root / pithos / api / util.py @ 53cff70c
History | View | Annotate | Download (34 kB)
1 |
# Copyright 2011-2012 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 |
from urllib import quote, unquote |
41 |
|
42 |
from django.conf import settings |
43 |
from django.http import HttpResponse |
44 |
from django.utils import simplejson as json |
45 |
from django.utils.http import http_date, parse_etags |
46 |
from django.utils.encoding import smart_unicode, smart_str |
47 |
from django.core.files.uploadhandler import FileUploadHandler |
48 |
from django.core.files.uploadedfile import UploadedFile |
49 |
|
50 |
from pithos.lib.compat import parse_http_date_safe, parse_http_date |
51 |
|
52 |
from pithos.api.faults import (Fault, NotModified, BadRequest, Unauthorized, Forbidden, ItemNotFound, |
53 |
Conflict, LengthRequired, PreconditionFailed, RequestEntityTooLarge, |
54 |
RangeNotSatisfiable, InternalServerError, NotImplemented)
|
55 |
from pithos.api.short_url import encode_url |
56 |
from pithos.backends import connect_backend |
57 |
from pithos.backends.base import NotAllowedError, QuotaError |
58 |
|
59 |
import logging |
60 |
import re |
61 |
import hashlib |
62 |
import uuid |
63 |
import decimal |
64 |
|
65 |
|
66 |
logger = logging.getLogger(__name__) |
67 |
|
68 |
|
69 |
class UTC(tzinfo): |
70 |
def utcoffset(self, dt): |
71 |
return timedelta(0) |
72 |
|
73 |
def tzname(self, dt): |
74 |
return 'UTC' |
75 |
|
76 |
def dst(self, dt): |
77 |
return timedelta(0) |
78 |
|
79 |
def json_encode_decimal(obj): |
80 |
if isinstance(obj, decimal.Decimal): |
81 |
return str(obj) |
82 |
raise TypeError(repr(obj) + " is not JSON serializable") |
83 |
|
84 |
def isoformat(d): |
85 |
"""Return an ISO8601 date string that includes a timezone."""
|
86 |
|
87 |
return d.replace(tzinfo=UTC()).isoformat()
|
88 |
|
89 |
def rename_meta_key(d, old, new): |
90 |
if old not in d: |
91 |
return
|
92 |
d[new] = d[old] |
93 |
del(d[old])
|
94 |
|
95 |
def printable_header_dict(d): |
96 |
"""Format a meta dictionary for printing out json/xml.
|
97 |
|
98 |
Convert all keys to lower case and replace dashes with underscores.
|
99 |
Format 'last_modified' timestamp.
|
100 |
"""
|
101 |
|
102 |
if 'last_modified' in d: |
103 |
d['last_modified'] = isoformat(datetime.fromtimestamp(d['last_modified'])) |
104 |
return dict([(k.lower().replace('-', '_'), v) for k, v in d.iteritems()]) |
105 |
|
106 |
def format_header_key(k): |
107 |
"""Convert underscores to dashes and capitalize intra-dash strings."""
|
108 |
return '-'.join([x.capitalize() for x in k.replace('_', '-').split('-')]) |
109 |
|
110 |
def get_header_prefix(request, prefix): |
111 |
"""Get all prefix-* request headers in a dict. Reformat keys with format_header_key()."""
