1 # Copyright 2011 GRNET S.A. All rights reserved.
3 # Redistribution and use in source and binary forms, with or
4 # without modification, are permitted provided that the following
7 # 1. Redistributions of source code must retain the above
8 # copyright notice, this list of conditions and the following
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.
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.
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.
34 from functools import wraps
36 from traceback import format_exc
37 from wsgiref.handlers import format_date_time
38 from binascii import hexlify
40 from django.conf import settings
41 from django.http import HttpResponse
42 from django.utils import simplejson as json
43 from django.utils.http import http_date, parse_etags
44 from django.utils.encoding import smart_str
46 from pithos.api.compat import parse_http_date_safe, parse_http_date
47 from pithos.api.faults import (Fault, NotModified, BadRequest, Unauthorized, ItemNotFound,
48 Conflict, LengthRequired, PreconditionFailed, RangeNotSatisfiable,
50 from pithos.backends import backend
51 from pithos.backends.base import NotAllowedError
60 logger = logging.getLogger(__name__)
63 def rename_meta_key(d, old, new):
69 def printable_header_dict(d):
70 """Format a meta dictionary for printing out json/xml.
72 Convert all keys to lower case and replace dashes with underscores.
73 Format 'last_modified' timestamp.
76 d['last_modified'] = datetime.datetime.fromtimestamp(int(d['last_modified'])).isoformat()
77 return dict([(k.lower().replace('-', '_'), v) for k, v in d.iteritems()])
79 def format_header_key(k):
80 """Convert underscores to dashes and capitalize intra-dash strings."""
81 return '-'.join([x.capitalize() for x in k.replace('_', '-').split('-')])
83 def get_header_prefix(request, prefix):
84 """Get all prefix-* request headers in a dict. Reformat keys with format_header_key()."""
86 prefix = 'HTTP_' + prefix.upper().replace('-', '_')
87 # TODO: Document or remove '~' replacing.
88 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)])
90 def get_account_headers(request):
91 meta = get_header_prefix(request, 'X-Account-Meta-')
93 for k, v in get_header_prefix(request, 'X-Account-Group-').iteritems():
95 if '-' in n or '_' in n:
96 raise BadRequest('Bad characters in group name')
97 groups[n] = v.replace(' ', '').split(',')
102 def put_account_headers(response, meta, groups):
104 response['X-Account-Container-Count'] = meta['count']
106 response['X-Account-Bytes-Used'] = meta['bytes']
107 response['Last-Modified'] = http_date(int(meta['modified']))
108 for k in [x for x in meta.keys() if x.startswith('X-Account-Meta-')]:
109 response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True)
110 if 'until_timestamp' in meta:
111 response['X-Account-Until-Timestamp'] = http_date(int(meta['until_timestamp']))
112 for k, v in groups.iteritems():
113 k = smart_str(k, strings_only=True)
114 k = format_header_key('X-Account-Group-' + k)
115 v = smart_str(','.join(v), strings_only=True)
118 def get_container_headers(request):
119 meta = get_header_prefix(request, 'X-Container-Meta-')
120 policy = dict([(k[19:].lower(), v.replace(' ', '')) for k, v in get_header_prefix(request, 'X-Container-Policy-').iteritems()])
123 def put_container_headers(response, meta, policy):
125 response['X-Container-Object-Count'] = meta['count']
127 response['X-Container-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-Container-Meta-')]:
130 response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True)
131 l = [smart_str(x, strings_only=True) for x in meta['object_meta'] if x.startswith('X-Object-Meta-')]
132 response['X-Container-Object-Meta'] = ','.join([x[14:] for x in l])
133 response['X-Container-Block-Size'] = backend.block_size
134 response['X-Container-Block-Hash'] = backend.hash_algorithm
135 if 'until_timestamp' in meta:
136 response['X-Container-Until-Timestamp'] = http_date(int(meta['until_timestamp']))
137 for k, v in policy.iteritems():
138 response[smart_str(format_header_key('X-Container-Policy-' + k), strings_only=True)] = smart_str(v, strings_only=True)
140 def get_object_headers(request):
