mediagoblin/media_types/image/processing.py

   1 # GNU MediaGoblin -- federated, autonomous media hosting
   2 # Copyright (C) 2011, 2012 MediaGoblin contributors.  See AUTHORS.
   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
  17 try:
  18     from PIL import Image
  19 except ImportError:
  20     import Image
  21 import os
  22 import logging
  23 import argparse
  24
  25 from mediagoblin import mg_globals as mgg
  26 from mediagoblin.db.models import MediaEntry
  27 from mediagoblin.processing import (
  28     BadMediaFail, FilenameBuilder,
  29     MediaProcessor, ProcessingManager,
  30     request_from_args)
  31 from mediagoblin.submit.lib import run_process_media
  32 from mediagoblin.tools.exif import exif_fix_image_orientation, \
  33     extract_exif, clean_exif, get_gps_data, get_useful, \
  34     exif_image_needs_rotation
  35 from mediagoblin.tools.translate import lazy_pass_to_ugettext as _
  36
  37 _log = logging.getLogger(__name__)
  38
  39 PIL_FILTERS = {
  40     'NEAREST': Image.NEAREST,
  41     'BILINEAR': Image.BILINEAR,
  42     'BICUBIC': Image.BICUBIC,
  43     'ANTIALIAS': Image.ANTIALIAS}
  44
  45 MEDIA_TYPE = 'mediagoblin.media_types.image'
  46
  47
  48 def resize_image(proc_state, resized, keyname, target_name, new_size,
  49                  exif_tags, workdir):
  50     """
  51     Store a resized version of an image and return its pathname.
  52
  53     Arguments:
  54     proc_state -- the processing state for the image to resize
  55     resized -- an image from Image.open() of the original image being resized
  56     keyname -- Under what key to save in the db.
  57     target_name -- public file path for the new resized image
  58     exif_tags -- EXIF data for the original image
  59     workdir -- directory path for storing converted image files
  60     new_size -- 2-tuple size for the resized image
  61     """
  62     config = mgg.global_config['media_type:mediagoblin.media_types.image']
  63
  64     resized = exif_fix_image_orientation(resized, exif_tags)  # Fix orientation
  65
  66     filter_config = config['resize_filter']
  67     try:
  68         resize_filter = PIL_FILTERS[filter_config.upper()]
  69     except KeyError:
  70         raise Exception('Filter "{0}" not found, choose one of {1}'.format(
  71             unicode(filter_config),
  72             u', '.join(PIL_FILTERS.keys())))
  73
  74     resized.thumbnail(new_size, resize_filter)
  75
  76     # Copy the new file to the conversion subdir, then remotely.
  77     tmp_resized_filename = os.path.join(workdir, target_name)
  78     with file(tmp_resized_filename, 'w') as resized_file:
  79         resized.save(resized_file, quality=config['quality'])
  80     proc_state.store_public(keyname, tmp_resized_filename, target_name)
  81
  82
  83 def resize_tool(proc_state, force, keyname, target_name,
  84                 conversions_subdir, exif_tags, new_size=None):
  85     # filename -- the filename of the original image being resized
  86     filename = proc_state.get_orig_filename()
  87
  88     # Use the default size if new_size was not given
  89     if not new_size:
  90         max_width = mgg.global_config['media:' + keyname]['max_width']
  91         max_height = mgg.global_config['media:' + keyname]['max_height']
  92         new_size = (max_width, max_height)
  93
  94     # If the size of the original file exceeds the specified size for the desized
  95     # file, a target_name file is created and later associated with the media
  96     # entry.
  97     # Also created if the file needs rotation, or if forced.
  98     try:
  99         im = Image.open(filename)
 100     except IOError:
 101         raise BadMediaFail()
 102     if force \
 103         or im.size[0] > new_size[0]\
 104         or im.size[1] > new_size[1]\
 105         or exif_image_needs_rotation(exif_tags):
 106         resize_image(
 107             proc_state, im, unicode(keyname), target_name,
 108             new_size,
 109             exif_tags, conversions_subdir)
 110
 111
 112 SUPPORTED_FILETYPES = ['png', 'gif', 'jpg', 'jpeg', 'tiff']
 113
 114
 115 def sniff_handler(media_file, **kw):
 116     _log.info('Sniffing {0}'.format(MEDIA_TYPE))
 117     if kw.get('media') is not None:  # That's a double negative!