|
112 |
|
113 |
prefix = 'HTTP_' + prefix.upper().replace('-', '_') |
114 |
# TODO: Document or remove '~' replacing.
|
115 |
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)]) |
116 |
|
117 |
def get_account_headers(request): |
118 |
meta = get_header_prefix(request, 'X-Account-Meta-')
|
119 |
groups = {} |
120 |
for k, v in get_header_prefix(request, 'X-Account-Group-').iteritems(): |
121 |
n = k[16:].lower()
|
122 |
if '-' in n or '_' in n: |
123 |
raise BadRequest('Bad characters in group name') |
124 |
groups[n] = v.replace(' ', '').split(',') |
125 |
while '' in groups[n]: |
126 |
groups[n].remove('')
|
127 |
return meta, groups
|
128 |
|
129 |
def put_account_headers(response, meta, groups, policy): |
130 |
if 'count' in meta: |
131 |
response['X-Account-Container-Count'] = meta['count'] |
132 |
if 'bytes' in meta: |
133 |
response['X-Account-Bytes-Used'] = meta['bytes'] |
134 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
135 |
for k in [x for x in meta.keys() if x.startswith('X-Account-Meta-')]: |
136 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
137 |
if 'until_timestamp' in meta: |
138 |
response['X-Account-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) |
139 |
for k, v in groups.iteritems(): |
140 |
k = smart_str(k, strings_only=True)
|
141 |
k = format_header_key('X-Account-Group-' + k)
|
142 |
v = smart_str(','.join(v), strings_only=True) |
143 |
response[k] = v |
144 |
for k, v in policy.iteritems(): |
145 |
response[smart_str(format_header_key('X-Account-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) |
146 |
|
147 |
def get_container_headers(request): |
148 |
meta = get_header_prefix(request, 'X-Container-Meta-')
|
149 |
policy = dict([(k[19:].lower(), v.replace(' ', '')) for k, v in get_header_prefix(request, 'X-Container-Policy-').iteritems()]) |
150 |
return meta, policy
|
151 |
|
152 |
def put_container_headers(request, response, meta, policy): |
153 |
if 'count' in meta: |
154 |
response['X-Container-Object-Count'] = meta['count'] |
155 |
if 'bytes' in meta: |
156 |
response['X-Container-Bytes-Used'] = meta['bytes'] |
157 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
158 |
for k in [x for x in meta.keys() if x.startswith('X-Container-Meta-')]: |
159 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
160 |
l = [smart_str(x, strings_only=True) for x in meta['object_meta'] if x.startswith('X-Object-Meta-')] |
161 |
response['X-Container-Object-Meta'] = ','.join([x[14:] for x in l]) |
162 |
response['X-Container-Block-Size'] = request.backend.block_size
|
163 |
response['X-Container-Block-Hash'] = request.backend.hash_algorithm
|
164 |
if 'until_timestamp' in meta: |
165 |
response['X-Container-Until-Timestamp'] = http_date(int(meta['until_timestamp'])) |
166 |
for k, v in policy.iteritems(): |
167 |
response[smart_str(format_header_key('X-Container-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True) |
168 |
|
169 |
def get_object_headers(request): |
170 |
meta = get_header_prefix(request, 'X-Object-Meta-')
|
171 |
if request.META.get('CONTENT_TYPE'): |
172 |
meta['Content-Type'] = request.META['CONTENT_TYPE'] |
173 |
if request.META.get('HTTP_CONTENT_ENCODING'): |
174 |
meta['Content-Encoding'] = request.META['HTTP_CONTENT_ENCODING'] |
175 |
if request.META.get('HTTP_CONTENT_DISPOSITION'): |
176 |
meta['Content-Disposition'] = request.META['HTTP_CONTENT_DISPOSITION'] |
177 |
if request.META.get('HTTP_X_OBJECT_MANIFEST'): |
178 |
meta['X-Object-Manifest'] = request.META['HTTP_X_OBJECT_MANIFEST'] |
179 |
return meta, get_sharing(request), get_public(request)
|
180 |
|
181 |
def put_object_headers(response, meta, restricted=False): |
182 |
if 'ETag' in meta: |
183 |
response['ETag'] = meta['ETag'] |
184 |
response['Content-Length'] = meta['bytes'] |
185 |
response['Content-Type'] = meta.get('Content-Type', 'application/octet-stream') |
186 |
response['Last-Modified'] = http_date(int(meta['modified'])) |
187 |
if not restricted: |
188 |
response['X-Object-Hash'] = meta['hash'] |
189 |
response['X-Object-UUID'] = meta['uuid'] |
190 |
response['X-Object-Modified-By'] = smart_str(meta['modified_by'], strings_only=True) |
191 |
response['X-Object-Version'] = meta['version'] |
192 |
response['X-Object-Version-Timestamp'] = http_date(int(meta['version_timestamp'])) |
193 |
for k in [x for x in meta.keys() if x.startswith('X-Object-Meta-')]: |
194 |
response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True) |
195 |
for k in ('Content-Encoding', 'Content-Disposition', 'X-Object-Manifest', |
196 |
'X-Object-Sharing', 'X-Object-Shared-By', 'X-Object-Allowed-To', |
197 |
'X-Object-Public'):
|
198 |
if k in meta: |
199 |
response[k] = smart_str(meta[k], strings_only=True)
|
200 |
else:
|
201 |
for k in ('Content-Encoding', 'Content-Disposition'): |
202 |
if k in meta: |
203 |
response[k] = smart_str(meta[k], strings_only=True)
|
204 |
|
205 |
def update_manifest_meta(request, v_account, meta): |
206 |
"""Update metadata if the object has an X-Object-Manifest."""
|
207 |
|
208 |
if 'X-Object-Manifest' in meta: |
209 |
etag = ''
|
210 |
bytes = 0
|
211 |
try:
|
212 |
src_container, src_name = split_container_object_string('/' + meta['X-Object-Manifest']) |
213 |
objects = request.backend.list_objects(request.user_uniq, v_account, |
214 |
src_container, prefix=src_name, virtual=False)
|
215 |
for x in objects: |
216 |
src_meta = request.backend.get_object_meta(request.user_uniq, |
217 |
v_account, src_container, x[0], 'pithos', x[1]) |
218 |
if 'ETag' in src_meta: |
219 |
etag += src_meta['ETag']
|
220 |
bytes += src_meta['bytes'] |
221 |
except:
|
222 |
# Ignore errors.
|
223 |
return
|
224 |
meta['bytes'] = bytes |
225 |
md5 = hashlib.md5() |
226 |
md5.update(etag) |
227 |
meta['ETag'] = md5.hexdigest().lower()
|
228 |
|
229 |
def update_sharing_meta(request, permissions, v_account, v_container, v_object, meta): |
230 |
if permissions is None: |
231 |
return
|
232 |
allowed, perm_path, perms = permissions |
233 |
if len(perms) == 0: |
234 |
return
|
235 |
ret = [] |
236 |
r = ','.join(perms.get('read', [])) |
237 |
if r:
|
238 |
ret.append('read=' + r)
|
239 |
w = ','.join(perms.get('write', [])) |
240 |
if w:
|
241 |
ret.append('write=' + w)
|
242 |
meta['X-Object-Sharing'] = '; '.join(ret) |
243 |
if '/'.join((v_account, v_container, v_object)) != perm_path: |
244 |
meta['X-Object-Shared-By'] = perm_path
|
245 |
if request.user_uniq != v_account:
|
246 |
meta['X-Object-Allowed-To'] = allowed
|
247 |
|
248 |
def update_public_meta(public, meta): |
249 |
if not public: |
250 |
return
|
251 |
meta['X-Object-Public'] = '/public/' + encode_url(public) |
252 |
|
253 |
def validate_modification_preconditions(request, meta): |
254 |
"""Check that the modified timestamp conforms with the preconditions set."""
|
255 |
|
256 |
if 'modified' not in meta: |
257 |
return # TODO: Always return? |
258 |
|
259 |
if_modified_since = request.META.get('HTTP_IF_MODIFIED_SINCE')
|
260 |
if if_modified_since is not None: |
261 |
if_modified_since = parse_http_date_safe(if_modified_since) |
262 |
if if_modified_since is not None and int(meta['modified']) <= if_modified_since: |
263 |
raise NotModified('Resource has not been modified') |
264 |
|
265 |
if_unmodified_since = request.META.get('HTTP_IF_UNMODIFIED_SINCE')
|
266 |
if if_unmodified_since is not None: |
267 |
if_unmodified_since = parse_http_date_safe(if_unmodified_since) |
268 |
if if_unmodified_since is not None and int(meta['modified']) > if_unmodified_since: |
269 |
raise PreconditionFailed('Resource has been modified') |
270 |
|
271 |
def validate_matching_preconditions(request, meta): |
272 |
"""Check that the ETag conforms with the preconditions set."""
|
273 |
|
274 |
etag = meta.get('ETag', None) |
275 |
|
276 |
if_match = request.META.get('HTTP_IF_MATCH')
|
277 |
if if_match is not None: |
278 |
if etag is None: |
279 |
raise PreconditionFailed('Resource does not exist') |
280 |
if if_match != '*' and etag not in [x.lower() for x in parse_etags(if_match)]: |
281 |
raise PreconditionFailed('Resource ETag does not match') |
282 |
|
283 |
if_none_match = request.META.get('HTTP_IF_NONE_MATCH')
|
284 |
if if_none_match is not None: |
285 |
# TODO: If this passes, must ignore If-Modified-Since header.
|
286 |
if etag is not None: |
287 |
if if_none_match == '*' or etag in [x.lower() for x in parse_etags(if_none_match)]: |
288 |
# TODO: Continue if an If-Modified-Since header is present.
|
289 |
if request.method in ('HEAD', 'GET'): |
290 |
raise NotModified('Resource ETag matches') |
291 |
raise PreconditionFailed('Resource exists or ETag matches') |
292 |
|
293 |
def split_container_object_string(s): |
294 |
if not len(s) > 0 or s[0] != '/': |
295 |
raise ValueError |
296 |
s = s[1:]
|
297 |
pos = s.find('/')
|
298 |
if pos == -1 or pos == len(s) - 1: |
299 |
raise ValueError |
300 |
return s[:pos], s[(pos + 1):] |
301 |
|
302 |
def copy_or_move_object(request, src_account, src_container, src_name, dest_account, dest_container, dest_name, move=False): |
303 |
"""Copy or move an object."""
