Commit | Line | Data |
---|---|---|
8e1e744d | 1 | # GNU MediaGoblin -- federated, autonomous media hosting |
a6b378ef CAW |
2 | # Copyright (C) 2011 Free Software Foundation, Inc |
3 | # | |
4 | # This program is free software: you can redistribute it and/or modify | |
5 | # it under the terms of the GNU Affero General Public License as published by | |
6 | # the Free Software Foundation, either version 3 of the License, or | |
7 | # (at your option) any later version. | |
8 | # | |
9 | # This program is distributed in the hope that it will be useful, | |
10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
12 | # GNU Affero General Public License for more details. | |
13 | # | |
14 | # You should have received a copy of the GNU Affero General Public License | |
15 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
16 | ||
779f2b94 | 17 | import os |
ffa22935 | 18 | import re |
6a07362d | 19 | import shutil |
f61a41b8 | 20 | import urlparse |
2fdec827 | 21 | import uuid |
851c51a3 | 22 | import cloudfiles |
a6b378ef CAW |
23 | |
24 | from werkzeug.utils import secure_filename | |
25 | ||
ffa22935 CAW |
26 | from mediagoblin import util |
27 | ||
d807b725 CAW |
28 | ######## |
29 | # Errors | |
30 | ######## | |
a6b378ef | 31 | |
770c12be | 32 | |
aa797ca1 JW |
33 | class Error(Exception): |
34 | pass | |
35 | ||
36 | ||
37 | class InvalidFilepath(Error): | |
38 | pass | |
39 | ||
40 | ||
41 | class NoWebServing(Error): | |
42 | pass | |
43 | ||
44 | ||
45 | class NotImplementedError(Error): | |
46 | pass | |
797be93c | 47 | |
770c12be | 48 | |
d807b725 CAW |
49 | ############################################### |
50 | # Storage interface & basic file implementation | |
51 | ############################################### | |
a6b378ef | 52 | |
797be93c CAW |
53 | class StorageInterface(object): |
54 | """ | |
55 | Interface for the storage API. | |
56 | ||
57 | This interface doesn't actually provide behavior, but it defines | |
58 | what kind of storage patterns subclasses should provide. | |
59 | ||
60 | It is important to note that the storage API idea of a "filepath" | |
61 | is actually like ['dir1', 'dir2', 'file.jpg'], so keep that in | |
62 | mind while reading method documentation. | |
5afb9227 CAW |
63 | |
64 | You should set up your __init__ method with whatever keyword | |
65 | arguments are appropriate to your storage system, but you should | |
66 | also passively accept all extraneous keyword arguments like: | |
67 | ||
68 | def __init__(self, **kwargs): | |
69 | pass | |
70 | ||
71 | See BasicFileStorage as a simple implementation of the | |
72 | StorageInterface. | |
797be93c | 73 | """ |
797be93c | 74 | |
3a89c23e CAW |
75 | # Whether this file store is on the local filesystem. |
76 | local_storage = False | |
77 | ||
797be93c CAW |
78 | def __raise_not_implemented(self): |
79 | """ | |
80 | Raise a warning about some component not implemented by a | |
81 | subclass of this interface. | |
82 | """ | |
83 | raise NotImplementedError( | |
84 | "This feature not implemented in this storage API implementation.") | |
85 | ||
86 | def file_exists(self, filepath): | |
87 | """ | |
88 | Return a boolean asserting whether or not file at filepath | |
89 | exists in our storage system. | |
90 | ||
91 | Returns: | |
92 | True / False depending on whether file exists or not. | |
93 | """ | |
94 | # Subclasses should override this method. | |
95 | self.__raise_not_implemented() | |
96 | ||
cee7a1c1 | 97 | def get_file(self, filepath, mode='r'): |
b0de01cf CAW |
98 | """ |
99 | Return a file-like object for reading/writing from this filepath. | |
100 | ||
101 | Should create directories, buckets, whatever, as necessary. | |
102 | """ | |
0b9cf289 CAW |
103 | # Subclasses should override this method. |
104 | self.__raise_not_implemented() | |
105 | ||
106 | def delete_file(self, filepath): | |
b0de01cf CAW |
107 | """ |
