|
1 """ |
|
2 Sync Media to S3 |
|
3 ================ |
|
4 |
|
5 Django command that scans all files in your settings.MEDIA_ROOT folder and |
|
6 uploads them to S3 with the same directory structure. |
|
7 |
|
8 This command can optionally do the following but it is off by default: |
|
9 * gzip compress any CSS and Javascript files it finds and adds the appropriate |
|
10 'Content-Encoding' header. |
|
11 * set a far future 'Expires' header for optimal caching. |
|
12 |
|
13 Note: This script requires the Python boto library and valid Amazon Web |
|
14 Services API keys. |
|
15 |
|
16 Required settings.py variables: |
|
17 AWS_ACCESS_KEY_ID = '' |
|
18 AWS_SECRET_ACCESS_KEY = '' |
|
19 AWS_BUCKET_NAME = '' |
|
20 |
|
21 Command options are: |
|
22 -p PREFIX, --prefix=PREFIX |
|
23 The prefix to prepend to the path on S3. |
|
24 --gzip Enables gzipping CSS and Javascript files. |
|
25 --expires Enables setting a far future expires header. |
|
26 --force Skip the file mtime check to force upload of all |
|
27 files. |
|
28 --filter-list Override default directory and file exclusion |
|
29 filters. (enter as comma seperated line) |
|
30 |
|
31 TODO: |
|
32 * Use fnmatch (or regex) to allow more complex FILTER_LIST rules. |
|
33 |
|
34 """ |
|
35 import datetime |
|
36 import email |
|
37 import mimetypes |
|
38 import optparse |
|
39 import os |
|
40 import sys |
|
41 import time |
|
42 |
|
43 from django.conf import settings |
|
44 from django.core.management.base import BaseCommand, CommandError |
|
45 |
|
46 # Make sure boto is available |
|
47 try: |
|
48 import boto |
|
49 import boto.exception |
|
50 except ImportError: |
|
51 raise ImportError, "The boto Python library is not installed." |
|
52 |
|
53 class Command(BaseCommand): |
|
54 |
|
55 # Extra variables to avoid passing these around |
|
56 AWS_ACCESS_KEY_ID = '' |
|
57 AWS_SECRET_ACCESS_KEY = '' |
|
58 AWS_BUCKET_NAME = '' |
|
59 DIRECTORY = '' |
|
60 FILTER_LIST = ['.DS_Store', '.svn', '.hg', '.git', 'Thumbs.db'] |
|
61 GZIP_CONTENT_TYPES = ( |
|
62 'text/css', |
|
63 'application/javascript', |
|
64 'application/x-javascript' |
|
65 ) |
|
66 |
|
67 upload_count = 0 |
|
68 skip_count = 0 |
|
69 |
|
70 option_list = BaseCommand.option_list + ( |
|
71 optparse.make_option('-p', '--prefix', |
|
72 dest='prefix', default='', |
|
73 help="The prefix to prepend to the path on S3."), |
|
74 optparse.make_option('-d', '--dir', |
|
75 dest='dir', default=settings.MEDIA_ROOT, |
|
76 help="The root directory to use instead of your MEDIA_ROOT"), |
|
77 optparse.make_option('--gzip', |
|
78 action='store_true', dest='gzip', default=False, |
|
79 help="Enables gzipping CSS and Javascript files."), |
|
80 optparse.make_option('--expires', |
|
81 action='store_true', dest='expires', default=False, |
|
82 help="Enables setting a far future expires header."), |
|
83 optparse.make_option('--force', |
|
84 action='store_true', dest='force', default=False, |
|
85 help="Skip the file mtime check to force upload of all files."), |
|
86 optparse.make_option('--filter-list', dest='filter_list', |
|
87 action='store', default='', |
|
88 help="Override default directory and file exclusion filters. (enter as comma seperated line)"), |
|
89 ) |
|
90 |
|
91 help = 'Syncs the complete MEDIA_ROOT structure and files to S3 into the given bucket name.' |
|