|
304 |
|
305 |
if 'ignore_content_type' in request.GET and 'CONTENT_TYPE' in request.META: |
306 |
del(request.META['CONTENT_TYPE']) |
307 |
meta, permissions, public = get_object_headers(request) |
308 |
src_version = request.META.get('HTTP_X_SOURCE_VERSION')
|
309 |
try:
|
310 |
if move:
|
311 |
version_id = request.backend.move_object(request.user_uniq, src_account, src_container, src_name, |
312 |
dest_account, dest_container, dest_name, |
313 |
'pithos', meta, False, permissions) |
314 |
else:
|
315 |
version_id = request.backend.copy_object(request.user_uniq, src_account, src_container, src_name, |
316 |
dest_account, dest_container, dest_name, |
317 |
'pithos', meta, False, permissions, src_version) |
318 |
except NotAllowedError:
|
319 |
raise Forbidden('Not allowed') |
320 |
except (NameError, IndexError): |
321 |
raise ItemNotFound('Container or object does not exist') |
322 |
except ValueError: |
323 |
raise BadRequest('Invalid sharing header') |
324 |
except AttributeError, e: |
325 |
raise Conflict('\n'.join(e.data) + '\n') |
326 |
except QuotaError:
|
327 |
raise RequestEntityTooLarge('Quota exceeded') |
328 |
if public is not None: |
329 |
try:
|
330 |
request.backend.update_object_public(request.user_uniq, dest_account, dest_container, dest_name, public) |
331 |
except NotAllowedError:
|
332 |
raise Forbidden('Not allowed') |
333 |
except NameError: |
334 |
raise ItemNotFound('Object does not exist') |
335 |
return version_id
|
336 |
|
337 |
def get_int_parameter(p): |
338 |
if p is not None: |
339 |
try:
|
340 |
p = int(p)
|
341 |
except ValueError: |
342 |
return None |
343 |
if p < 0: |
344 |
return None |
345 |
return p
|
346 |
|
347 |
def get_content_length(request): |
348 |
content_length = get_int_parameter(request.META.get('CONTENT_LENGTH'))
|
349 |
if content_length is None: |
350 |
raise LengthRequired('Missing or invalid Content-Length header') |
351 |
return content_length
|
352 |
|
353 |
def get_range(request, size): |
354 |
"""Parse a Range header from the request.
|
355 |
|
356 |
Either returns None, when the header is not existent or should be ignored,
|
357 |
or a list of (offset, length) tuples - should be further checked.
|
358 |
"""
|
359 |
|
360 |
ranges = request.META.get('HTTP_RANGE', '').replace(' ', '') |
361 |
if not ranges.startswith('bytes='): |
362 |
return None |
363 |
|
364 |
ret = [] |
365 |
for r in (x.strip() for x in ranges[6:].split(',')): |
366 |
p = re.compile('^(?P<offset>\d*)-(?P<upto>\d*)$')
|
367 |
m = p.match(r) |
368 |
if not m: |
369 |
return None |
370 |
offset = m.group('offset')
|
371 |
upto = m.group('upto')
|
372 |
if offset == '' and upto == '': |
373 |
return None |
374 |
|
375 |
if offset != '': |
376 |
offset = int(offset)
|
377 |
if upto != '': |
378 |
upto = int(upto)
|
379 |
if offset > upto:
|
380 |
return None |
381 |
ret.append((offset, upto - offset + 1))
|
382 |
else:
|
383 |
ret.append((offset, size - offset)) |
384 |
else:
|
385 |
length = int(upto)
|
386 |
ret.append((size - length, length)) |
387 |
|
388 |
return ret
|
389 |
|
390 |
def get_content_range(request): |
391 |
"""Parse a Content-Range header from the request.
|
392 |
|
393 |
Either returns None, when the header is not existent or should be ignored,
|
394 |
or an (offset, length, total) tuple - check as length, total may be None.
|
395 |
Returns (None, None, None) if the provided range is '*/*'.
|
396 |
"""
|
397 |
|
398 |
ranges = request.META.get('HTTP_CONTENT_RANGE', '') |
399 |
if not ranges: |
400 |
return None |
401 |
|
402 |
p = re.compile('^bytes (?P<offset>\d+)-(?P<upto>\d*)/(?P<total>(\d+|\*))$')
|
403 |
m = p.match(ranges) |
404 |
if not m: |
405 |
if ranges == 'bytes */*': |
406 |
return (None, None, None) |
407 |
return None |
408 |
offset = int(m.group('offset')) |
409 |
upto = m.group('upto')
|
410 |
total = m.group('total')
|
411 |
if upto != '': |
412 |
upto = int(upto)
|
413 |
else:
|
414 |
upto = None
|
415 |
if total != '*': |
416 |
total = int(total)
|
417 |
else:
|
418 |
total = None
|
419 |
if (upto is not None and offset > upto) or \ |
420 |
(total is not None and offset >= total) or \ |
421 |
(total is not None and upto is not None and upto >= total): |
422 |
return None |
423 |
|
424 |
if upto is None: |
425 |
length = None
|
426 |
else:
|
427 |
length = upto - offset + 1
|
428 |
return (offset, length, total)
|
429 |
|
430 |
def get_sharing(request): |
431 |
"""Parse an X-Object-Sharing header from the request.
|
432 |
|
433 |
Raises BadRequest on error.