108 | Delete or dereference the file at filepath. | |
109 | ||
110 | This might need to delete directories, buckets, whatever, for | |
111 | cleanliness. (Be sure to avoid race conditions on that though) | |
112 | """ | |
0b9cf289 CAW |
113 | # Subclasses should override this method. |
114 | self.__raise_not_implemented() | |
115 | ||
f61a41b8 | 116 | def file_url(self, filepath): |
644614d4 CAW |
117 | """ |
118 | Get the URL for this file. This assumes our storage has been | |
119 | mounted with some kind of URL which makes this possible. | |
120 | """ | |
121 | # Subclasses should override this method. | |
122 | self.__raise_not_implemented() | |
123 | ||
2d1a6073 | 124 | def get_unique_filepath(self, filepath): |
797be93c CAW |
125 | """ |
126 | If a filename at filepath already exists, generate a new name. | |
127 | ||
128 | Eg, if the filename doesn't exist: | |
129 | >>> storage_handler.get_unique_filename(['dir1', 'dir2', 'fname.jpg']) | |
130 | [u'dir1', u'dir2', u'fname.jpg'] | |
aa797ca1 | 131 | |
797be93c CAW |
132 | But if a file does exist, let's get one back with at uuid tacked on: |
133 | >>> storage_handler.get_unique_filename(['dir1', 'dir2', 'fname.jpg']) | |
ef10e3a2 | 134 | [u'dir1', u'dir2', u'd02c3571-dd62-4479-9d62-9e3012dada29-fname.jpg'] |
797be93c | 135 | """ |
b0bfb766 CAW |
136 | # Make sure we have a clean filepath to start with, since |
137 | # we'll be possibly tacking on stuff to the filename. | |
138 | filepath = clean_listy_filepath(filepath) | |
139 | ||
0b9cf289 CAW |
140 | if self.file_exists(filepath): |
141 | return filepath[:-1] + ["%s-%s" % (uuid.uuid4(), filepath[-1])] | |
142 | else: | |
143 | return filepath | |
779f2b94 | 144 | |
3a89c23e CAW |
145 | def get_local_path(self, filepath): |
146 | """ | |
147 | If this is a local_storage implementation, give us a link to | |
148 | the local filesystem reference to this file. | |
149 | ||
150 | >>> storage_handler.get_local_path(['foo', 'bar', 'baz.jpg']) | |
151 | u'/path/to/mounting/foo/bar/baz.jpg' | |
152 | """ | |
153 | # Subclasses should override this method, if applicable. | |
154 | self.__raise_not_implemented() | |
155 | ||
6a07362d CAW |
156 | def copy_locally(self, filepath, dest_path): |
157 | """ | |
158 | Copy this file locally. | |
159 | ||
160 | A basic working method for this is provided that should | |
161 | function both for local_storage systems and remote storge | |
162 | systems, but if more efficient systems for copying locally | |
163 | apply to your system, override this method with something more | |
164 | appropriate. | |
165 | """ | |
166 | if self.local_storage: | |
167 | shutil.copy( | |
168 | self.get_local_path(filepath), dest_path) | |
169 | else: | |
170 | with self.get_file(filepath, 'rb') as source_file: | |
171 | with file(dest_path, 'wb') as dest_file: | |
172 | dest_file.write(source_file.read()) | |
173 | ||
779f2b94 CAW |
174 | |
175 | class BasicFileStorage(StorageInterface): | |
176 | """ | |
177 | Basic local filesystem implementation of storage API | |
178 | """ | |
179 | ||
3a89c23e CAW |
180 | local_storage = True |
181 | ||
5afb9227 | 182 | def __init__(self, base_dir, base_url=None, **kwargs): |
779f2b94 CAW |
183 | """ |
184 | Keyword arguments: | |
185 | - base_dir: Base directory things will be served out of. MUST | |
186 | be an absolute path. | |
b1bb050b | 187 | - base_url: URL files will be served from |
779f2b94 CAW |
188 | """ |
189 | self.base_dir = base_dir | |
b1bb050b | 190 | self.base_url = base_url |
779f2b94 CAW |
191 | |
192 | def _resolve_filepath(self, filepath): | |
193 | """ | |
194 | Transform the given filepath into a local filesystem filepath. | |
195 | """ | |
196 | return os.path.join( | |
197 | self.base_dir, *clean_listy_filepath(filepath)) | |
aa797ca1 | 198 | |
779f2b94 CAW |
199 | def file_exists(self, filepath): |
200 | return os.path.exists(self._resolve_filepath(filepath)) | |
201 | ||
cee7a1c1 CAW |
