root / pithos / api / util.py @ 61efb530
History | View | Annotate | Download (30.6 kB)
1 |
# Copyright 2011 GRNET S.A. All rights reserved.
|
---|---|
2 |
#
|
3 |
# Redistribution and use in source and binary forms, with or
|
4 |
# without modification, are permitted provided that the following
|
5 |
# conditions are met:
|
6 |
#
|
7 |
# 1. Redistributions of source code must retain the above
|
8 |
# copyright notice, this list of conditions and the following
|
9 |
# disclaimer.
|
10 |
#
|
11 |
# 2. Redistributions in binary form must reproduce the above
|
12 |
# copyright notice, this list of conditions and the following
|
13 |
# disclaimer in the documentation and/or other materials
|
14 |
# provided with the distribution.
|
15 |
#
|
16 |
# THIS SOFTWARE IS PROVIDED BY GRNET S.A. ``AS IS'' AND ANY EXPRESS
|
17 |
# OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
|
18 |
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
|
19 |
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL GRNET S.A OR
|
20 |
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
21 |
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
22 |
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
|
23 |
# USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
|
24 |
# AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
25 |
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
|
26 |
# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
27 |
# POSSIBILITY OF SUCH DAMAGE.
|
28 |
#
|
29 |
# The views and conclusions contained in the software and
|
30 |
# documentation are those of the authors and should not be
|
31 |
# interpreted as representing official policies, either expressed
|
32 |
# or implied, of GRNET S.A.
|
33 |
|
34 |
from functools import wraps |
35 |
from time import time |
36 |
from traceback import format_exc |
37 |
from wsgiref.handlers import format_date_time |
38 |
from binascii import hexlify, unhexlify |
39 |
from datetime import datetime, tzinfo, timedelta |
40 |
|
41 |
from django.conf import settings |
42 |
from django.http import HttpResponse |
43 |
from django.utils import simplejson as json |
44 |
from django.utils.http import http_date, parse_etags |
45 |
from django.utils.encoding import smart_str |
46 |
|
47 |
from pithos.api.compat import parse_http_date_safe, parse_http_date |
48 |
from pithos.api.faults import (Fault, NotModified, BadRequest, Unauthorized, Forbidden, ItemNotFound, |
49 |
Conflict, LengthRequired, PreconditionFailed, RequestEntityTooLarge, |
50 |
RangeNotSatisfiable, ServiceUnavailable) |
51 |
from pithos.backends import connect_backend |
52 |
from pithos.backends.base import NotAllowedError, QuotaError |
53 |
|
54 |
import logging |
55 |
import re |
56 |
import hashlib |
57 |
import uuid |
58 |
import decimal |
59 |
|
60 |
logger = logging.getLogger(__name__) |
61 |
|
62 |
|
63 |
class UTC(tzinfo): |
64 |
def utcoffset(self, dt): |
65 |
return timedelta(0) |
66 |
|
67 |
def tzname(self, dt): |
68 |
return 'UTC' |
69 |
|
70 |
def dst(self, dt): |
71 |
return timedelta(0) |
72 |
|
73 |
def json_encode_decimal(obj): |
74 |
if isinstance(obj, decimal.Decimal): |
75 |
return str(obj) |
76 |
raise TypeError(repr(obj) + " is not JSON serializable") |
77 |
|
78 |
def isoformat(d): |
79 |
"""Return an ISO8601 date string that includes a timezone."""
|
80 |
|
81 |
return d.replace(tzinfo=UTC()).isoformat()
|
82 |
|
83 |
def rename_meta_key(d, old, new): |
84 |
if old not in d: |
85 |
return
|
86 |
d[new] = d[old] |
87 |
del(d[old])
|
88 |
|
89 |
def printable_header_dict(d): |
90 |
"""Format a meta dictionary for printing out json/xml.
|
91 |
|
92 |
Convert all keys to lower case and replace dashes with underscores.
|
93 |
Format 'last_modified' timestamp.
|
94 |
"""
|
95 |
|
96 |
d['last_modified'] = isoformat(datetime.fromtimestamp(d['last_modified'])) |
97 |
return dict([(k.lower().replace('-', '_'), v) for k, v in d.iteritems()]) |
98 |
|
99 |
def format_header_key(k): |
100 |
"""Convert underscores to dashes and capitalize intra-dash strings."""
|
101 |
return '-'.join([x.capitalize() for x in k.replace('_', '-').split('-')]) |
102 |
|
103 |
def get_header_prefix(request, prefix): |
104 |
"""Get all prefix-* request headers in a dict. Reformat keys with format_header_key()."""
