root / pithos / api / util.py @ 4a1c29ea
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 |
|
61 |
logger = logging.getLogger(__name__) |
62 |
|
63 |
|
64 |
class UTC(tzinfo): |
65 |
def utcoffset(self, dt): |
66 |
return timedelta(0) |
67 |
|
68 |
def tzname(self, dt): |
69 |
return 'UTC' |
70 |
|
71 |
def dst(self, dt): |
72 |
return timedelta(0) |
73 |
|
74 |
def json_encode_decimal(obj): |
75 |
if isinstance(obj, decimal.Decimal): |
76 |
return str(obj) |
77 |
raise TypeError(repr(obj) + " is not JSON serializable") |
78 |
|
79 |
def isoformat(d): |
80 |
"""Return an ISO8601 date string that includes a timezone."""
|
81 |
|
82 |
return d.replace(tzinfo=UTC()).isoformat()
|
83 |
|
84 |
def rename_meta_key(d, old, new): |
85 |
if old not in d: |
86 |
return
|
87 |
d[new] = d[old] |
88 |
del(d[old])
|
89 |
|
90 |
def printable_header_dict(d): |
91 |
"""Format a meta dictionary for printing out json/xml.
|
92 |
|
93 |
Convert all keys to lower case and replace dashes with underscores.
|
94 |
Format 'last_modified' timestamp.
|
95 |
"""
|
96 |
|
97 |
d['last_modified'] = isoformat(datetime.fromtimestamp(d['last_modified'])) |
98 |
return dict([(k.lower().replace('-', '_'), v) for k, v in d.iteritems()]) |
99 |
|
100 |
def format_header_key(k): |
101 |
"""Convert underscores to dashes and capitalize intra-dash strings."""
|
102 |
return '-'.join([x.capitalize() for x in k.replace('_', '-').split('-')]) |
103 |
|
104 |
def get_header_prefix(request, prefix): |
105 |
"""Get all prefix-* request headers in a dict. Reformat keys with format_header_key()."""
|
106 |
|
107 |
prefix = 'HTTP_' + prefix.upper().replace('-', '_') |
108 |
# TODO: Document or remove '~' replacing.
|
109 |
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)]) |
110 |
|
111 |
def get_account_headers(request): |
112 |
meta = get_header_prefix(request, 'X-Account-Meta-')
|
113 |
groups = {} |
114 |
for k, v in get_header_prefix(request, 'X-Account-Group-').iteritems(): |
115 |
n = k[16:].lower()
|
116 |
if '-' in n or '_' in n: |
117 |
raise BadRequest('Bad characters in group name') |
118 |
groups[n] = v.replace(' ', '').split(',') |
119 |
while '' in groups[n]: |
120 |
groups[n].remove('')
|
121 |
return meta, groups
|
122 |
|
123 |
def put_account_headers(response, meta, groups, policy): |
124 |
if 'count' in meta: |
125 |
response['X-Account-Container-Count'] = meta['count'] |
126 |
if 'bytes' in meta: |
127 |
response['X-Account-Bytes-Used'] = meta['bytes'] |
128 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
129 |
for k in [x for x in meta.keys() if x.startswith('X-Account-Meta-')]: |
130 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
131 |
if 'until_timestamp' in meta: |
132 |
response['X-Account-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) |
133 |
for k, v in groups.iteritems(): |
134 |
k = smart_str(k, strings_only=True)
|
135 |
k = format_header_key('X-Account-Group-' + k)
|
136 |
v = smart_str(','.join(v), strings_only=True) |
137 |
response[k] = v |
138 |
for k, v in policy.iteritems(): |
139 |
response[smart_str(format_header_key('X-Account-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) |
140 |
|
141 |
def get_container_headers(request): |
142 |
meta = get_header_prefix(request, 'X-Container-Meta-')
|
143 |
policy = dict([(k[19:].lower(), v.replace(' ', '')) for k, v in get_header_prefix(request, 'X-Container-Policy-').iteritems()]) |
144 |
return meta, policy
|
145 |
|
146 |
def put_container_headers(request, response, meta, policy): |