202 | def get_file(self, filepath, mode='r'): |
203 | # Make directories if necessary | |
204 | if len(filepath) > 1: | |
205 | directory = self._resolve_filepath(filepath[:-1]) | |
d0e3a534 | 206 | if not os.path.exists(directory): |
cee7a1c1 CAW |
207 | os.makedirs(directory) |
208 | ||
209 | # Grab and return the file in the mode specified | |
210 | return open(self._resolve_filepath(filepath), mode) | |
211 | ||
779f2b94 | 212 | def delete_file(self, filepath): |
b1bb050b CAW |
213 | # TODO: Also delete unused directories if empty (safely, with |
214 | # checks to avoid race conditions). | |
215 | os.remove(self._resolve_filepath(filepath)) | |
644614d4 | 216 | |
f61a41b8 | 217 | def file_url(self, filepath): |
b1bb050b CAW |
218 | if not self.base_url: |
219 | raise NoWebServing( | |
220 | "base_url not set, cannot provide file urls") | |
221 | ||
222 | return urlparse.urljoin( | |
223 | self.base_url, | |
224 | '/'.join(clean_listy_filepath(filepath))) | |
ffa22935 | 225 | |
3a89c23e CAW |
226 | def get_local_path(self, filepath): |
227 | return self._resolve_filepath(filepath) | |
228 | ||
ffa22935 | 229 | |
aa797ca1 JW |
230 | class CloudFilesStorage(StorageInterface): |
231 | def __init__(self, **kwargs): | |
232 | self.param_container = kwargs.get('cloudfiles_container') | |
233 | self.param_user = kwargs.get('cloudfiles_user') | |
234 | self.param_api_key = kwargs.get('cloudfiles_api_key') | |
235 | self.param_host = kwargs.get('cloudfiles_host') | |
236 | self.param_use_servicenet = kwargs.get('cloudfiles_use_servicenet') | |
237 | ||
238 | if not self.param_host: | |
239 | print('No CloudFiles host URL specified, ' | |
240 | 'defaulting to Rackspace US') | |
241 | ||
242 | self.connection = cloudfiles.get_connection( | |
243 | username=self.param_user, | |
244 | api_key=self.param_api_key, | |
245 | servicenet=True if self.param_use_servicenet == 'true' or \ | |
246 | self.param_use_servicenet == True else False) | |
247 | ||
248 | if not self.param_container == \ | |
249 | self.connection.get_container(self.param_container): | |
250 | self.container = self.connection.create_container( | |
251 | self.param_container) | |
252 | self.container.make_public( | |
253 | ttl=60 * 60 * 2) | |
254 | else: | |
255 | self.container = self.connection.get_container( | |
256 | self.param_container) | |
257 | ||
258 | def _resolve_filepath(self, filepath): | |
259 | return '/'.join( | |
260 | clean_listy_filepath(filepath)) | |
261 | ||
262 | def file_exists(self, filepath): | |
263 | try: | |
264 | object = self.container.get_object( | |
265 | self._resolve_filepath(filepath)) | |
266 | return True | |
267 | except cloudfiles.errors.NoSuchObject: | |
268 | return False | |
269 | ||
270 | def get_file(self, filepath, mode='r'): | |
271 | try: | |
272 | obj = self.container.get_object( | |
273 | self._resolve_filepath(filepath)) | |
274 | except cloudfiles.errors.NoSuchObject: | |
275 | obj = self.container.create_object( | |
276 | self._resolve_filepath(filepath)) | |
277 | ||
278 | return obj | |
279 | ||
280 | def delete_file(self, filepath): | |
281 | # TODO: Also delete unused directories if empty (safely, with | |
282 | # checks to avoid race conditions). | |
283 | self.container.delete_object(filepath) | |
284 | ||
285 | def file_url(self, filepath): | |
286 | return self.get_file(filepath).public_uri() | |
287 | ||
288 | ||
68cf996c | 289 | class MountStorage(StorageInterface): |
255f02c4 E |
290 | """ |
291 | Experimental "Mount" virtual Storage Interface | |
292 | ||
f016fc65 CAW |
293 | This isn't an interface to some real storage, instead it's a |
294 | redirecting interface, that redirects requests to other | |
295 | "StorageInterface"s. | |
296 | ||
297 | For example, say you have the paths: | |
298 | ||
299 | 1. ['user_data', 'cwebber', 'avatar.jpg'] | |
300 | 2. ['user_data', 'elrond', 'avatar.jpg'] | |
301 | 3. ['media_entries', '34352f304c3f4d0ad8ad0f043522b6f2', 'thumb.jpg'] | |