|
105 |
|
106 |
prefix = 'HTTP_' + prefix.upper().replace('-', '_') |
107 |
# TODO: Document or remove '~' replacing.
|
108 |
return dict([(format_header_key(k[5:]), v.replace('~', '')) for k, v in request.META.iteritems() if k.startswith(prefix) and len(k) > len(prefix)]) |
109 |
|
110 |
def get_account_headers(request): |
111 |
meta = get_header_prefix(request, 'X-Account-Meta-')
|
112 |
groups = {} |
113 |
for k, v in get_header_prefix(request, 'X-Account-Group-').iteritems(): |
114 |
n = k[16:].lower()
|
115 |
if '-' in n or '_' in n: |
116 |
raise BadRequest('Bad characters in group name') |
117 |
groups[n] = v.replace(' ', '').split(',') |
118 |
while '' in groups[n]: |
119 |
groups[n].remove('')
|
120 |
return meta, groups
|
121 |
|
122 |
def put_account_headers(response, meta, groups, policy): |
123 |
if 'count' in meta: |
124 |
response['X-Account-Container-Count'] = meta['count'] |
125 |
if 'bytes' in meta: |
126 |
response['X-Account-Bytes-Used'] = meta['bytes'] |
127 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
128 |
for k in [x for x in meta.keys() if x.startswith('X-Account-Meta-')]: |
129 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
130 |
if 'until_timestamp' in meta: |
131 |
response['X-Account-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) |
132 |
for k, v in groups.iteritems(): |
133 |
k = smart_str(k, strings_only=True)
|
134 |
k = format_header_key('X-Account-Group-' + k)
|
135 |
v = smart_str(','.join(v), strings_only=True) |
136 |
response[k] = v |
137 |
for k, v in policy.iteritems(): |
138 |
response[smart_str(format_header_key('X-Account-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) |
139 |
|
140 |
def get_container_headers(request): |
141 |
meta = get_header_prefix(request, 'X-Container-Meta-')
|
142 |
policy = dict([(k[19:].lower(), v.replace(' ', '')) for k, v in get_header_prefix(request, 'X-Container-Policy-').iteritems()]) |
143 |
return meta, policy
|
144 |
|
145 |
def put_container_headers(request, response, meta, policy): |
146 |
if 'count' in meta: |
147 |
response['X-Container-Object-Count'] = meta['count'] |
148 |
if 'bytes' in meta: |
149 |
response['X-Container-Bytes-Used'] = meta['bytes'] |
150 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
151 |
for k in [x for x in meta.keys() if x.startswith('X-Container-Meta-')]: |
152 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
153 |
l = [smart_str(x, strings_only=True) for x in meta['object_meta'] if x.startswith('X-Object-Meta-')] |
154 |
response['X-Container-Object-Meta'] = ','.join([x[14:] for x in l]) |
155 |
response['X-Container-Block-Size'] = request.backend.block_size
|
156 |
response['X-Container-Block-Hash'] = request.backend.hash_algorithm
|
157 |
if 'until_timestamp' in meta: |
158 |
response['X-Container-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) |
159 |
for k, v in policy.iteritems(): |
160 |
response[smart_str(format_header_key('X-Container-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) |
161 |
|
162 |
def get_object_headers(request): |
163 |
meta = get_header_prefix(request, 'X-Object-Meta-')
|
164 |
if request.META.get('CONTENT_TYPE'): |
165 |
meta['Content-Type'] = request.META['CONTENT_TYPE'] |
166 |
if request.META.get('HTTP_CONTENT_ENCODING'): |
167 |
meta['Content-Encoding'] = request.META['HTTP_CONTENT_ENCODING'] |
168 |
if request.META.get('HTTP_CONTENT_DISPOSITION'): |
169 |
meta['Content-Disposition'] = request.META['HTTP_CONTENT_DISPOSITION'] |
170 |
if request.META.get('HTTP_X_OBJECT_MANIFEST'): |
171 |
meta['X-Object-Manifest'] = request.META['HTTP_X_OBJECT_MANIFEST'] |
172 |
return meta, get_sharing(request), get_public(request)
|
173 |
|
174 |
def put_object_headers(response, meta, restricted=False): |
175 |
response['ETag'] = meta['hash'] |
176 |
response['Content-Length'] = meta['bytes'] |
177 |
response['Content-Type'] = meta.get('Content-Type', 'application/octet-stream') |
178 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
179 |
if not restricted: |
180 |
response['X-Object-Modified-By'] = smart_str(meta['modified_by'], strings_only=True) |
181 |
response['X-Object-Version'] = meta['version'] |
182 |
response['X-Object-Version-Timestamp'] = http_date(int(meta['version_timestamp'])) |
183 |
for k in [x for x in meta.keys() if x.startswith('X-Object-Meta-')]: |
184 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
185 |
for k in ('Content-Encoding', 'Content-Disposition', 'X-Object-Manifest', |
186 |
'X-Object-Sharing', 'X-Object-Shared-By', 'X-Object-Allowed-To', |
187 |
'X-Object-Public'):
|
188 |
if k in meta: |
189 |
response[k] = smart_str(meta[k], strings_only=True)
|
190 |
else:
|
191 |
for k in ('Content-Encoding', 'Content-Disposition'): |
192 |
if k in meta: |
193 |
response[k] = meta[k] |
194 |
|
195 |
def update_manifest_meta(request, v_account, meta): |
196 |
"""Update metadata if the object has an X-Object-Manifest."""