 118         name, ext = os.path.splitext(kw['media'].filename)
 119         clean_ext = ext[1:].lower()  # Strip the . from ext and make lowercase
 120
 121         if clean_ext in SUPPORTED_FILETYPES:
 122             _log.info('Found file extension in supported filetypes')
 123             return MEDIA_TYPE
 124         else:
 125             _log.debug('Media present, extension not found in {0}'.format(
 126                     SUPPORTED_FILETYPES))
 127     else:
 128         _log.warning('Need additional information (keyword argument \'media\')'
 129                      ' to be able to handle sniffing')
 130
 131     return None
 132
 133
 134 class ProcessImage(object):
 135     """Code to process an image. Will be run by celery.
 136
 137     A Workbench() represents a local tempory dir. It is automatically
 138     cleaned up when this function exits.
 139     """
 140     def __init__(self, proc_state=None):
 141         if proc_state:
 142             self.proc_state = proc_state
 143             self.entry = proc_state.entry
 144             self.workbench = proc_state.workbench
 145
 146             # Conversions subdirectory to avoid collisions
 147             self.conversions_subdir = os.path.join(
 148                 self.workbench.dir, 'convirsions')
 149
 150             self.orig_filename = proc_state.get_orig_filename()
 151             self.name_builder = FilenameBuilder(self.orig_filename)
 152
 153             # Exif extraction
 154             self.exif_tags = extract_exif(self.orig_filename)
 155
 156             os.mkdir(self.conversions_subdir)
 157
 158     def reprocess_action(self, args):
 159         """
 160         List the available actions for media in a given state
 161         """
 162         if args[0].state == 'processed':
 163             print _('\n Available reprocessing actions for processed images:'
 164                     '\n \t --resize: thumb or medium'
 165                     '\n Options:'
 166                     '\n \t --size: max_width max_height (defaults to'
 167                     'config specs)')
 168             return True
 169
 170     def _parser(self, args):
 171         """
 172         Parses the unknown args from the gmg parser
 173         """
 174         parser = argparse.ArgumentParser()
 175         parser.add_argument(
 176             '--resize',
 177             choices=['thumb', 'medium'])
 178         parser.add_argument(
 179             '--size',
 180             nargs=2,
 181             metavar=('max_width', 'max_height'),
 182             type=int)
 183         parser.add_argument(
 184             '--initial_processing',
 185             action='store_true')
 186
 187         return parser.parse_args(args[1])
 188
 189     def _check_eligible(self, entry_args, reprocess_args):
 190         """
 191         Check to see if we can actually process the given media as requested
 192         """
 193
 194         if entry_args.state == 'processed':
 195             if reprocess_args.initial_processing:
 196                 raise Exception(_('You can not run --initial_processing on'
 197                                   ' media that has already been processed.'))
 198
 199         if entry_args.state == 'failed':
 200             if reprocess_args.resize:
 201                 raise Exception(_('You can not run --resize on media that has'
 202                                   ' not been processed.'))
 203             if reprocess_args.size:
 204                 _log.warn('With --initial_processing, the --size flag will be'
 205                           ' ignored.')
 206
 207         if entry_args.state == 'processing':
 208             raise Exception(_('We currently do not support reprocessing on'
 209                               ' media that is in the "processing" state.'))