141 meta = get_header_prefix(request, 'X-Object-Meta-')
142 if request.META.get('CONTENT_TYPE'):
143 meta['Content-Type'] = request.META['CONTENT_TYPE']
144 if request.META.get('HTTP_CONTENT_ENCODING'):
145 meta['Content-Encoding'] = request.META['HTTP_CONTENT_ENCODING']
146 if request.META.get('HTTP_CONTENT_DISPOSITION'):
147 meta['Content-Disposition'] = request.META['HTTP_CONTENT_DISPOSITION']
148 if request.META.get('HTTP_X_OBJECT_MANIFEST'):
149 meta['X-Object-Manifest'] = request.META['HTTP_X_OBJECT_MANIFEST']
150 return meta, get_sharing(request), get_public(request)
152 def put_object_headers(response, meta, restricted=False):
153 response['ETag'] = meta['hash']
154 response['Content-Length'] = meta['bytes']
155 response['Content-Type'] = meta.get('Content-Type', 'application/octet-stream')
156 response['Last-Modified'] = http_date(int(meta['modified']))
158 response['X-Object-Modified-By'] = smart_str(meta['modified_by'], strings_only=True)
159 response['X-Object-Version'] = meta['version']
160 response['X-Object-Version-Timestamp'] = http_date(int(meta['version_timestamp']))
161 for k in [x for x in meta.keys() if x.startswith('X-Object-Meta-')]:
162 response[smart_str(k, strings_only=True)] = smart_str(meta[k], strings_only=True)
163 for k in ('Content-Encoding', 'Content-Disposition', 'X-Object-Manifest', 'X-Object-Sharing', 'X-Object-Shared-By', 'X-Object-Public'):
165 response[k] = smart_str(meta[k], strings_only=True)
167 for k in ('Content-Encoding', 'Content-Disposition'):
169 response[k] = meta[k]
171 def update_manifest_meta(request, v_account, meta):
172 """Update metadata if the object has an X-Object-Manifest."""
174 if 'X-Object-Manifest' in meta:
178 src_container, src_name = split_container_object_string('/' + meta['X-Object-Manifest'])
179 objects = backend.list_objects(request.user, v_account, src_container, prefix=src_name, virtual=False)
181 src_meta = backend.get_object_meta(request.user, v_account, src_container, x[0], x[1])
182 hash += src_meta['hash']
183 bytes += src_meta['bytes']
187 meta['bytes'] = bytes
190 meta['hash'] = md5.hexdigest().lower()
192 def update_sharing_meta(permissions, v_account, v_container, v_object, meta):
193 if permissions is None:
195 perm_path, perms = permissions
199 r = ','.join(perms.get('read', []))
201 ret.append('read=' + r)
202 w = ','.join(perms.get('write', []))
204 ret.append('write=' + w)
205 meta['X-Object-Sharing'] = '; '.join(ret)
206 if '/'.join((v_account, v_container, v_object)) != perm_path:
207 meta['X-Object-Shared-By'] = perm_path
209 def update_public_meta(public, meta):
212 meta['X-Object-Public'] = public
214 def validate_modification_preconditions(request, meta):
215 """Check that the modified timestamp conforms with the preconditions set."""
217 if 'modified' not in meta:
218 return # TODO: Always return?
220 if_modified_since = request.META.get('HTTP_IF_MODIFIED_SINCE')
221 if if_modified_since is not None:
222 if_modified_since = parse_http_date_safe(if_modified_since)
223 if if_modified_since is not None and int(meta['modified']) <= if_modified_since:
224 raise NotModified('Resource has not been modified')
226 if_unmodified_since = request.META.get('HTTP_IF_UNMODIFIED_SINCE')
227 if if_unmodified_since is not None:
228 if_unmodified_since = parse_http_date_safe(if_unmodified_since)
229 if if_unmodified_since is not None and int(meta['modified']) > if_unmodified_since:
230 raise PreconditionFailed('Resource has been modified')
232 def validate_matching_preconditions(request, meta):
233 """Check that the ETag conforms with the preconditions set."""
235 hash = meta.get('hash', None)
237 if_match = request.META.get('HTTP_IF_MATCH')
238 if if_match is not None:
240 raise PreconditionFailed('Resource does not exist')
241 if if_match != '*' and hash not in [x.lower() for x in parse_etags(if_match)]:
242 raise PreconditionFailed('Resource ETag does not match')
244 if_none_match = request.META.get('HTTP_IF_NONE_MATCH')
245 if if_none_match is not None:
246 # TODO: If this passes, must ignore If-Modified-Since header.
248 if if_none_match == '*' or hash in [x.lower() for x in parse_etags(if_none_match)]:
249 # TODO: Continue if an If-Modified-Since header is present.