|
197 |
|
198 |
if 'X-Object-Manifest' in meta: |
199 |
hash = ''
|
200 |
bytes = 0
|
201 |
try:
|
202 |
src_container, src_name = split_container_object_string('/' + meta['X-Object-Manifest']) |
203 |
objects = request.backend.list_objects(request.user_uniq, v_account, |
204 |
src_container, prefix=src_name, virtual=False)
|
205 |
for x in objects: |
206 |
src_meta = request.backend.get_object_meta(request.user_uniq, |
207 |
v_account, src_container, x[0], x[1]) |
208 |
hash += src_meta['hash'] |
209 |
bytes += src_meta['bytes'] |
210 |
except:
|
211 |
# Ignore errors.
|
212 |
return
|
213 |
meta['bytes'] = bytes |
214 |
md5 = hashlib.md5() |
215 |
md5.update(hash)
|
216 |
meta['hash'] = md5.hexdigest().lower()
|
217 |
|
218 |
def update_sharing_meta(request, permissions, v_account, v_container, v_object, meta): |
219 |
if permissions is None: |
220 |
return
|
221 |
allowed, perm_path, perms = permissions |
222 |
if len(perms) == 0: |
223 |
return
|
224 |
ret = [] |
225 |
r = ','.join(perms.get('read', [])) |
226 |
if r:
|
227 |
ret.append('read=' + r)
|
228 |
w = ','.join(perms.get('write', [])) |
229 |
if w:
|
230 |
ret.append('write=' + w)
|
231 |
meta['X-Object-Sharing'] = '; '.join(ret) |
232 |
if '/'.join((v_account, v_container, v_object)) != perm_path: |
233 |
meta['X-Object-Shared-By'] = perm_path
|
234 |
if request.user_uniq != v_account:
|
235 |
meta['X-Object-Allowed-To'] = allowed
|
236 |
|
237 |
def update_public_meta(public, meta): |
238 |
if not public: |
239 |
return
|
240 |
meta['X-Object-Public'] = public
|
241 |
|
242 |
def validate_modification_preconditions(request, meta): |
243 |
"""Check that the modified timestamp conforms with the preconditions set."""
|
244 |
|
245 |
if 'modified' not in meta: |
246 |
return # TODO: Always return? |
247 |
|
248 |
if_modified_since = request.META.get('HTTP_IF_MODIFIED_SINCE')
|
249 |
if if_modified_since is not None: |
250 |
if_modified_since = parse_http_date_safe(if_modified_since) |
251 |
if if_modified_since is not None and int(meta['modified']) <= if_modified_since: |
252 |
raise NotModified('Resource has not been modified') |
253 |
|
254 |
if_unmodified_since = request.META.get('HTTP_IF_UNMODIFIED_SINCE')
|
255 |
if if_unmodified_since is not None: |
256 |
if_unmodified_since = parse_http_date_safe(if_unmodified_since) |
257 |
if if_unmodified_since is not None and int(meta['modified']) > if_unmodified_since: |
258 |
raise PreconditionFailed('Resource has been modified') |
259 |
|
260 |
def validate_matching_preconditions(request, meta): |
261 |
"""Check that the ETag conforms with the preconditions set."""