147 |
if 'count' in meta: |
148 |
response['X-Container-Object-Count'] = meta['count'] |
149 |
if 'bytes' in meta: |
150 |
response['X-Container-Bytes-Used'] = meta['bytes'] |
151 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
152 |
for k in [x for x in meta.keys() if x.startswith('X-Container-Meta-')]: |
153 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
154 |
l = [smart_str(x, strings_only=True) for x in meta['object_meta'] if x.startswith('X-Object-Meta-')] |
155 |
response['X-Container-Object-Meta'] = ','.join([x[14:] for x in l]) |
156 |
response['X-Container-Block-Size'] = request.backend.block_size
|
157 |
response['X-Container-Block-Hash'] = request.backend.hash_algorithm
|
158 |
if 'until_timestamp' in meta: |
159 |
response['X-Container-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) |
160 |
for k, v in policy.iteritems(): |
161 |
response[smart_str(format_header_key('X-Container-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) |
162 |
|
163 |
def get_object_headers(request): |
164 |
meta = get_header_prefix(request, 'X-Object-Meta-')
|
165 |
if request.META.get('CONTENT_TYPE'): |
166 |
meta['Content-Type'] = request.META['CONTENT_TYPE'] |
167 |
if request.META.get('HTTP_CONTENT_ENCODING'): |
168 |
meta['Content-Encoding'] = request.META['HTTP_CONTENT_ENCODING'] |
169 |
if request.META.get('HTTP_CONTENT_DISPOSITION'): |
170 |
meta['Content-Disposition'] = request.META['HTTP_CONTENT_DISPOSITION'] |
171 |
if request.META.get('HTTP_X_OBJECT_MANIFEST'): |
172 |
meta['X-Object-Manifest'] = request.META['HTTP_X_OBJECT_MANIFEST'] |
173 |
return meta, get_sharing(request), get_public(request)
|
174 |
|
175 |
def put_object_headers(response, meta, restricted=False): |
176 |
response['ETag'] = meta['ETag'] |
177 |
response['Content-Length'] = meta['bytes'] |
178 |
response['Content-Type'] = meta.get('Content-Type', 'application/octet-stream') |
179 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
180 |
if not restricted: |
181 |
response['X-Object-Hash'] = meta['hash'] |
182 |
response['X-Object-Modified-By'] = smart_str(meta['modified_by'], strings_only=True) |
183 |
response['X-Object-Version'] = meta['version'] |
184 |
response['X-Object-Version-Timestamp'] = http_date(int(meta['version_timestamp'])) |
185 |
for k in [x for x in meta.keys() if x.startswith('X-Object-Meta-')]: |
186 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
187 |
for k in ('Content-Encoding', 'Content-Disposition', 'X-Object-Manifest', |
188 |
'X-Object-Sharing', 'X-Object-Shared-By', 'X-Object-Allowed-To', |
189 |
'X-Object-Public'):
|
190 |
if k in meta: |
191 |
response[k] = smart_str(meta[k], strings_only=True)
|
192 |
else:
|
193 |
for k in ('Content-Encoding', 'Content-Disposition'): |
194 |
if k in meta: |
195 |
response[k] = meta[k] |
196 |
|
197 |
def update_manifest_meta(request, v_account, meta): |
198 |
"""Update metadata if the object has an X-Object-Manifest."""
|
199 |
|
200 |
if 'X-Object-Manifest' in meta: |
201 |
etag = ''
|
202 |
bytes = 0
|
203 |
try:
|
204 |
src_container, src_name = split_container_object_string('/' + meta['X-Object-Manifest']) |
205 |
objects = request.backend.list_objects(request.user_uniq, v_account, |
206 |
src_container, prefix=src_name, virtual=False)
|
207 |
for x in objects: |
208 |
src_meta = request.backend.get_object_meta(request.user_uniq, |
209 |
v_account, src_container, x[0], x[1]) |
210 |
etag += src_meta['ETag']
|
211 |
bytes += src_meta['bytes'] |
212 |
except:
|
213 |
# Ignore errors.
|
214 |
return
|
215 |
meta['bytes'] = bytes |
216 |
md5 = hashlib.md5() |
217 |
md5.update(etag) |
218 |
meta['ETag'] = md5.hexdigest().lower()
|
219 |
|
220 |
def update_sharing_meta(request, permissions, v_account, v_container, v_object, meta): |
221 |
if permissions is None: |
222 |
return
|
223 |
allowed, perm_path, perms = permissions |
224 |
if len(perms) == 0: |
225 |
return
|
226 |
ret = [] |
227 |
r = ','.join(perms.get('read', [])) |
228 |
if r:
|
229 |
ret.append('read=' + r)
|
230 |
w = ','.join(perms.get('write', [])) |
231 |
if w:
|
232 |
ret.append('write=' + w)
|
233 |
meta['X-Object-Sharing'] = '; '.join(ret) |
234 |
if '/'.join((v_account, v_container, v_object)) != perm_path: |
235 |
meta['X-Object-Shared-By'] = perm_path
|
236 |
if request.user_uniq != v_account:
|
237 |
meta['X-Object-Allowed-To'] = allowed
|
238 |
|
239 |
def update_public_meta(public, meta): |
240 |
if not public: |
241 |
return
|
242 |
meta['X-Object-Public'] = public
|
243 |
|
244 |
def validate_modification_preconditions(request, meta): |
245 |
"""Check that the modified timestamp conforms with the preconditions set."""
|
246 |
|
247 |
if 'modified' not in meta: |
248 |
return # TODO: Always return? |
249 |
|
250 |
if_modified_since = request.META.get('HTTP_IF_MODIFIED_SINCE')
|
251 |
if if_modified_since is not None: |
252 |
if_modified_since = parse_http_date_safe(if_modified_since) |
253 |
if if_modified_since is not None and int(meta['modified']) <= if_modified_since: |
254 |
raise NotModified('Resource has not been modified') |
255 |
|
256 |
if_unmodified_since = request.META.get('HTTP_IF_UNMODIFIED_SINCE')
|
257 |
if if_unmodified_since is not None: |
258 |
if_unmodified_since = parse_http_date_safe(if_unmodified_since) |
259 |
if if_unmodified_since is not None and int(meta['modified']) > if_unmodified_since: |
260 |
raise PreconditionFailed('Resource has been modified') |
261 |
|
262 |
def validate_matching_preconditions(request, meta): |
263 |
"""Check that the ETag conforms with the preconditions set."""
|
264 |
|
265 |
etag = meta.get('ETag', None) |
266 |
|
267 |
if_match = request.META.get('HTTP_IF_MATCH')
|
268 |
if if_match is not None: |
269 |
if etag is None: |
270 |
raise PreconditionFailed('Resource does not exist') |
271 |
if if_match != '*' and etag not in [x.lower() for x in parse_etags(if_match)]: |
272 |
raise PreconditionFailed('Resource ETag does not match') |
273 |
|
274 |
if_none_match = request.META.get('HTTP_IF_NONE_MATCH')
|
275 |
if if_none_match is not None: |
276 |
# TODO: If this passes, must ignore If-Modified-Since header.
|
277 |
if etag is not None: |
278 |
if if_none_match == '*' or etag in [x.lower() for x in parse_etags(if_none_match)]: |
279 |
# TODO: Continue if an If-Modified-Since header is present.
|
280 |
if request.method in ('HEAD', 'GET'): |
281 |
raise NotModified('Resource ETag matches') |
282 |
raise PreconditionFailed('Resource exists or ETag matches') |
283 |
|
284 |
def split_container_object_string(s): |
285 |
if not len(s) > 0 or s[0] != '/': |
286 |
raise ValueError |
287 |
s = s[1:]
|
288 |
pos = s.find('/')
|
289 |
if pos == -1: |
290 |
raise ValueError |
291 |
return s[:pos], s[(pos + 1):] |
292 |
|
293 |
def copy_or_move_object(request, src_account, src_container, src_name, dest_account, dest_container, dest_name, move=False): |
294 |
"""Copy or move an object."""