|
434 |
"""
|
435 |
|
436 |
permissions = request.META.get('HTTP_X_OBJECT_SHARING')
|
437 |
if permissions is None: |
438 |
return None |
439 |
|
440 |
# TODO: Document or remove '~' replacing.
|
441 |
permissions = permissions.replace('~', '') |
442 |
|
443 |
ret = {} |
444 |
permissions = permissions.replace(' ', '') |
445 |
if permissions == '': |
446 |
return ret
|
447 |
for perm in (x for x in permissions.split(';')): |
448 |
if perm.startswith('read='): |
449 |
ret['read'] = list(set([v.replace(' ','').lower() for v in perm[5:].split(',')])) |
450 |
if '' in ret['read']: |
451 |
ret['read'].remove('') |
452 |
if '*' in ret['read']: |
453 |
ret['read'] = ['*'] |
454 |
if len(ret['read']) == 0: |
455 |
raise BadRequest('Bad X-Object-Sharing header value') |
456 |
elif perm.startswith('write='): |
457 |
ret['write'] = list(set([v.replace(' ','').lower() for v in perm[6:].split(',')])) |
458 |
if '' in ret['write']: |
459 |
ret['write'].remove('') |
460 |
if '*' in ret['write']: |
461 |
ret['write'] = ['*'] |
462 |
if len(ret['write']) == 0: |
463 |
raise BadRequest('Bad X-Object-Sharing header value') |
464 |
else:
|
465 |
raise BadRequest('Bad X-Object-Sharing header value') |
466 |
|
467 |
# Keep duplicates only in write list.
|
468 |
dups = [x for x in ret.get('read', []) if x in ret.get('write', []) and x != '*'] |
469 |
if dups:
|
470 |
for x in dups: |
471 |
ret['read'].remove(x)
|
472 |
if len(ret['read']) == 0: |
473 |
del(ret['read']) |
474 |
|
475 |
return ret
|
476 |
|
477 |
def get_public(request): |
478 |
"""Parse an X-Object-Public header from the request.
|
479 |
|
480 |
Raises BadRequest on error.
|
481 |
"""
|
482 |
|
483 |
public = request.META.get('HTTP_X_OBJECT_PUBLIC')
|
484 |
if public is None: |
485 |
return None |
486 |
|
487 |
public = public.replace(' ', '').lower() |
488 |
if public == 'true': |
489 |
return True |
490 |
elif public == 'false' or public == '': |
491 |
return False |
492 |
raise BadRequest('Bad X-Object-Public header value') |
493 |
|
494 |
def raw_input_socket(request): |
495 |
"""Return the socket for reading the rest of the request."""
|
496 |
|
497 |
server_software = request.META.get('SERVER_SOFTWARE')
|
498 |
if server_software and server_software.startswith('mod_python'): |
499 |
return request._req
|
500 |
if 'wsgi.input' in request.environ: |
501 |
return request.environ['wsgi.input'] |
502 |
raise NotImplemented('Unknown server software') |
503 |
|
504 |
MAX_UPLOAD_SIZE = 5 * (1024 * 1024 * 1024) # 5GB |
505 |
|
506 |
def socket_read_iterator(request, length=0, blocksize=4096): |
507 |
"""Return a maximum of blocksize data read from the socket in each iteration.
|
508 |
|
509 |
Read up to 'length'. If 'length' is negative, will attempt a chunked read.
|
510 |
The maximum ammount of data read is controlled by MAX_UPLOAD_SIZE.
|
511 |
"""
|
512 |
|
513 |
sock = raw_input_socket(request) |
514 |
if length < 0: # Chunked transfers |
515 |
# Small version (server does the dechunking).
|
516 |
if request.environ.get('mod_wsgi.input_chunked', None) or request.META['SERVER_SOFTWARE'].startswith('gunicorn'): |
517 |
while length < MAX_UPLOAD_SIZE:
|
518 |
data = sock.read(blocksize) |
519 |
if data == '': |
520 |
return
|
521 |
yield data
|
522 |
raise BadRequest('Maximum size is reached') |
523 |
|
524 |
# Long version (do the dechunking).
|
525 |
data = ''
|
526 |
while length < MAX_UPLOAD_SIZE:
|
527 |
# Get chunk size.
|
528 |
if hasattr(sock, 'readline'): |
529 |
chunk_length = sock.readline() |
530 |
else:
|
531 |
chunk_length = ''
|
532 |
while chunk_length[-1:] != '\n': |
533 |
chunk_length += sock.read(1)
|
534 |
chunk_length.strip() |
535 |
pos = chunk_length.find(';')
|
536 |
if pos >= 0: |
537 |
chunk_length = chunk_length[:pos] |
538 |
try:
|
539 |
chunk_length = int(chunk_length, 16) |
540 |
except Exception, e: |
541 |
raise BadRequest('Bad chunk size') # TODO: Change to something more appropriate. |
542 |
# Check if done.
|
543 |
if chunk_length == 0: |
544 |
if len(data) > 0: |
545 |
yield data
|
546 |
return
|
547 |
# Get the actual data.
|
548 |
while chunk_length > 0: |
549 |
chunk = sock.read(min(chunk_length, blocksize))
|
550 |
chunk_length -= len(chunk)
|
551 |
if length > 0: |
552 |
length += len(chunk)
|
553 |
data += chunk |
554 |
if len(data) >= blocksize: |
555 |
ret = data[:blocksize] |
556 |
data = data[blocksize:] |
557 |
yield ret
|
558 |
sock.read(2) # CRLF |
559 |
raise BadRequest('Maximum size is reached') |
560 |
else:
|
561 |
if length > MAX_UPLOAD_SIZE:
|
562 |
raise BadRequest('Maximum size is reached') |
563 |
while length > 0: |
564 |
data = sock.read(min(length, blocksize))
|
565 |
if not data: |
566 |
raise BadRequest()
|
567 |
length -= len(data)
|
568 |
yield data
|
569 |
|
570 |
class SaveToBackendHandler(FileUploadHandler): |
571 |
"""Handle a file from an HTML form the django way."""