|
262 |
|
263 |
hash = meta.get('hash', None) |
264 |
|
265 |
if_match = request.META.get('HTTP_IF_MATCH')
|
266 |
if if_match is not None: |
267 |
if hash is None: |
268 |
raise PreconditionFailed('Resource does not exist') |
269 |
if if_match != '*' and hash not in [x.lower() for x in parse_etags(if_match)]: |
270 |
raise PreconditionFailed('Resource ETag does not match') |
271 |
|
272 |
if_none_match = request.META.get('HTTP_IF_NONE_MATCH')
|
273 |
if if_none_match is not None: |
274 |
# TODO: If this passes, must ignore If-Modified-Since header.
|
275 |
if hash is not None: |
276 |
if if_none_match == '*' or hash in [x.lower() for x in parse_etags(if_none_match)]: |
277 |
# TODO: Continue if an If-Modified-Since header is present.
|
278 |
if request.method in ('HEAD', 'GET'): |
279 |
raise NotModified('Resource ETag matches') |
280 |
raise PreconditionFailed('Resource exists or ETag matches') |
281 |
|
282 |
def split_container_object_string(s): |
283 |
if not len(s) > 0 or s[0] != '/': |
284 |
raise ValueError |
285 |
s = s[1:]
|
286 |
pos = s.find('/')
|
287 |
if pos == -1: |
288 |
raise ValueError |
289 |
return s[:pos], s[(pos + 1):] |
290 |
|
291 |
def copy_or_move_object(request, src_account, src_container, src_name, dest_account, dest_container, dest_name, move=False): |
292 |
"""Copy or move an object."""
|
293 |
|
294 |
meta, permissions, public = get_object_headers(request) |
295 |
src_version = request.META.get('HTTP_X_SOURCE_VERSION')
|
296 |
try:
|
297 |
if move:
|
298 |
version_id = request.backend.move_object(request.user_uniq, src_account, src_container, src_name, |
299 |
dest_account, dest_container, dest_name, |
300 |
meta, False, permissions)
|
301 |
else:
|
302 |
version_id = request.backend.copy_object(request.user_uniq, src_account, src_container, src_name, |
303 |
dest_account, dest_container, dest_name, |
304 |
meta, False, permissions, src_version)
|
305 |
except NotAllowedError:
|
306 |
raise Forbidden('Not allowed') |
307 |
except (NameError, IndexError): |
308 |
raise ItemNotFound('Container or object does not exist') |
309 |
except ValueError: |
310 |
raise BadRequest('Invalid sharing header') |
311 |
except AttributeError, e: |
312 |
raise Conflict('\n'.join(e.data) + '\n') |
313 |
except QuotaError:
|
314 |
raise RequestEntityTooLarge('Quota exceeded') |
315 |
if public is not None: |
316 |
try:
|
317 |
request.backend.update_object_public(request.user_uniq, dest_account, dest_container, dest_name, public) |
318 |
except NotAllowedError:
|
319 |
raise Forbidden('Not allowed') |
320 |
except NameError: |
321 |
raise ItemNotFound('Object does not exist') |
322 |
return version_id
|
323 |
|
324 |
def get_int_parameter(p): |
325 |
if p is not None: |
326 |
try:
|
327 |
p = int(p)
|
328 |
except ValueError: |
329 |
return None |
330 |
if p < 0: |
331 |
return None |
332 |
return p
|
333 |
|
334 |
def get_content_length(request): |
335 |
content_length = get_int_parameter(request.META.get('CONTENT_LENGTH'))
|
336 |
if content_length is None: |
337 |
raise LengthRequired('Missing or invalid Content-Length header') |
338 |
return content_length
|
339 |
|
340 |
def get_range(request, size): |
341 |
"""Parse a Range header from the request.
|
342 |
|
343 |
Either returns None, when the header is not existent or should be ignored,
|
344 |
or a list of (offset, length) tuples - should be further checked.
|
345 |
"""
|
346 |
|
347 |
ranges = request.META.get('HTTP_RANGE', '').replace(' ', '') |
348 |
if not ranges.startswith('bytes='): |
349 |
return None |
350 |
|
351 |
ret = [] |
352 |
for r in (x.strip() for x in ranges[6:].split(',')): |
353 |
p = re.compile('^(?P<offset>\d*)-(?P<upto>\d*)$')
|
354 |
m = p.match(r) |
355 |
if not m: |
356 |
return None |
357 |
offset = m.group('offset')
|
358 |
upto = m.group('upto')
|
359 |
if offset == '' and upto == '': |
360 |
return None |
361 |
|
362 |
if offset != '': |
363 |
offset = int(offset)
|
364 |
if upto != '': |
365 |
upto = int(upto)
|
366 |
if offset > upto:
|
367 |
return None |
368 |
ret.append((offset, upto - offset + 1))
|
369 |
else:
|
370 |
ret.append((offset, size - offset)) |
371 |
else:
|
372 |
length = int(upto)
|
373 |
ret.append((size - length, length)) |
374 |
|
375 |
return ret
|
376 |
|
377 |
def get_content_range(request): |
378 |
"""Parse a Content-Range header from the request.