250 if request.method in ('HEAD', 'GET'):
251 raise NotModified('Resource ETag matches')
252 raise PreconditionFailed('Resource exists or ETag matches')
254 def split_container_object_string(s):
255 if not len(s) > 0 or s[0] != '/':
261 return s[:pos], s[(pos + 1):]
263 def copy_or_move_object(request, v_account, src_container, src_name, dest_container, dest_name, move=False):
264 """Copy or move an object."""
266 meta, permissions, public = get_object_headers(request)
267 src_version = request.META.get('HTTP_X_SOURCE_VERSION')
270 backend.move_object(request.user, v_account, src_container, src_name, dest_container, dest_name, meta, False, permissions)
272 backend.copy_object(request.user, v_account, src_container, src_name, dest_container, dest_name, meta, False, permissions, src_version)
273 except NotAllowedError:
274 raise Unauthorized('Access denied')
275 except (NameError, IndexError):
276 raise ItemNotFound('Container or object does not exist')
278 raise BadRequest('Invalid sharing header')
279 except AttributeError, e:
280 raise Conflict(json.dumps(e.data))
281 if public is not None:
283 backend.update_object_public(request.user, v_account, dest_container, dest_name, public)
284 except NotAllowedError:
285 raise Unauthorized('Access denied')
287 raise ItemNotFound('Object does not exist')
289 def get_int_parameter(p):
299 def get_content_length(request):
300 content_length = get_int_parameter(request.META.get('CONTENT_LENGTH'))
301 if content_length is None:
302 raise LengthRequired('Missing or invalid Content-Length header')
303 return content_length
305 def get_range(request, size):
306 """Parse a Range header from the request.
308 Either returns None, when the header is not existent or should be ignored,
309 or a list of (offset, length) tuples - should be further checked.
312 ranges = request.META.get('HTTP_RANGE', '').replace(' ', '')
313 if not ranges.startswith('bytes='):
317 for r in (x.strip() for x in ranges[6:].split(',')):
318 p = re.compile('^(?P<offset>\d*)-(?P<upto>\d*)$')
322 offset = m.group('offset')
323 upto = m.group('upto')
324 if offset == '' and upto == '':
333 ret.append((offset, upto - offset + 1))
335 ret.append((offset, size - offset))
338 ret.append((size - length, length))
342 def get_content_range(request):
343 """Parse a Content-Range header from the request.
345 Either returns None, when the header is not existent or should be ignored,
346 or an (offset, length, total) tuple - check as length, total may be None.
347 Returns (None, None, None) if the provided range is '*/*'.
350 ranges = request.META.get('HTTP_CONTENT_RANGE', '')
354 p = re.compile('^bytes (?P<offset>\d+)-(?P<upto>\d*)/(?P<total>(\d+|\*))$')
357 if ranges == 'bytes */*':
358 return (None, None, None)
360 offset = int(m.group('offset'))
361 upto = m.group('upto')
362 total = m.group('total')
371 if (upto is not None and offset > upto) or \
372 (total is not None and offset >= total) or \
373 (total is not None and upto is not None and upto >= total):
379 length = upto - offset + 1
380 return (offset, length, total)
382 def get_sharing(request):
383 """Parse an X-Object-Sharing header from the request.
385 Raises BadRequest on error.
388 permissions = request.META.get('HTTP_X_OBJECT_SHARING')
389 if permissions is None:
393 permissions = permissions.replace(' ', '')
394 if permissions == '':
396 for perm in (x for x in permissions.split(';')):
397 if perm.startswith('read='):
398 ret['read'] = list(set([v.replace(' ','').lower() for v in perm[5:].split(',')]))
399 if '' in ret['read']:
400 ret['read'].remove('')
401 if '*' in ret['read']:
403 if len(ret['read']) == 0:
404 raise BadRequest('Bad X-Object-Sharing header value')
405 elif perm.startswith('write='):
406 ret['write'] = list(set([v.replace(' ','').lower() for v in perm[6:].split(',')]))
407 if '' in ret['write']:
408 ret['write'].remove('')
409 if '*' in ret['write']:
411 if len(ret['write']) == 0:
412 raise BadRequest('Bad X-Object-Sharing header value')
414 raise BadRequest('Bad X-Object-Sharing header value')
417 def get_public(request):
418 """Parse an X-Object-Public header from the request.