|
572 |
|
573 |
def __init__(self, request=None): |
574 |
super(SaveToBackendHandler, self).__init__(request) |
575 |
self.backend = request.backend
|
576 |
|
577 |
def put_data(self, length): |
578 |
if len(self.data) >= length: |
579 |
block = self.data[:length]
|
580 |
self.file.hashmap.append(self.backend.put_block(block)) |
581 |
self.md5.update(block)
|
582 |
self.data = self.data[length:] |
583 |
|
584 |
def new_file(self, field_name, file_name, content_type, content_length, charset=None): |
585 |
self.md5 = hashlib.md5()
|
586 |
self.data = '' |
587 |
self.file = UploadedFile(name=file_name, content_type=content_type, charset=charset)
|
588 |
self.file.size = 0 |
589 |
self.file.hashmap = []
|
590 |
|
591 |
def receive_data_chunk(self, raw_data, start): |
592 |
self.data += raw_data
|
593 |
self.file.size += len(raw_data) |
594 |
self.put_data(self.request.backend.block_size) |
595 |
return None |
596 |
|
597 |
def file_complete(self, file_size): |
598 |
l = len(self.data) |
599 |
if l > 0: |
600 |
self.put_data(l)
|
601 |
self.file.etag = self.md5.hexdigest().lower() |
602 |
return self.file |
603 |
|
604 |
class ObjectWrapper(object): |
605 |
"""Return the object's data block-per-block in each iteration.
|
606 |
|
607 |
Read from the object using the offset and length provided in each entry of the range list.
|
608 |
"""
|
609 |
|
610 |
def __init__(self, backend, ranges, sizes, hashmaps, boundary): |
611 |
self.backend = backend
|
612 |
self.ranges = ranges
|
613 |
self.sizes = sizes
|
614 |
self.hashmaps = hashmaps
|
615 |
self.boundary = boundary
|
616 |
self.size = sum(self.sizes) |
617 |
|
618 |
self.file_index = 0 |
619 |
self.block_index = 0 |
620 |
self.block_hash = -1 |
621 |
self.block = '' |
622 |
|
623 |
self.range_index = -1 |
624 |
self.offset, self.length = self.ranges[0] |
625 |
|
626 |
def __iter__(self): |
627 |
return self |
628 |
|
629 |
def part_iterator(self): |
630 |
if self.length > 0: |
631 |
# Get the file for the current offset.
|
632 |
file_size = self.sizes[self.file_index] |
633 |
while self.offset >= file_size: |
634 |
self.offset -= file_size
|
635 |
self.file_index += 1 |
636 |
file_size = self.sizes[self.file_index] |
637 |
|
638 |
# Get the block for the current position.
|
639 |
self.block_index = int(self.offset / self.backend.block_size) |
640 |
if self.block_hash != self.hashmaps[self.file_index][self.block_index]: |
641 |
self.block_hash = self.hashmaps[self.file_index][self.block_index] |
642 |
try:
|
643 |
self.block = self.backend.get_block(self.block_hash) |
644 |
except NameError: |
645 |
raise ItemNotFound('Block does not exist') |
646 |
|
647 |
# Get the data from the block.
|
648 |
bo = self.offset % self.backend.block_size |
649 |
bl = min(self.length, len(self.block) - bo) |
650 |
data = self.block[bo:bo + bl]
|
651 |
self.offset += bl
|
652 |
self.length -= bl
|
653 |
return data
|
654 |
else:
|
655 |
raise StopIteration |
656 |
|
657 |
def next(self): |
658 |
if len(self.ranges) == 1: |
659 |
return self.part_iterator() |
660 |
if self.range_index == len(self.ranges): |
661 |
raise StopIteration |
662 |
try:
|
663 |
if self.range_index == -1: |
664 |
raise StopIteration |
665 |
return self.part_iterator() |
666 |
except StopIteration: |
667 |
self.range_index += 1 |
668 |
out = [] |
669 |
if self.range_index < len(self.ranges): |
670 |
# Part header.
|
671 |
self.offset, self.length = self.ranges[self.range_index] |
672 |
self.file_index = 0 |
673 |
if self.range_index > 0: |
674 |
out.append('')
|
675 |
out.append('--' + self.boundary) |
676 |
out.append('Content-Range: bytes %d-%d/%d' % (self.offset, self.offset + self.length - 1, self.size)) |
677 |
out.append('Content-Transfer-Encoding: binary')
|
678 |
out.append('')
|
679 |
out.append('')
|
680 |
return '\r\n'.join(out) |
681 |
else:
|
682 |
# Footer.
|
683 |
out.append('')
|
684 |
out.append('--' + self.boundary + '--') |
685 |
out.append('')
|
686 |
return '\r\n'.join(out) |
687 |
|
688 |
def object_data_response(request, sizes, hashmaps, meta, public=False): |
689 |
"""Get the HttpResponse object for replying with the object's data."""