|
295 |
|
296 |
meta, permissions, public = get_object_headers(request) |
297 |
src_version = request.META.get('HTTP_X_SOURCE_VERSION')
|
298 |
try:
|
299 |
if move:
|
300 |
version_id = request.backend.move_object(request.user_uniq, src_account, src_container, src_name, |
301 |
dest_account, dest_container, dest_name, |
302 |
meta, False, permissions)
|
303 |
else:
|
304 |
version_id = request.backend.copy_object(request.user_uniq, src_account, src_container, src_name, |
305 |
dest_account, dest_container, dest_name, |
306 |
meta, False, permissions, src_version)
|
307 |
except NotAllowedError:
|
308 |
raise Forbidden('Not allowed') |
309 |
except (NameError, IndexError): |
310 |
raise ItemNotFound('Container or object does not exist') |
311 |
except ValueError: |
312 |
raise BadRequest('Invalid sharing header') |
313 |
except AttributeError, e: |
314 |
raise Conflict('\n'.join(e.data) + '\n') |
315 |
except QuotaError:
|
316 |
raise RequestEntityTooLarge('Quota exceeded') |
317 |
if public is not None: |
318 |
try:
|
319 |
request.backend.update_object_public(request.user_uniq, dest_account, dest_container, dest_name, public) |
320 |
except NotAllowedError:
|
321 |
raise Forbidden('Not allowed') |
322 |
except NameError: |
323 |
raise ItemNotFound('Object does not exist') |
324 |
return version_id
|
325 |
|
326 |
def get_int_parameter(p): |
327 |
if p is not None: |
328 |
try:
|
329 |
p = int(p)
|
330 |
except ValueError: |
331 |
return None |
332 |
if p < 0: |
333 |
return None |
334 |
return p
|
335 |
|
336 |
def get_content_length(request): |
337 |
content_length = get_int_parameter(request.META.get('CONTENT_LENGTH'))
|
338 |
if content_length is None: |
339 |
raise LengthRequired('Missing or invalid Content-Length header') |
340 |
return content_length
|
341 |
|
342 |
def get_range(request, size): |
343 |
"""Parse a Range header from the request.
|
344 |
|
345 |
Either returns None, when the header is not existent or should be ignored,
|
346 |
or a list of (offset, length) tuples - should be further checked.
|
347 |
"""
|
348 |
|
349 |
ranges = request.META.get('HTTP_RANGE', '').replace(' ', '') |
350 |
if not ranges.startswith('bytes='): |
351 |
return None |
352 |
|
353 |
ret = [] |
354 |
for r in (x.strip() for x in ranges[6:].split(',')): |
355 |
p = re.compile('^(?P<offset>\d*)-(?P<upto>\d*)$')
|
356 |
m = p.match(r) |
357 |
if not m: |
358 |
return None |
359 |
offset = m.group('offset')
|
360 |
upto = m.group('upto')
|
361 |
if offset == '' and upto == '': |
362 |
return None |
363 |
|
364 |
if offset != '': |
365 |
offset = int(offset)
|
366 |
if upto != '': |
367 |
upto = int(upto)
|
368 |
if offset > upto:
|
369 |
return None |
370 |
ret.append((offset, upto - offset + 1))
|
371 |
else:
|
372 |
ret.append((offset, size - offset)) |
373 |
else:
|
374 |
length = int(upto)
|
375 |
ret.append((size - length, length)) |
376 |
|
377 |
return ret
|
378 |
|
379 |
def get_content_range(request): |
380 |
"""Parse a Content-Range header from the request.
|
381 |
|
382 |
Either returns None, when the header is not existent or should be ignored,
|
383 |
or an (offset, length, total) tuple - check as length, total may be None.
|
384 |
Returns (None, None, None) if the provided range is '*/*'.
|
385 |
"""
|
386 |
|
387 |
ranges = request.META.get('HTTP_CONTENT_RANGE', '') |
388 |
if not ranges: |
389 |
return None |
390 |
|
391 |
p = re.compile('^bytes (?P<offset>\d+)-(?P<upto>\d*)/(?P<total>(\d+|\*))$')
|
392 |
m = p.match(ranges) |
393 |
if not m: |
394 |
if ranges == 'bytes */*': |
395 |
return (None, None, None) |
396 |
return None |
397 |
offset = int(m.group('offset')) |
398 |
upto = m.group('upto')
|
399 |
total = m.group('total')
|
400 |
if upto != '': |
401 |
upto = int(upto)
|
402 |
else:
|
403 |
upto = None
|
404 |
if total != '*': |
405 |
total = int(total)
|
406 |
else:
|
407 |
total = None
|
408 |
if (upto is not None and offset > upto) or \ |
409 |
(total is not None and offset >= total) or \ |
410 |
(total is not None and upto is not None and upto >= total): |
411 |
return None |
412 |
|
413 |
if upto is None: |
414 |
length = None
|
415 |
else:
|
416 |
length = upto - offset + 1
|
417 |
return (offset, length, total)
|
418 |
|
419 |
def get_sharing(request): |
420 |
"""Parse an X-Object-Sharing header from the request.