92 args = 'bucket_name' |
|
93 |
|
94 can_import_settings = True |
|
95 |
|
96 def handle(self, *args, **options): |
|
97 |
|
98 # Check for AWS keys in settings |
|
99 if not hasattr(settings, 'AWS_ACCESS_KEY_ID') or \ |
|
100 not hasattr(settings, 'AWS_SECRET_ACCESS_KEY'): |
|
101 raise CommandError('Missing AWS keys from settings file. Please' + |
|
102 'supply both AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY.') |
|
103 else: |
|
104 self.AWS_ACCESS_KEY_ID = settings.AWS_ACCESS_KEY_ID |
|
105 self.AWS_SECRET_ACCESS_KEY = settings.AWS_SECRET_ACCESS_KEY |
|
106 |
|
107 if not hasattr(settings, 'AWS_BUCKET_NAME'): |
|
108 raise CommandError('Missing bucket name from settings file. Please' + |
|
109 ' add the AWS_BUCKET_NAME to your settings file.') |
|
110 else: |
|
111 if not settings.AWS_BUCKET_NAME: |
|
112 raise CommandError('AWS_BUCKET_NAME cannot be empty.') |
|
113 self.AWS_BUCKET_NAME = settings.AWS_BUCKET_NAME |
|
114 |
|
115 if not hasattr(settings, 'MEDIA_ROOT'): |
|
116 raise CommandError('MEDIA_ROOT must be set in your settings.') |
|
117 else: |
|
118 if not settings.MEDIA_ROOT: |
|
119 raise CommandError('MEDIA_ROOT must be set in your settings.') |
|
120 |
|
121 self.verbosity = int(options.get('verbosity')) |
|
122 self.prefix = options.get('prefix') |
|
123 self.do_gzip = options.get('gzip') |
|
124 self.do_expires = options.get('expires') |
|
125 self.do_force = options.get('force') |
|
126 self.DIRECTORY = options.get('dir') |
|
127 self.FILTER_LIST = getattr(settings, 'FILTER_LIST', self.FILTER_LIST) |
|
128 filter_list = options.get('filter_list').split(',') |
|
129 if filter_list: |
|
130 # command line option overrides default filter_list and |
|
131 # settings.filter_list |
|
132 self.FILTER_LIST = filter_list |
|
133 |
|
134 # Now call the syncing method to walk the MEDIA_ROOT directory and |
|
135 # upload all files found. |
|
136 self.sync_s3() |
|
137 |
|
138 print |
|
139 print "%d files uploaded." % (self.upload_count) |
|
140 print "%d files skipped." % (self.skip_count) |
|
141 |
|
142 def sync_s3(self): |
|
143 """ |
|
144 Walks the media directory and syncs files to S3 |
|
145 """ |
|
146 bucket, key = self.open_s3() |
|
147 os.path.walk(self.DIRECTORY, self.upload_s3, |
|
148 (bucket, key, self.AWS_BUCKET_NAME, self.DIRECTORY)) |
|
149 |
|
150 def compress_string(self, s): |
|
151 """Gzip a given string.""" |
|
152 import cStringIO, gzip |
|
153 zbuf = cStringIO.StringIO() |
|
154 zfile = gzip.GzipFile(mode='wb', compresslevel=6, fileobj=zbuf) |
|
155 zfile.write(s) |
|
156 zfile.close() |
|
157 return zbuf.getvalue() |
|
158 |
|
159 def open_s3(self): |
|
160 """ |
|
161 Opens connection to S3 returning bucket and key |
|
162 """ |
|
163 conn = boto.connect_s3(self.AWS_ACCESS_KEY_ID, self.AWS_SECRET_ACCESS_KEY) |
|
164 try: |
|
165 bucket = conn.get_bucket(self.AWS_BUCKET_NAME) |
|
166 except boto.exception.S3ResponseError: |
|
167 bucket = conn.create_bucket(self.AWS_BUCKET_NAME) |
|
168 return bucket, boto.s3.key.Key(bucket) |
|
169 |
|
170 def upload_s3(self, arg, dirname, names): |
|
171 """ |
|
172 This is the callback to os.path.walk and where much of the work happens |
|
173 """ |
|
174 bucket, key, bucket_name, root_dir = arg # expand arg tuple |
|
175 |
|
176 # Skip directories we don't want to sync |
|
177 if os.path.basename(dirname) in self.FILTER_LIST: |