302 | ||
303 | You could mount media_entries under CloudFileStorage and user_data | |
304 | under BasicFileStorage. Then 1 would be passed to | |
305 | BasicFileStorage under the path ['cwebber', 'avatar.jpg'] and 3 | |
306 | would be passed to CloudFileStorage under | |
307 | ['34352f304c3f4d0ad8ad0f043522b6f2', 'thumb.jpg']. | |
308 | ||
309 | In other words, this is kind of like mounting /home/ and /etc/ | |
310 | under different filesystems on your operating system... but with | |
311 | mediagoblin filestorages :) | |
255f02c4 | 312 | |
f016fc65 CAW |
313 | To set this up, you currently need to call the mount() method with |
314 | the target path and a backend, that shall be available under that | |
315 | target path. You have to mount things in a sensible order, | |
255f02c4 E |
316 | especially you can't mount ["a", "b"] before ["a"]. |
317 | """ | |
68cf996c E |
318 | def __init__(self, **kwargs): |
319 | self.mounttab = {} | |
320 | ||
321 | def mount(self, dirpath, backend): | |
322 | """ | |
323 | Mount a new backend under dirpath | |
324 | """ | |
325 | new_ent = clean_listy_filepath(dirpath) | |
68cf996c E |
326 | |
327 | print "Mounting:", repr(new_ent) | |
eea31562 | 328 | already, rem_1, table, rem_2 = self._resolve_to_backend(new_ent, True) |
620fca54 E |
329 | print "===", repr(already), repr(rem_1), repr(rem_2), len(table) |
330 | ||
331 | assert (len(rem_2) > 0) or (None not in table), \ | |
332 | "That path is already mounted" | |
333 | assert (len(rem_2) > 0) or (len(table)==0), \ | |
334 | "A longer path is already mounted here" | |
68cf996c | 335 | |
68cf996c E |
336 | for part in rem_2: |
337 | table[part] = {} | |
338 | table = table[part] | |
339 | table[None] = backend | |
340 | ||
eea31562 | 341 | def _resolve_to_backend(self, filepath, extra_info = False): |
68cf996c E |
342 | """ |
343 | extra_info = True is for internal use! | |
344 | ||
345 | Normally, returns the backend and the filepath inside that backend. | |
346 | ||
347 | With extra_info = True it returns the last directory node and the | |
348 | remaining filepath from there in addition. | |
349 | """ | |
350 | table = self.mounttab | |
351 | filepath = filepath[:] | |
352 | res_fp = None | |
353 | while True: | |
354 | new_be = table.get(None) | |
355 | if (new_be is not None) or res_fp is None: | |
356 | res_be = new_be | |
357 | res_fp = filepath[:] | |
358 | res_extra = (table, filepath[:]) | |
359 | # print "... New res: %r, %r, %r" % (res_be, res_fp, res_extra) | |
360 | if len(filepath) == 0: | |
361 | break | |
362 | query = filepath.pop(0) | |
363 | entry = table.get(query) | |
364 | if entry is not None: | |
365 | table = entry | |
366 | res_extra = (table, filepath[:]) | |
367 | else: | |
368 | break | |
369 | if extra_info: | |
370 | return (res_be, res_fp) + res_extra | |
371 | else: | |
372 | return (res_be, res_fp) | |
373 | ||
eea31562 E |
374 | def resolve_to_backend(self, filepath): |
375 | backend, filepath = self._resolve_to_backend(filepath) | |
376 | if backend is None: | |
377 | raise Error("Path not mounted") | |
378 | return backend, filepath | |
379 | ||
93b2796c | 380 | def __repr__(self, table = None, indent = []): |
68cf996c E |
381 | res = [] |
382 | if table is None: | |
383 | res.append("MountStorage<") | |
384 | table = self.mounttab | |
385 | v = table.get(None) | |
386 | if v: | |
93b2796c | 387 | res.append(" " * len(indent) + repr(indent) + ": " + repr(v)) |
68cf996c E |
388 | for k, v in table.iteritems(): |
389 | if k == None: | |
390 | continue | |
93b2796c E |
391 | res.append(" " * len(indent) + repr(k) + ":") |
392 | res += self.__repr__(v, indent + [k]) | |
68cf996c E |
393 | if table is self.mounttab: |
394 | res.append(">") | |
395 | return "\n".join(res) | |
396 | else: | |
397 | return res | |
398 | ||
937e2c88 E |
399 | def file_exists(self, filepath): |