|
690 |
|
691 |
# Range handling.
|
692 |
size = sum(sizes)
|
693 |
ranges = get_range(request, size) |
694 |
if ranges is None: |
695 |
ranges = [(0, size)]
|
696 |
ret = 200
|
697 |
else:
|
698 |
check = [True for offset, length in ranges if |
699 |
length <= 0 or length > size or |
700 |
offset < 0 or offset >= size or |
701 |
offset + length > size] |
702 |
if len(check) > 0: |
703 |
raise RangeNotSatisfiable('Requested range exceeds object limits') |
704 |
ret = 206
|
705 |
if_range = request.META.get('HTTP_IF_RANGE')
|
706 |
if if_range:
|
707 |
try:
|
708 |
# Modification time has passed instead.
|
709 |
last_modified = parse_http_date(if_range) |
710 |
if last_modified != meta['modified']: |
711 |
ranges = [(0, size)]
|
712 |
ret = 200
|
713 |
except ValueError: |
714 |
if if_range != meta['ETag']: |
715 |
ranges = [(0, size)]
|
716 |
ret = 200
|
717 |
|
718 |
if ret == 206 and len(ranges) > 1: |
719 |
boundary = uuid.uuid4().hex |
720 |
else:
|
721 |
boundary = ''
|
722 |
wrapper = ObjectWrapper(request.backend, ranges, sizes, hashmaps, boundary) |
723 |
response = HttpResponse(wrapper, status=ret) |
724 |
put_object_headers(response, meta, public) |
725 |
if ret == 206: |
726 |
if len(ranges) == 1: |
727 |
offset, length = ranges[0]
|
728 |
response['Content-Length'] = length # Update with the correct length. |
729 |
response['Content-Range'] = 'bytes %d-%d/%d' % (offset, offset + length - 1, size) |
730 |
else:
|
731 |
del(response['Content-Length']) |
732 |
response['Content-Type'] = 'multipart/byteranges; boundary=%s' % (boundary,) |
733 |
return response
|
734 |
|
735 |
def put_object_block(request, hashmap, data, offset): |
736 |
"""Put one block of data at the given offset."""
|
737 |
|
738 |
bi = int(offset / request.backend.block_size)
|
739 |
bo = offset % request.backend.block_size |
740 |
bl = min(len(data), request.backend.block_size - bo) |
741 |
if bi < len(hashmap): |
742 |
hashmap[bi] = request.backend.update_block(hashmap[bi], data[:bl], bo) |
743 |
else:
|
744 |
hashmap.append(request.backend.put_block(('\x00' * bo) + data[:bl]))
|
745 |
return bl # Return ammount of data written. |
746 |
|
747 |
def hashmap_md5(request, hashmap, size): |
748 |
"""Produce the MD5 sum from the data in the hashmap."""
|
749 |
|
750 |
# TODO: Search backend for the MD5 of another object with the same hashmap and size...
|
751 |
md5 = hashlib.md5() |
752 |
bs = request.backend.block_size |
753 |
for bi, hash in enumerate(hashmap): |
754 |
data = request.backend.get_block(hash)
|
755 |
if bi == len(hashmap) - 1: |
756 |
bs = size % bs |
757 |
pad = bs - min(len(data), bs) |
758 |
md5.update(data + ('\x00' * pad))
|
759 |
return md5.hexdigest().lower()