420 Raises BadRequest on error.
423 public = request.META.get('HTTP_X_OBJECT_PUBLIC')
427 public = public.replace(' ', '').lower()
430 elif public == 'false' or public == '':
432 raise BadRequest('Bad X-Object-Public header value')
434 def raw_input_socket(request):
435 """Return the socket for reading the rest of the request."""
437 server_software = request.META.get('SERVER_SOFTWARE')
438 if server_software and server_software.startswith('mod_python'):
440 if 'wsgi.input' in request.environ:
441 return request.environ['wsgi.input']
442 raise ServiceUnavailable('Unknown server software')
444 MAX_UPLOAD_SIZE = 10 * (1024 * 1024) # 10MB
446 def socket_read_iterator(request, length=0, blocksize=4096):
447 """Return a maximum of blocksize data read from the socket in each iteration.
449 Read up to 'length'. If 'length' is negative, will attempt a chunked read.
450 The maximum ammount of data read is controlled by MAX_UPLOAD_SIZE.
453 sock = raw_input_socket(request)
454 if length < 0: # Chunked transfers
455 # Small version (server does the dechunking).
456 if request.environ.get('mod_wsgi.input_chunked', None):
457 while length < MAX_UPLOAD_SIZE:
458 data = sock.read(blocksize)
462 raise BadRequest('Maximum size is reached')
464 # Long version (do the dechunking).
466 while length < MAX_UPLOAD_SIZE:
468 if hasattr(sock, 'readline'):
469 chunk_length = sock.readline()
472 while chunk_length[-1:] != '\n':
473 chunk_length += sock.read(1)
475 pos = chunk_length.find(';')
477 chunk_length = chunk_length[:pos]
479 chunk_length = int(chunk_length, 16)
481 raise BadRequest('Bad chunk size') # TODO: Change to something more appropriate.
483 if chunk_length == 0:
487 # Get the actual data.
488 while chunk_length > 0:
489 chunk = sock.read(min(chunk_length, blocksize))
490 chunk_length -= len(chunk)
494 if len(data) >= blocksize:
495 ret = data[:blocksize]
496 data = data[blocksize:]
499 raise BadRequest('Maximum size is reached')
501 if length > MAX_UPLOAD_SIZE:
502 raise BadRequest('Maximum size is reached')
504 data = sock.read(min(length, blocksize))
508 class ObjectWrapper(object):
509 """Return the object's data block-per-block in each iteration.
511 Read from the object using the offset and length provided in each entry of the range list.
514 def __init__(self, ranges, sizes, hashmaps, boundary):
517 self.hashmaps = hashmaps
518 self.boundary = boundary
519 self.size = sum(self.sizes)
526 self.range_index = -1
527 self.offset, self.length = self.ranges[0]
532 def part_iterator(self):
534 # Get the file for the current offset.
535 file_size = self.sizes[self.file_index]
536 while self.offset >= file_size:
537 self.offset -= file_size
539 file_size = self.sizes[self.file_index]
541 # Get the block for the current position.
542 self.block_index = int(self.offset / backend.block_size)
543 if self.block_hash != self.hashmaps[self.file_index][self.block_index]:
544 self.block_hash = self.hashmaps[self.file_index][self.block_index]
546 self.block = backend.get_block(self.block_hash)
548 raise ItemNotFound('Block does not exist')
550 # Get the data from the block.
551 bo = self.offset % backend.block_size
552 bl = min(self.length, len(self.block) - bo)
553 data = self.block[bo:bo + bl]
561 if len(self.ranges) == 1:
562 return self.part_iterator()
563 if self.range_index == len(self.ranges):
566 if self.range_index == -1:
568 return self.part_iterator()
569 except StopIteration:
570 self.range_index += 1
572 if self.range_index < len(self.ranges):
574 self.offset, self.length = self.ranges[self.range_index]
576 if self.range_index > 0:
578 out.append('--' + self.boundary)
579 out.append('Content-Range: bytes %d-%d/%d' % (self.offset, self.offset + self.length - 1, self.size))
580 out.append('Content-Transfer-Encoding: binary')
583 return '\r\n'.join(out)
587 out.append('--' + self.boundary + '--')
589 return '\r\n'.join(out)
591 def object_data_response(request, sizes, hashmaps, meta, public=False):
592 """Get the HttpResponse object for replying with the object's data."""