|
379 |
|
380 |
Either returns None, when the header is not existent or should be ignored,
|
381 |
or an (offset, length, total) tuple - check as length, total may be None.
|
382 |
Returns (None, None, None) if the provided range is '*/*'.
|
383 |
"""
|
384 |
|
385 |
ranges = request.META.get('HTTP_CONTENT_RANGE', '') |
386 |
if not ranges: |
387 |
return None |
388 |
|
389 |
p = re.compile('^bytes (?P<offset>\d+)-(?P<upto>\d*)/(?P<total>(\d+|\*))$')
|
390 |
m = p.match(ranges) |
391 |
if not m: |
392 |
if ranges == 'bytes */*': |
393 |
return (None, None, None) |
394 |
return None |
395 |
offset = int(m.group('offset')) |
396 |
upto = m.group('upto')
|
397 |
total = m.group('total')
|
398 |
if upto != '': |
399 |
upto = int(upto)
|
400 |
else:
|
401 |
upto = None
|
402 |
if total != '*': |
403 |
total = int(total)
|
404 |
else:
|
405 |
total = None
|
406 |
if (upto is not None and offset > upto) or \ |
407 |
(total is not None and offset >= total) or \ |
408 |
(total is not None and upto is not None and upto >= total): |
409 |
return None |
410 |
|
411 |
if upto is None: |
412 |
length = None
|
413 |
else:
|
414 |
length = upto - offset + 1
|
415 |
return (offset, length, total)
|
416 |
|
417 |
def get_sharing(request): |
418 |
"""Parse an X-Object-Sharing header from the request.
|
419 |
|
420 |
Raises BadRequest on error.
|
421 |
"""
|
422 |
|
423 |
permissions = request.META.get('HTTP_X_OBJECT_SHARING')
|
424 |
if permissions is None: |
425 |
return None |
426 |
|
427 |
# TODO: Document or remove '~' replacing.
|
428 |
permissions = permissions.replace('~', '') |
429 |
|
430 |
ret = {} |
431 |
permissions = permissions.replace(' ', '') |
432 |
if permissions == '': |
433 |
return ret
|
434 |
for perm in (x for x in permissions.split(';')): |
435 |
if perm.startswith('read='): |
436 |
ret['read'] = list(set([v.replace(' ','').lower() for v in perm[5:].split(',')])) |
437 |
if '' in ret['read']: |
438 |
ret['read'].remove('') |
439 |
if '*' in ret['read']: |
440 |
ret['read'] = ['*'] |
441 |
if len(ret['read']) == 0: |
442 |
raise BadRequest('Bad X-Object-Sharing header value') |
443 |
elif perm.startswith('write='): |
444 |
ret['write'] = list(set([v.replace(' ','').lower() for v in perm[6:].split(',')])) |
445 |
if '' in ret['write']: |
446 |
ret['write'].remove('') |
447 |
if '*' in ret['write']: |
448 |
ret['write'] = ['*'] |
449 |
if len(ret['write']) == 0: |
450 |
raise BadRequest('Bad X-Object-Sharing header value') |
451 |
else:
|
452 |
raise BadRequest('Bad X-Object-Sharing header value') |
453 |
|
454 |
# Keep duplicates only in write list.
|
455 |
dups = [x for x in ret.get('read', []) if x in ret.get('write', []) and x != '*'] |
456 |
if dups:
|
457 |
for x in dups: |
458 |
ret['read'].remove(x)
|
459 |
if len(ret['read']) == 0: |
460 |
del(ret['read']) |
461 |
|
462 |
return ret
|
463 |
|
464 |
def get_public(request): |
465 |
"""Parse an X-Object-Public header from the request.
|
466 |
|
467 |
Raises BadRequest on error.
|
468 |
"""
|
469 |
|
470 |
public = request.META.get('HTTP_X_OBJECT_PUBLIC')
|
471 |
if public is None: |
472 |
return None |
473 |
|
474 |
public = public.replace(' ', '').lower() |
475 |
if public == 'true': |
476 |
return True |
477 |
elif public == 'false' or public == '': |
478 |
return False |
479 |
raise BadRequest('Bad X-Object-Public header value') |
480 |
|
481 |
def raw_input_socket(request): |
482 |
"""Return the socket for reading the rest of the request."""
|
483 |
|
484 |
server_software = request.META.get('SERVER_SOFTWARE')
|
485 |
if server_software and server_software.startswith('mod_python'): |
486 |
return request._req
|
487 |
if 'wsgi.input' in request.environ: |
488 |
return request.environ['wsgi.input'] |
489 |
raise ServiceUnavailable('Unknown server software') |
490 |
|
491 |
MAX_UPLOAD_SIZE = 5 * (1024 * 1024 * 1024) # 5GB |
492 |
|
493 |
def socket_read_iterator(request, length=0, blocksize=4096): |
494 |
"""Return a maximum of blocksize data read from the socket in each iteration.