|
421 |
|
422 |
Raises BadRequest on error.
|
423 |
"""
|
424 |
|
425 |
permissions = request.META.get('HTTP_X_OBJECT_SHARING')
|
426 |
if permissions is None: |
427 |
return None |
428 |
|
429 |
# TODO: Document or remove '~' replacing.
|
430 |
permissions = permissions.replace('~', '') |
431 |
|
432 |
ret = {} |
433 |
permissions = permissions.replace(' ', '') |
434 |
if permissions == '': |
435 |
return ret
|
436 |
for perm in (x for x in permissions.split(';')): |
437 |
if perm.startswith('read='): |
438 |
ret['read'] = list(set([v.replace(' ','').lower() for v in perm[5:].split(',')])) |
439 |
if '' in ret['read']: |
440 |
ret['read'].remove('') |
441 |
if '*' in ret['read']: |
442 |
ret['read'] = ['*'] |
443 |
if len(ret['read']) == 0: |
444 |
raise BadRequest('Bad X-Object-Sharing header value') |
445 |
elif perm.startswith('write='): |
446 |
ret['write'] = list(set([v.replace(' ','').lower() for v in perm[6:].split(',')])) |
447 |
if '' in ret['write']: |
448 |
ret['write'].remove('') |
449 |
if '*' in ret['write']: |
450 |
ret['write'] = ['*'] |
451 |
if len(ret['write']) == 0: |
452 |
raise BadRequest('Bad X-Object-Sharing header value') |
453 |
else:
|
454 |
raise BadRequest('Bad X-Object-Sharing header value') |
455 |
|
456 |
# Keep duplicates only in write list.
|
457 |
dups = [x for x in ret.get('read', []) if x in ret.get('write', []) and x != '*'] |
458 |
if dups:
|
459 |
for x in dups: |
460 |
ret['read'].remove(x)
|
461 |
if len(ret['read']) == 0: |
462 |
del(ret['read']) |
463 |
|
464 |
return ret
|
465 |
|
466 |
def get_public(request): |
467 |
"""Parse an X-Object-Public header from the request.
|
468 |
|
469 |
Raises BadRequest on error.
|
470 |
"""
|
471 |
|
472 |
public = request.META.get('HTTP_X_OBJECT_PUBLIC')
|
473 |
if public is None: |
474 |
return None |
475 |
|
476 |
public = public.replace(' ', '').lower() |
477 |
if public == 'true': |
478 |
return True |
479 |
elif public == 'false' or public == '': |
480 |
return False |
481 |
raise BadRequest('Bad X-Object-Public header value') |
482 |
|
483 |
def raw_input_socket(request): |
484 |
"""Return the socket for reading the rest of the request."""
|
485 |
|
486 |
server_software = request.META.get('SERVER_SOFTWARE')
|
487 |
if server_software and server_software.startswith('mod_python'): |
488 |
return request._req
|
489 |
if 'wsgi.input' in request.environ: |
490 |
return request.environ['wsgi.input'] |
491 |
raise ServiceUnavailable('Unknown server software') |
492 |
|
493 |
MAX_UPLOAD_SIZE = 5 * (1024 * 1024 * 1024) # 5GB |
494 |
|
495 |
def socket_read_iterator(request, length=0, blocksize=4096): |
496 |
"""Return a maximum of blocksize data read from the socket in each iteration.
|
497 |
|
498 |
Read up to 'length'. If 'length' is negative, will attempt a chunked read.
|
499 |
The maximum ammount of data read is controlled by MAX_UPLOAD_SIZE.