|
178 # prevent walk from processing subfiles/subdirs below the ignored one |
|
179 del names[:] |
|
180 return |
|
181 |
|
182 # Later we assume the MEDIA_ROOT ends with a trailing slash |
|
183 if not root_dir.endswith(os.path.sep): |
|
184 root_dir = root_dir + os.path.sep |
|
185 |
|
186 for file in names: |
|
187 headers = {} |
|
188 |
|
189 if file in self.FILTER_LIST: |
|
190 continue # Skip files we don't want to sync |
|
191 |
|
192 filename = os.path.join(dirname, file) |
|
193 if os.path.isdir(filename): |
|
194 continue # Don't try to upload directories |
|
195 |
|
196 file_key = filename[len(root_dir):] |
|
197 if self.prefix: |
|
198 file_key = '%s/%s' % (self.prefix, file_key) |
|
199 |
|
200 # Check if file on S3 is older than local file, if so, upload |
|
201 if not self.do_force: |
|
202 s3_key = bucket.get_key(file_key) |
|
203 if s3_key: |
|
204 s3_datetime = datetime.datetime(*time.strptime( |
|
205 s3_key.last_modified, '%a, %d %b %Y %H:%M:%S %Z')[0:6]) |
|
206 local_datetime = datetime.datetime.utcfromtimestamp( |
|
207 os.stat(filename).st_mtime) |
|
208 if local_datetime < s3_datetime: |
|
209 self.skip_count += 1 |
|
210 if self.verbosity > 1: |
|
211 print "File %s hasn't been modified since last " \ |
|
212 "being uploaded" % (file_key) |
|
213 continue |
|
214 |
|
215 # File is newer, let's process and upload |
|
216 if self.verbosity > 0: |
|
217 print "Uploading %s..." % (file_key) |
|
218 |
|
219 content_type = mimetypes.guess_type(filename)[0] |
|
220 if content_type: |
|
221 headers['Content-Type'] = content_type |
|
222 file_obj = open(filename, 'rb') |
|
223 file_size = os.fstat(file_obj.fileno()).st_size |
|
224 filedata = file_obj.read() |
|
225 if self.do_gzip: |
|
226 # Gzipping only if file is large enough (>1K is recommended) |
|
227 # and only if file is a common text type (not a binary file) |
|
228 if file_size > 1024 and content_type in self.GZIP_CONTENT_TYPES: |
|
229 filedata = self.compress_string(filedata) |
|
230 headers['Content-Encoding'] = 'gzip' |
|
231 if self.verbosity > 1: |
|
232 print "\tgzipped: %dk to %dk" % \ |
|
233 (file_size/1024, len(filedata)/1024) |
|
234 if self.do_expires: |
|
235 # HTTP/1.0 |
|
236 headers['Expires'] = '%s GMT' % (email.Utils.formatdate( |
|
237 time.mktime((datetime.datetime.now() + |
|
238 datetime.timedelta(days=365*2)).timetuple()))) |
|
239 # HTTP/1.1 |
|
240 headers['Cache-Control'] = 'max-age %d' % (3600 * 24 * 365 * 2) |
|
241 if self.verbosity > 1: |
|
242 print "\texpires: %s" % (headers['Expires']) |
|
243 print "\tcache-control: %s" % (headers['Cache-Control']) |
|
244 |
|
245 try: |
|
246 key.name = file_key |
|
247 key.set_contents_from_string(filedata, headers, replace=True) |
|
248 key.set_acl('public-read') |
|
249 except boto.s3.connection.S3CreateError, e: |
|
250 print "Failed: %s" % e |
|
251 except Exception, e: |
|
252 print e |
|
253 raise |
|
254 else: |
|
255 self.upload_count += 1 |
|
256 |
|
257 file_obj.close() |
|
258 |
|
259 # Backwards compatibility for Django r9110 |
|
260 if not [opt for opt in Command.option_list if opt.dest=='verbosity']: |
|
261 Command.option_list += ( |
|
262 optparse.make_option('-v', '--verbosity', |
|
263 dest='verbosity', default=1, action='count', |
|
264 help="Verbose mode. Multiple -v options increase the verbosity."), |
|
265 ) |