|
495 |
|
496 |
Read up to 'length'. If 'length' is negative, will attempt a chunked read.
|
497 |
The maximum ammount of data read is controlled by MAX_UPLOAD_SIZE.
|
498 |
"""
|
499 |
|
500 |
sock = raw_input_socket(request) |
501 |
if length < 0: # Chunked transfers |
502 |
# Small version (server does the dechunking).
|
503 |
if request.environ.get('mod_wsgi.input_chunked', None) or request.META['SERVER_SOFTWARE'].startswith('gunicorn'): |
504 |
while length < MAX_UPLOAD_SIZE:
|
505 |
data = sock.read(blocksize) |
506 |
if data == '': |
507 |
return
|
508 |
yield data
|
509 |
raise BadRequest('Maximum size is reached') |
510 |
|
511 |
# Long version (do the dechunking).
|
512 |
data = ''
|
513 |
while length < MAX_UPLOAD_SIZE:
|
514 |
# Get chunk size.
|
515 |
if hasattr(sock, 'readline'): |
516 |
chunk_length = sock.readline() |
517 |
else:
|
518 |
chunk_length = ''
|
519 |
while chunk_length[-1:] != '\n': |
520 |
chunk_length += sock.read(1)
|
521 |
chunk_length.strip() |
522 |
pos = chunk_length.find(';')
|
523 |
if pos >= 0: |
524 |
chunk_length = chunk_length[:pos] |
525 |
try:
|
526 |
chunk_length = int(chunk_length, 16) |
527 |
except Exception, e: |
528 |
raise BadRequest('Bad chunk size') # TODO: Change to something more appropriate. |
529 |
# Check if done.
|
530 |
if chunk_length == 0: |
531 |
if len(data) > 0: |
532 |
yield data
|
533 |
return
|
534 |
# Get the actual data.
|
535 |
while chunk_length > 0: |
536 |
chunk = sock.read(min(chunk_length, blocksize))
|
537 |
chunk_length -= len(chunk)
|
538 |
if length > 0: |
539 |
length += len(chunk)
|
540 |
data += chunk |
541 |
if len(data) >= blocksize: |
542 |
ret = data[:blocksize] |
543 |
data = data[blocksize:] |
544 |
yield ret
|
545 |
sock.read(2) # CRLF |
546 |
raise BadRequest('Maximum size is reached') |
547 |
else:
|
548 |
if length > MAX_UPLOAD_SIZE:
|
549 |
raise BadRequest('Maximum size is reached') |
550 |
while length > 0: |
551 |
data = sock.read(min(length, blocksize))
|
552 |
if not data: |
553 |
raise BadRequest()
|
554 |
length -= len(data)
|
555 |
yield data
|
556 |
|
557 |
class ObjectWrapper(object): |
558 |
"""Return the object's data block-per-block in each iteration.
|
559 |
|
560 |
Read from the object using the offset and length provided in each entry of the range list.
|
561 |
"""
|
562 |
|
563 |
def __init__(self, backend, ranges, sizes, hashmaps, boundary): |
564 |
self.backend = backend
|
565 |
self.ranges = ranges
|
566 |
self.sizes = sizes
|
567 |
self.hashmaps = hashmaps
|
568 |
self.boundary = boundary
|
569 |
self.size = sum(self.sizes) |
570 |
|
571 |
self.file_index = 0 |
572 |
self.block_index = 0 |
573 |
self.block_hash = -1 |
574 |
self.block = '' |
575 |
|
576 |
self.range_index = -1 |
577 |
self.offset, self.length = self.ranges[0] |
578 |
|
579 |
def __iter__(self): |
580 |
return self |
581 |
|
582 |
def part_iterator(self): |
583 |
if self.length > 0: |
584 |
# Get the file for the current offset.
|
585 |
file_size = self.sizes[self.file_index] |
586 |
while self.offset >= file_size: |
587 |
self.offset -= file_size
|
588 |
self.file_index += 1 |
589 |
file_size = self.sizes[self.file_index] |
590 |
|
591 |
# Get the block for the current position.
|
592 |
self.block_index = int(self.offset / self.backend.block_size) |
593 |
if self.block_hash != self.hashmaps[self.file_index][self.block_index]: |
594 |
self.block_hash = self.hashmaps[self.file_index][self.block_index] |
595 |
try:
|
596 |
self.block = self.backend.get_block(self.block_hash) |
597 |
except NameError: |
598 |
raise ItemNotFound('Block does not exist') |
599 |
|
600 |
# Get the data from the block.
|
601 |
bo = self.offset % self.backend.block_size |
602 |
bl = min(self.length, len(self.block) - bo) |
603 |
data = self.block[bo:bo + bl]
|
604 |
self.offset += bl
|
605 |
self.length -= bl
|
606 |
return data
|
607 |
else:
|
608 |
raise StopIteration |
609 |
|
610 |
def next(self): |
611 |
if len(self.ranges) == 1: |
612 |
return self.part_iterator() |
613 |
if self.range_index == len(self.ranges): |
614 |
raise StopIteration |
615 |
try:
|
616 |
if self.range_index == -1: |
617 |
raise StopIteration |
618 |
return self.part_iterator() |
619 |
except StopIteration: |
620 |
self.range_index += 1 |
621 |
out = [] |
622 |
if self.range_index < len(self.ranges): |
623 |
# Part header.
|
624 |
self.offset, self.length = self.ranges[self.range_index] |
625 |
self.file_index = 0 |
626 |
if self.range_index > 0: |
627 |
out.append('')
|
628 |
out.append('--' + self.boundary) |
629 |
out.append('Content-Range: bytes %d-%d/%d' % (self.offset, self.offset + self.length - 1, self.size)) |
630 |
out.append('Content-Transfer-Encoding: binary')
|
631 |
out.append('')
|
632 |
out.append('')
|
633 |
return '\r\n'.join(out) |
634 |
else:
|
635 |
# Footer.
|
636 |
out.append('')
|
637 |
out.append('--' + self.boundary + '--') |
638 |
out.append('')
|
639 |
return '\r\n'.join(out) |
640 |
|
641 |
def object_data_response(request, sizes, hashmaps, meta, public=False): |
642 |
"""Get the HttpResponse object for replying with the object's data."""
|
643 |
|
644 |
# Range handling.
|
645 |
size = sum(sizes)
|
646 |
ranges = get_range(request, size) |
647 |
if ranges is None: |
648 |
ranges = [(0, size)]
|
649 |
ret = 200
|
650 |
else:
|
651 |
check = [True for offset, length in ranges if |
652 |
length <= 0 or length > size or |
653 |
offset < 0 or offset >= size or |
654 |
offset + length > size] |
655 |
if len(check) > 0: |
656 |
raise RangeNotSatisfiable('Requested range exceeds object limits') |
657 |
ret = 206
|
658 |
if_range = request.META.get('HTTP_IF_RANGE')
|
659 |
if if_range:
|
660 |
try:
|
661 |
# Modification time has passed instead.
|
662 |
last_modified = parse_http_date(if_range) |
663 |
if last_modified != meta['modified']: |
664 |
ranges = [(0, size)]
|
665 |
ret = 200
|
666 |
except ValueError: |
667 |
if if_range != meta['hash']: |
668 |
ranges = [(0, size)]
|
669 |
ret = 200
|
670 |
|
671 |
if ret == 206 and len(ranges) > 1: |
672 |
boundary = uuid.uuid4().hex |
673 |
else:
|
674 |
boundary = ''
|
675 |
wrapper = ObjectWrapper(request.backend, ranges, sizes, hashmaps, boundary) |
676 |
response = HttpResponse(wrapper, status=ret) |
677 |
put_object_headers(response, meta, public) |
678 |
if ret == 206: |
679 |
if len(ranges) == 1: |
680 |
offset, length = ranges[0]
|
681 |
response['Content-Length'] = length # Update with the correct length. |
682 |
response['Content-Range'] = 'bytes %d-%d/%d' % (offset, offset + length - 1, size) |
683 |
else:
|
684 |
del(response['Content-Length']) |
685 |
response['Content-Type'] = 'multipart/byteranges; boundary=%s' % (boundary,) |
686 |
return response
|
687 |
|
688 |
def put_object_block(request, hashmap, data, offset): |
689 |
"""Put one block of data at the given offset."""
|
690 |
|
691 |
bi = int(offset / request.backend.block_size)
|
692 |
bo = offset % request.backend.block_size |
693 |
bl = min(len(data), request.backend.block_size - bo) |
694 |
if bi < len(hashmap): |
695 |
hashmap[bi] = request.backend.update_block(hashmap[bi], data[:bl], bo) |
696 |
else:
|
697 |
hashmap.append(request.backend.put_block(('\x00' * bo) + data[:bl]))
|
698 |
return bl # Return ammount of data written. |
699 |
|
700 |
def hashmap_hash(request, hashmap): |
701 |
"""Produce the root hash, treating the hashmap as a Merkle-like tree."""