|
500 |
"""
|
501 |
|
502 |
sock = raw_input_socket(request) |
503 |
if length < 0: # Chunked transfers |
504 |
# Small version (server does the dechunking).
|
505 |
if request.environ.get('mod_wsgi.input_chunked', None) or request.META['SERVER_SOFTWARE'].startswith('gunicorn'): |
506 |
while length < MAX_UPLOAD_SIZE:
|
507 |
data = sock.read(blocksize) |
508 |
if data == '': |
509 |
return
|
510 |
yield data
|
511 |
raise BadRequest('Maximum size is reached') |
512 |
|
513 |
# Long version (do the dechunking).
|
514 |
data = ''
|
515 |
while length < MAX_UPLOAD_SIZE:
|
516 |
# Get chunk size.
|
517 |
if hasattr(sock, 'readline'): |
518 |
chunk_length = sock.readline() |
519 |
else:
|
520 |
chunk_length = ''
|
521 |
while chunk_length[-1:] != '\n': |
522 |
chunk_length += sock.read(1)
|
523 |
chunk_length.strip() |
524 |
pos = chunk_length.find(';')
|
525 |
if pos >= 0: |
526 |
chunk_length = chunk_length[:pos] |
527 |
try:
|
528 |
chunk_length = int(chunk_length, 16) |
529 |
except Exception, e: |
530 |
raise BadRequest('Bad chunk size') # TODO: Change to something more appropriate. |
531 |
# Check if done.
|
532 |
if chunk_length == 0: |
533 |
if len(data) > 0: |
534 |
yield data
|
535 |
return
|
536 |
# Get the actual data.
|
537 |
while chunk_length > 0: |
538 |
chunk = sock.read(min(chunk_length, blocksize))
|
539 |
chunk_length -= len(chunk)
|
540 |
if length > 0: |
541 |
length += len(chunk)
|
542 |
data += chunk |
543 |
if len(data) >= blocksize: |
544 |
ret = data[:blocksize] |
545 |
data = data[blocksize:] |
546 |
yield ret
|
547 |
sock.read(2) # CRLF |
548 |
raise BadRequest('Maximum size is reached') |
549 |
else:
|
550 |
if length > MAX_UPLOAD_SIZE:
|
551 |
raise BadRequest('Maximum size is reached') |
552 |
while length > 0: |
553 |
data = sock.read(min(length, blocksize))
|
554 |
if not data: |
555 |
raise BadRequest()
|
556 |
length -= len(data)
|
557 |
yield data
|
558 |
|
559 |
class ObjectWrapper(object): |
560 |
"""Return the object's data block-per-block in each iteration.
|
561 |
|
562 |
Read from the object using the offset and length provided in each entry of the range list.
|
563 |
"""
|
564 |
|
565 |
def __init__(self, backend, ranges, sizes, hashmaps, boundary): |
566 |
self.backend = backend
|
567 |
self.ranges = ranges
|
568 |
self.sizes = sizes
|
569 |
self.hashmaps = hashmaps
|
570 |
self.boundary = boundary
|
571 |
self.size = sum(self.sizes) |
572 |
|
573 |
self.file_index = 0 |
574 |
self.block_index = 0 |
575 |
self.block_hash = -1 |
576 |
self.block = '' |
577 |
|
578 |
self.range_index = -1 |
579 |
self.offset, self.length = self.ranges[0] |
580 |
|
581 |
def __iter__(self): |
582 |
return self |
583 |
|
584 |
def part_iterator(self): |
585 |
if self.length > 0: |
586 |
# Get the file for the current offset.
|
587 |
file_size = self.sizes[self.file_index] |
588 |
while self.offset >= file_size: |
589 |
self.offset -= file_size
|
590 |
self.file_index += 1 |
591 |
file_size = self.sizes[self.file_index] |
592 |
|
593 |
# Get the block for the current position.
|
594 |
self.block_index = int(self.offset / self.backend.block_size) |
595 |
if self.block_hash != self.hashmaps[self.file_index][self.block_index]: |
596 |
self.block_hash = self.hashmaps[self.file_index][self.block_index] |
597 |
try:
|
598 |
self.block = self.backend.get_block(self.block_hash) |
599 |
except NameError: |
600 |
raise ItemNotFound('Block does not exist') |
601 |
|
602 |
# Get the data from the block.
|
603 |
bo = self.offset % self.backend.block_size |
604 |
bl = min(self.length, len(self.block) - bo) |
605 |
data = self.block[bo:bo + bl]
|
606 |
self.offset += bl
|
607 |
self.length -= bl
|
608 |
return data
|
609 |
else:
|
610 |
raise StopIteration |
611 |
|
612 |
def next(self): |
613 |
if len(self.ranges) == 1: |
614 |
return self.part_iterator() |
615 |
if self.range_index == len(self.ranges): |
616 |
raise StopIteration |
617 |
try:
|
618 |
if self.range_index == -1: |
619 |
raise StopIteration |
620 |
return self.part_iterator() |
621 |
except StopIteration: |
622 |
self.range_index += 1 |
623 |
out = [] |
624 |
if self.range_index < len(self.ranges): |
625 |
# Part header.
|
626 |
self.offset, self.length = self.ranges[self.range_index] |
627 |
self.file_index = 0 |
628 |
if self.range_index > 0: |
629 |
out.append('')
|
630 |
out.append('--' + self.boundary) |
631 |
out.append('Content-Range: bytes %d-%d/%d' % (self.offset, self.offset + self.length - 1, self.size)) |
632 |
out.append('Content-Transfer-Encoding: binary')
|
633 |
out.append('')
|
634 |
out.append('')
|
635 |
return '\r\n'.join(out) |
636 |
else:
|
637 |
# Footer.
|
638 |
out.append('')
|
639 |
out.append('--' + self.boundary + '--') |
640 |
out.append('')
|
641 |
return '\r\n'.join(out) |
642 |
|
643 |
def object_data_response(request, sizes, hashmaps, meta, public=False): |
644 |
"""Get the HttpResponse object for replying with the object's data."""
|
645 |
|
646 |
# Range handling.
|
647 |
size = sum(sizes)
|
648 |
ranges = get_range(request, size) |
649 |
if ranges is None: |
650 |
ranges = [(0, size)]
|
651 |
ret = 200
|
652 |
else:
|
653 |
check = [True for offset, length in ranges if |
654 |
length <= 0 or length > size or |
655 |
offset < 0 or offset >= size or |
656 |
offset + length > size] |
657 |
if len(check) > 0: |
658 |
raise RangeNotSatisfiable('Requested range exceeds object limits') |
659 |
ret = 206
|
660 |
if_range = request.META.get('HTTP_IF_RANGE')
|
661 |
if if_range:
|
662 |
try:
|
663 |
# Modification time has passed instead.
|
664 |
last_modified = parse_http_date(if_range) |
665 |
if last_modified != meta['modified']: |
666 |
ranges = [(0, size)]
|
667 |
ret = 200
|
668 |
except ValueError: |
669 |
if if_range != meta['ETag']: |
670 |
ranges = [(0, size)]
|
671 |
ret = 200
|
672 |
|
673 |
if ret == 206 and len(ranges) > 1: |
674 |
boundary = uuid.uuid4().hex |
675 |
else:
|
676 |
boundary = ''
|
677 |
wrapper = ObjectWrapper(request.backend, ranges, sizes, hashmaps, boundary) |
678 |
response = HttpResponse(wrapper, status=ret) |
679 |
put_object_headers(response, meta, public) |
680 |
if ret == 206: |
681 |
if len(ranges) == 1: |
682 |
offset, length = ranges[0]
|
683 |
response['Content-Length'] = length # Update with the correct length. |
684 |
response['Content-Range'] = 'bytes %d-%d/%d' % (offset, offset + length - 1, size) |
685 |
else:
|
686 |
del(response['Content-Length']) |
687 |
response['Content-Type'] = 'multipart/byteranges; boundary=%s' % (boundary,) |
688 |
return response
|
689 |
|
690 |
def put_object_block(request, hashmap, data, offset): |
691 |
"""Put one block of data at the given offset."""
|
692 |
|
693 |
bi = int(offset / request.backend.block_size)
|
694 |
bo = offset % request.backend.block_size |
695 |
bl = min(len(data), request.backend.block_size - bo) |
696 |
if bi < len(hashmap): |
697 |
hashmap[bi] = request.backend.update_block(hashmap[bi], data[:bl], bo) |
698 |
else:
|
699 |
hashmap.append(request.backend.put_block(('\x00' * bo) + data[:bl]))
|
700 |
return bl # Return ammount of data written. |
701 |
|
702 |
def hashmap_hash(request, hashmap): |
703 |
"""Produce the root hash, treating the hashmap as a Merkle-like tree."""
