude 308s from the status codes
1910 # it treats as redirects
Bu Sun Kima480d532020-03-13 12:52:22 -07001911 try:
1912 http.redirect_codes = http.redirect_codes - {308}
1913 except AttributeError:
1914 # Apache Beam tests depend on this library and cannot
1915 # currently upgrade their httplib2 version
1916 # http.redirect_codes does not exist in previous versions
1917 # of httplib2, so pass
1918 pass
Bu Sun Kimb3b773f2020-03-11 12:58:16 -07001919
1920 return http