|
702 |
|
703 |
def subhash(d): |
704 |
h = hashlib.new(request.backend.hash_algorithm) |
705 |
h.update(d) |
706 |
return h.digest()
|
707 |
|
708 |
if len(hashmap) == 0: |
709 |
return hexlify(subhash('')) |
710 |
if len(hashmap) == 1: |
711 |
return hashmap[0] |
712 |
|
713 |
s = 2
|
714 |
while s < len(hashmap): |
715 |
s = s * 2
|
716 |
h = [unhexlify(x) for x in hashmap] |
717 |
h += [('\x00' * len(h[0]))] * (s - len(hashmap)) |
718 |
while len(h) > 1: |
719 |
h = [subhash(h[x] + h[x + 1]) for x in range(0, len(h), 2)] |
720 |
return hexlify(h[0]) |
721 |
|
722 |
def update_response_headers(request, response): |
723 |
if request.serialization == 'xml': |
724 |
response['Content-Type'] = 'application/xml; charset=UTF-8' |
725 |
elif request.serialization == 'json': |
726 |
response['Content-Type'] = 'application/json; charset=UTF-8' |
727 |
elif not response['Content-Type']: |
728 |
response['Content-Type'] = 'text/plain; charset=UTF-8' |
729 |
|
730 |
if not response.has_header('Content-Length') and not (response.has_header('Content-Type') and response['Content-Type'].startswith('multipart/byteranges')): |
731 |
response['Content-Length'] = len(response.content) |
732 |
|
733 |
if settings.TEST:
|
734 |
response['Date'] = format_date_time(time())
|
735 |
|
736 |
def render_fault(request, fault): |
737 |
if settings.DEBUG or settings.TEST: |
738 |
fault.details = format_exc(fault) |
739 |
|
740 |
request.serialization = 'text'
|
741 |
data = '\n'.join((fault.message, fault.details)) + '\n' |
742 |
response = HttpResponse(data, status=fault.code) |
743 |
update_response_headers(request, response) |
744 |
return response
|
745 |
|
746 |
def request_serialization(request, format_allowed=False): |
747 |
"""Return the serialization format requested.
|
748 |
|
749 |
Valid formats are 'text' and 'json', 'xml' if 'format_allowed' is True.
|
750 |
"""
|
751 |
|
752 |
if not format_allowed: |
753 |
return 'text' |
754 |
|
755 |
format = request.GET.get('format')
|
756 |
if format == 'json': |
757 |
return 'json' |
758 |
elif format == 'xml': |
759 |
return 'xml' |
760 |
|
761 |
for item in request.META.get('HTTP_ACCEPT', '').split(','): |
762 |
accept, sep, rest = item.strip().partition(';')
|
763 |
if accept == 'application/json': |
764 |
return 'json' |
765 |
elif accept == 'application/xml' or accept == 'text/xml': |
766 |
return 'xml' |
767 |
|
768 |
return 'text' |
769 |
|
770 |
def api_method(http_method=None, format_allowed=False, user_required=True): |
771 |
"""Decorator function for views that implement an API method."""
|
772 |
|
773 |
def decorator(func): |
774 |
@wraps(func)
|
775 |
def wrapper(request, *args, **kwargs): |
776 |
try:
|
777 |
if http_method and request.method != http_method: |
778 |
raise BadRequest('Method not allowed.') |
779 |
if user_required and getattr(request, 'user', None) is None: |
780 |
raise Unauthorized('Access denied') |
781 |
|
782 |
# The args variable may contain up to (account, container, object).
|
783 |
if len(args) > 1 and len(args[1]) > 256: |
784 |
raise BadRequest('Container name too large.') |
785 |
if len(args) > 2 and len(args[2]) > 1024: |
786 |
raise BadRequest('Object name too large.') |
787 |
|
788 |
# Fill in custom request variables.
|
789 |
request.serialization = request_serialization(request, format_allowed) |
790 |
request.backend = connect_backend() |
791 |
|
792 |
response = func(request, *args, **kwargs) |
793 |
update_response_headers(request, response) |
794 |
return response
|
795 |
except Fault, fault:
|
796 |
return render_fault(request, fault)
|
797 |
except BaseException, e: |
798 |
logger.exception('Unexpected error: %s' % e)
|
799 |
fault = ServiceUnavailable('Unexpected error')
|
800 |
return render_fault(request, fault)
|
801 |
finally:
|
802 |
if getattr(request, 'backend', None) is not None: |
803 |
request.backend.wrapper.conn.close() |
804 |
return wrapper
|
805 |
return decorator
|