|
704 |
|
705 |
def subhash(d): |
706 |
h = hashlib.new(request.backend.hash_algorithm) |
707 |
h.update(d) |
708 |
return h.digest()
|
709 |
|
710 |
if len(hashmap) == 0: |
711 |
return hexlify(subhash('')) |
712 |
if len(hashmap) == 1: |
713 |
return hashmap[0] |
714 |
|
715 |
s = 2
|
716 |
while s < len(hashmap): |
717 |
s = s * 2
|
718 |
h = [unhexlify(x) for x in hashmap] |
719 |
h += [('\x00' * len(h[0]))] * (s - len(hashmap)) |
720 |
while len(h) > 1: |
721 |
h = [subhash(h[x] + h[x + 1]) for x in range(0, len(h), 2)] |
722 |
return hexlify(h[0]) |
723 |
|
724 |
def update_response_headers(request, response): |
725 |
if request.serialization == 'xml': |
726 |
response['Content-Type'] = 'application/xml; charset=UTF-8' |
727 |
elif request.serialization == 'json': |
728 |
response['Content-Type'] = 'application/json; charset=UTF-8' |
729 |
elif not response['Content-Type']: |
730 |
response['Content-Type'] = 'text/plain; charset=UTF-8' |
731 |
|
732 |
if not response.has_header('Content-Length') and not (response.has_header('Content-Type') and response['Content-Type'].startswith('multipart/byteranges')): |
733 |
response['Content-Length'] = len(response.content) |
734 |
|
735 |
if settings.TEST:
|
736 |
response['Date'] = format_date_time(time())
|
737 |
|
738 |
def render_fault(request, fault): |
739 |
if settings.DEBUG or settings.TEST: |
740 |
fault.details = format_exc(fault) |
741 |
|
742 |
request.serialization = 'text'
|
743 |
data = '\n'.join((fault.message, fault.details)) + '\n' |
744 |
response = HttpResponse(data, status=fault.code) |
745 |
update_response_headers(request, response) |
746 |
return response
|
747 |
|
748 |
def request_serialization(request, format_allowed=False): |
749 |
"""Return the serialization format requested.
|
750 |
|
751 |
Valid formats are 'text' and 'json', 'xml' if 'format_allowed' is True.
|
752 |
"""
|
753 |
|
754 |
if not format_allowed: |
755 |
return 'text' |
756 |
|
757 |
format = request.GET.get('format')
|
758 |
if format == 'json': |
759 |
return 'json' |
760 |
elif format == 'xml': |
761 |
return 'xml' |
762 |
|
763 |
for item in request.META.get('HTTP_ACCEPT', '').split(','): |
764 |
accept, sep, rest = item.strip().partition(';')
|
765 |
if accept == 'application/json': |
766 |
return 'json' |
767 |
elif accept == 'application/xml' or accept == 'text/xml': |
768 |
return 'xml' |
769 |
|
770 |
return 'text' |
771 |
|
772 |
def api_method(http_method=None, format_allowed=False, user_required=True): |
773 |
"""Decorator function for views that implement an API method."""
|
774 |
|
775 |
def decorator(func): |
776 |
@wraps(func)
|
777 |
def wrapper(request, *args, **kwargs): |
778 |
try:
|
779 |
if http_method and request.method != http_method: |
780 |
raise BadRequest('Method not allowed.') |
781 |
if user_required and getattr(request, 'user', None) is None: |
782 |
raise Unauthorized('Access denied') |
783 |
|
784 |
# The args variable may contain up to (account, container, object).
|
785 |
if len(args) > 1 and len(args[1]) > 256: |
786 |
raise BadRequest('Container name too large.') |
787 |
if len(args) > 2 and len(args[2]) > 1024: |
788 |
raise BadRequest('Object name too large.') |
789 |
|
790 |
# Fill in custom request variables.
|
791 |
request.serialization = request_serialization(request, format_allowed) |
792 |
request.backend = connect_backend() |
793 |
|
794 |
response = func(request, *args, **kwargs) |
795 |
update_response_headers(request, response) |
796 |
return response
|
797 |
except Fault, fault:
|
798 |
return render_fault(request, fault)
|
799 |
except BaseException, e: |
800 |
logger.exception('Unexpected error: %s' % e)
|
801 |
fault = ServiceUnavailable('Unexpected error')
|
802 |
return render_fault(request, fault)
|
803 |
finally:
|
804 |
if getattr(request, 'backend', None) is not None: |
805 |
request.backend.close() |
806 |
return wrapper
|
807 |
return decorator
|