400 | backend, filepath = self.resolve_to_backend(filepath) | |
401 | return backend.file_exists(filepath) | |
402 | ||
403 | def get_file(self, filepath, mode='r'): | |
404 | backend, filepath = self.resolve_to_backend(filepath) | |
405 | return backend.get_file(filepath, mode) | |
406 | ||
407 | def delete_file(self, filepath): | |
408 | backend, filepath = self.resolve_to_backend(filepath) | |
409 | return backend.delete_file(filepath) | |
410 | ||
411 | def file_url(self, filepath): | |
412 | backend, filepath = self.resolve_to_backend(filepath) | |
413 | return backend.file_url(filepath) | |
414 | ||
415 | def get_local_path(self, filepath): | |
416 | backend, filepath = self.resolve_to_backend(filepath) | |
417 | return backend.get_local_path(filepath) | |
418 | ||
419 | def copy_locally(self, filepath, dest_path): | |
420 | """ | |
421 | Need to override copy_locally, because the local_storage | |
422 | attribute is not correct. | |
423 | """ | |
424 | backend, filepath = self.resolve_to_backend(filepath) | |
425 | backend.copy_locally(filepath, dest_path) | |
426 | ||
68cf996c | 427 | |
d807b725 CAW |
428 | ########### |
429 | # Utilities | |
430 | ########### | |
431 | ||
432 | def clean_listy_filepath(listy_filepath): | |
433 | """ | |
434 | Take a listy filepath (like ['dir1', 'dir2', 'filename.jpg']) and | |
435 | clean out any nastiness from it. | |
436 | ||
3a89c23e | 437 | |
d807b725 CAW |
438 | >>> clean_listy_filepath([u'/dir1/', u'foo/../nasty', u'linooks.jpg']) |
439 | [u'dir1', u'foo_.._nasty', u'linooks.jpg'] | |
440 | ||
441 | Args: | |
442 | - listy_filepath: a list of filepath components, mediagoblin | |
443 | storage API style. | |
444 | ||
445 | Returns: | |
446 | A cleaned list of unicode objects. | |
447 | """ | |
448 | cleaned_filepath = [ | |
449 | unicode(secure_filename(filepath)) | |
450 | for filepath in listy_filepath] | |
451 | ||
452 | if u'' in cleaned_filepath: | |
453 | raise InvalidFilepath( | |
454 | "A filename component could not be resolved into a usable name.") | |
455 | ||
456 | return cleaned_filepath | |
457 | ||
458 | ||
3c7d11ff | 459 | def storage_system_from_config(paste_config, storage_prefix): |
ffa22935 CAW |
460 | """ |
461 | Utility for setting up a storage system from the paste app config. | |
462 | ||
463 | Note that a special argument may be passed in to the paste_config | |
464 | which is "${storage_prefix}_storage_class" which will provide an | |
465 | import path to a storage system. This defaults to | |
466 | "mediagoblin.storage:BasicFileStorage" if otherwise undefined. | |
467 | ||
468 | Arguments: | |
469 | - paste_config: dictionary of config parameters | |
470 | - storage_prefix: the storage system we're setting up / will be | |
471 | getting keys/arguments from. For example 'publicstore' will | |
472 | grab all arguments that are like 'publicstore_FOO'. | |
473 | ||
474 | Returns: | |
475 | An instantiated storage system. | |
476 | ||
477 | Example: | |
3c7d11ff | 478 | storage_system_from_config( |
ffa22935 CAW |
479 | {'publicstore_base_url': '/media/', |
480 | 'publicstore_base_dir': '/var/whatever/media/'}, | |
481 | 'publicstore') | |
482 | ||
483 | Will return: | |
484 | BasicFileStorage( | |
485 | base_url='/media/', | |
486 | base_dir='/var/whatever/media') | |
487 | """ | |
488 | prefix_re = re.compile('^%s_(.+)$' % re.escape(storage_prefix)) | |
489 | ||
490 | config_params = dict( | |
491 | [(prefix_re.match(key).groups()[0], value) | |
492 | for key, value in paste_config.iteritems() | |
493 | if prefix_re.match(key)]) | |
494 | ||
aa797ca1 | 495 | if 'storage_class' in config_params: |
ffa22935 CAW |
496 | storage_class = config_params['storage_class'] |
497 | config_params.pop('storage_class') | |
498 | else: | |
499 | storage_class = "mediagoblin.storage:BasicFileStorage" | |
500 | ||
501 | storage_class = util.import_component(storage_class) | |
502 | return storage_class(**config_params) |