|
760 |
|
761 |
def get_backend(): |
762 |
backend = connect_backend(db_module=settings.BACKEND_DB_MODULE, |
763 |
db_connection=settings.BACKEND_DB_CONNECTION, |
764 |
block_module=settings.BACKEND_BLOCK_MODULE, |
765 |
block_path=settings.BACKEND_BLOCK_PATH) |
766 |
backend.default_policy['quota'] = settings.BACKEND_QUOTA
|
767 |
backend.default_policy['versioning'] = settings.BACKEND_VERSIONING
|
768 |
return backend
|
769 |
|
770 |
def update_request_headers(request): |
771 |
# Handle URL-encoded keys and values.
|
772 |
# Handle URL-encoded keys and values.
|
773 |
meta = dict([(k, v) for k, v in request.META.iteritems() if k.startswith('HTTP_')]) |
774 |
if len(meta) > 90: |
775 |
raise BadRequest('Too many headers.') |
776 |
for k, v in meta.iteritems(): |
777 |
if len(k) > 128: |
778 |
raise BadRequest('Header name too large.') |
779 |
if len(v) > 256: |
780 |
raise BadRequest('Header value too large.') |
781 |
try:
|
782 |
k.decode('ascii')
|
783 |
v.decode('ascii')
|
784 |
except UnicodeDecodeError: |
785 |
raise BadRequest('Bad character in headers.') |
786 |
if '%' in k or '%' in v: |
787 |
del(request.META[k])
|
788 |
request.META[unquote(k)] = smart_unicode(unquote(v), strings_only=True)
|
789 |
|
790 |
def update_response_headers(request, response): |
791 |
if request.serialization == 'xml': |
792 |
response['Content-Type'] = 'application/xml; charset=UTF-8' |
793 |
elif request.serialization == 'json': |
794 |
response['Content-Type'] = 'application/json; charset=UTF-8' |
795 |
elif not response['Content-Type']: |
796 |
response['Content-Type'] = 'text/plain; charset=UTF-8' |
797 |
|
798 |
if (not response.has_header('Content-Length') and |
799 |
not (response.has_header('Content-Type') and |
800 |
response['Content-Type'].startswith('multipart/byteranges'))): |
801 |
response['Content-Length'] = len(response.content) |
802 |
|
803 |
# URL-encode unicode in headers.
|
804 |
meta = response.items() |
805 |
for k, v in meta: |
806 |
if (k.startswith('X-Account-') or k.startswith('X-Container-') or |
807 |
k.startswith('X-Object-') or k.startswith('Content-')): |
808 |
del(response[k])
|
809 |
response[quote(k)] = quote(v, safe='/=,:@; ')
|
810 |
|
811 |
if settings.TEST:
|
812 |
response['Date'] = format_date_time(time())
|
813 |
|
814 |
def render_fault(request, fault): |
815 |
if isinstance(fault, InternalServerError) and (settings.DEBUG or settings.TEST): |
816 |
fault.details = format_exc(fault) |
817 |
|
818 |
request.serialization = 'text'
|
819 |
data = fault.message + '\n'
|
820 |
if fault.details:
|
821 |
data += '\n' + fault.details
|
822 |
response = HttpResponse(data, status=fault.code) |
823 |
update_response_headers(request, response) |
824 |
return response
|
825 |
|
826 |
def request_serialization(request, format_allowed=False): |
827 |
"""Return the serialization format requested.
|
828 |
|
829 |
Valid formats are 'text' and 'json', 'xml' if 'format_allowed' is True.
|
830 |
"""
|
831 |
|
832 |
if not format_allowed: |
833 |
return 'text' |
834 |
|
835 |
format = request.GET.get('format')
|
836 |
if format == 'json': |
837 |
return 'json' |
838 |
elif format == 'xml': |
839 |
return 'xml' |
840 |
|
841 |
for item in request.META.get('HTTP_ACCEPT', '').split(','): |
842 |
accept, sep, rest = item.strip().partition(';')
|
843 |
if accept == 'application/json': |
844 |
return 'json' |
845 |
elif accept == 'application/xml' or accept == 'text/xml': |
846 |
return 'xml' |
847 |
|
848 |
return 'text' |
849 |
|
850 |
def api_method(http_method=None, format_allowed=False, user_required=True): |
851 |
"""Decorator function for views that implement an API method."""
|
852 |
|
853 |
def decorator(func): |
854 |
@wraps(func)
|
855 |
def wrapper(request, *args, **kwargs): |
856 |
try:
|
857 |
if http_method and request.method != http_method: |
858 |
raise BadRequest('Method not allowed.') |
859 |
if user_required and getattr(request, 'user', None) is None: |
860 |
raise Unauthorized('Access denied') |
861 |
|
862 |
# The args variable may contain up to (account, container, object).
|
863 |
if len(args) > 1 and len(args[1]) > 256: |
864 |
raise BadRequest('Container name too large.') |
865 |
if len(args) > 2 and len(args[2]) > 1024: |
866 |
raise BadRequest('Object name too large.') |
867 |
|
868 |
# Format and check headers.
|
869 |
update_request_headers(request) |
870 |
|
871 |
# Fill in custom request variables.
|
872 |
request.serialization = request_serialization(request, format_allowed) |
873 |
request.backend = get_backend() |
874 |
|
875 |
response = func(request, *args, **kwargs) |
876 |
update_response_headers(request, response) |
877 |
return response
|
878 |
except Fault, fault:
|
879 |
return render_fault(request, fault)
|
880 |
except BaseException, e: |
881 |
logger.exception('Unexpected error: %s' % e)
|
882 |
fault = InternalServerError('Unexpected error')
|
883 |
return render_fault(request, fault)
|
884 |
finally:
|
885 |
if getattr(request, 'backend', None) is not None: |
886 |
request.backend.close() |
887 |
return wrapper
|
888 |
return decorator
|