596 ranges = get_range(request, size)
601 check = [True for offset, length in ranges if
602 length <= 0 or length > size or
603 offset < 0 or offset >= size or
604 offset + length > size]
606 raise RangeNotSatisfiable('Requested range exceeds object limits')
608 if_range = request.META.get('HTTP_IF_RANGE')
611 # Modification time has passed instead.
612 last_modified = parse_http_date(if_range)
613 if last_modified != meta['modified']:
617 if if_range != meta['hash']:
621 if ret == 206 and len(ranges) > 1:
622 boundary = uuid.uuid4().hex
625 wrapper = ObjectWrapper(ranges, sizes, hashmaps, boundary)
626 response = HttpResponse(wrapper, status=ret)
627 put_object_headers(response, meta, public)
630 offset, length = ranges[0]
631 response['Content-Length'] = length # Update with the correct length.
632 response['Content-Range'] = 'bytes %d-%d/%d' % (offset, offset + length - 1, size)
634 del(response['Content-Length'])
635 response['Content-Type'] = 'multipart/byteranges; boundary=%s' % (boundary,)
638 def put_object_block(hashmap, data, offset):
639 """Put one block of data at the given offset."""
641 bi = int(offset / backend.block_size)
642 bo = offset % backend.block_size
643 bl = min(len(data), backend.block_size - bo)
644 if bi < len(hashmap):
645 hashmap[bi] = backend.update_block(hashmap[bi], data[:bl], bo)
647 hashmap.append(backend.put_block(('\x00' * bo) + data[:bl]))
648 return bl # Return ammount of data written.
650 def hashmap_hash(hashmap):
651 """Produce the root hash, treating the hashmap as a Merkle-like tree."""
654 h = hashlib.new(backend.hash_algorithm)
658 if len(hashmap) == 0:
659 return hexlify(subhash(''))
660 if len(hashmap) == 1:
663 while s < len(hashmap):
665 h = hashmap + ([('\x00' * len(hashmap[0]))] * (s - len(hashmap)))
666 h = [subhash(h[x] + (h[x + 1] if x + 1 < len(h) else '')) for x in range(0, len(h), 2)]
668 h = [subhash(h[x] + (h[x + 1] if x + 1 < len(h) else '')) for x in range(0, len(h), 2)]
671 def update_response_headers(request, response):
672 if request.serialization == 'xml':
673 response['Content-Type'] = 'application/xml; charset=UTF-8'
674 elif request.serialization == 'json':
675 response['Content-Type'] = 'application/json; charset=UTF-8'
676 elif not response['Content-Type']:
677 response['Content-Type'] = 'text/plain; charset=UTF-8'
679 if not response.has_header('Content-Length') and not (response.has_header('Content-Type') and response['Content-Type'].startswith('multipart/byteranges')):
680 response['Content-Length'] = len(response.content)
683 response['Date'] = format_date_time(time())
685 def render_fault(request, fault):
686 if settings.DEBUG or settings.TEST:
687 fault.details = format_exc(fault)
689 request.serialization = 'text'
690 data = '\n'.join((fault.message, fault.details)) + '\n'
691 response = HttpResponse(data, status=fault.code)
692 update_response_headers(request, response)
695 def request_serialization(request, format_allowed=False):
696 """Return the serialization format requested.
698 Valid formats are 'text' and 'json', 'xml' if 'format_allowed' is True.
701 if not format_allowed:
704 format = request.GET.get('format')
707 elif format == 'xml':
710 # for item in request.META.get('HTTP_ACCEPT', '').split(','):
711 # accept, sep, rest = item.strip().partition(';')
712 # if accept == 'application/json':
714 # elif accept == 'application/xml' or accept == 'text/xml':
719 def api_method(http_method=None, format_allowed=False):
720 """Decorator function for views that implement an API method."""
724 def wrapper(request, *args, **kwargs):
726 if http_method and request.method != http_method:
727 raise BadRequest('Method not allowed.')
729 # The args variable may contain up to (account, container, object).
730 if len(args) > 1 and len(args[1]) > 256:
731 raise BadRequest('Container name too large.')
732 if len(args) > 2 and len(args[2]) > 1024:
733 raise BadRequest('Object name too large.')
735 # Fill in custom request variables.
736 request.serialization = request_serialization(request, format_allowed)
738 response = func(request, *args, **kwargs)
739 update_response_headers(request, response)
742 return render_fault(request, fault)
743 except BaseException, e:
744 logger.exception('Unexpected error: %s' % e)
745 fault = ServiceUnavailable('Unexpected error')
746 return render_fault(request, fault)