 210
 211     def initial_processing(self):
 212         # Is there any GPS data
 213         gps_data = get_gps_data(self.exif_tags)
 214
 215          # Always create a small thumbnail
 216         resize_tool(self.proc_state, True, 'thumb', self.orig_filename,
 217                     self.name_builder.fill('{basename}.thumbnail{ext}'),
 218                     self.conversions_subdir, self.exif_tags)
 219
 220         # Possibly create a medium
 221         resize_tool(self.proc_state, False, 'medium', self.orig_filename,
 222                     self.name_builder.fill('{basename}.medium{ext}'),
 223                     self.conversions_subdir, self.exif_tags)
 224
 225         # Copy our queued local workbench to its final destination
 226         self.proc_state.copy_original(self.name_builder.fill('{basename}{ext}'))
 227
 228         # Remove queued media file from storage and database
 229         self.proc_state.delete_queue_file()
 230
 231         # Insert exif data into database
 232         exif_all = clean_exif(self.exif_tags)
 233
 234         if len(exif_all):
 235             self.entry.media_data_init(exif_all=exif_all)
 236
 237         if len(gps_data):
 238             for key in list(gps_data.keys()):
 239                 gps_data['gps_' + key] = gps_data.pop(key)
 240             self.entry.media_data_init(**gps_data)
 241
 242     def reprocess(self, reprocess_info):
 243         """
 244         This function actually does the reprocessing when called by
 245         ProcessMedia in gmg/processing/task.py
 246         """
 247         new_size = None
 248
 249         # Did they specify a size? They must specify either both or none, so
 250         # we only need to check if one is present
 251         if reprocess_info.get('max_width'):
 252             max_width = reprocess_info['max_width']
 253             max_height = reprocess_info['max_height']
 254
 255             new_size = (max_width, max_height)
 256
 257         resize_tool(self.proc_state, False, reprocess_info['resize'],
 258                     self.name_builder.fill('{basename}.medium{ext}'),
 259                     self.conversions_subdir, self.exif_tags, new_size)
 260
 261     def media_reprocess(self, args):
 262         """
 263         This function handles the all of the reprocessing logic, before calling
 264         gmg/submit/lib/run_process_media
 265         """
 266         reprocess_args = self._parser(args)
 267         entry_args = args[0]
 268
 269         # Can we actually process the given media as requested?
 270         self._check_eligible(entry_args, reprocess_args)
 271
 272         # Do we want to re-try initial processing?
 273         if reprocess_args.initial_processing:
 274             for id in entry_args.media_id:
 275                 entry = MediaEntry.query.filter_by(id=id).first()
 276                 run_process_media(entry)
 277
 278         # Are we wanting to resize the thumbnail or medium?
 279         elif reprocess_args.resize:
 280
 281             # reprocess all given media entries
 282             for id in entry_args.media_id:
 283                 entry = MediaEntry.query.filter_by(id=id).first()
 284
 285                 # For now we can only reprocess with the original file
 286                 if not entry.media_files.get('original'):
 287                     raise Exception(_('The original file for this media entry'
 288                                       ' does not exist.'))
 289
 290                 reprocess_info = self._get_reprocess_info(reprocess_args)
 291                 run_process_media(entry, reprocess_info=reprocess_info)
 292
 293         # If we are here, they forgot to tell us how to reprocess
 294         else:
 295             _log.warn('You must set either --resize or --initial_processing'
 296                       ' flag to reprocess an image.')
 297
 298     def _get_reprocess_info(self, args):
 299         """ Returns a dict with the info needed for reprocessing"""
 300         reprocess_info = {'resize': args.resize}
 301
 302         if args.size:
 303             reprocess_info['max_width'] = args.size[0]
 304             reprocess_info['max_height'] = args.size[1]
 305
 306         return reprocess_info
 307
 308
 309 class CommonImageProcessor(MediaProcessor):
 310     """
 311     Provides a base for various media processing steps
 312     """
 313     # Common resizing step
 314     def resize_step(self):
 315         pass
 316
 317     @classmethod
 318     def _add_width_height_args(cls, parser):
 319         parser.add_argument(
 320             "--width", default=None,
 321             help=(
 322                 "Width of the resized image (if not using defaults)"))
 323         parser.add_argument(
 324             "--height", default=None,
 325             help=(
 326                 "Height of the resized image (if not using defaults)"))
 327
 328
 329 class InitialProcessor(CommonImageProcessor):
 330     """
 331     Initial processing step for new images
 332     """
 333     name = "initial"
 334     description = "Initial processing"
 335
 336     @classmethod
 337     def media_is_eligible(cls, media_entry):
 338         """
 339         Determine if this media type is eligible for processing
 340         """
 341         return media_entry.state in (
 342             "unprocessed", "failed")
 343
 344     ###############################
 345     # Command line interface things
 346     ###############################
 347
 348     @classmethod
 349     def generate_parser(cls):
 350         parser = argparse.ArgumentParser(
 351             description=cls.description,
 352             prog=cls.name)
 353
 354         cls._add_width_height_args(parser)
 355
 356         return parser
 357
 358     @classmethod
 359     def args_to_request(cls, args):
 360         return request_from_args(
 361             args, ['width', 'height'])
 362
 363
 364
 365 class ImageProcessingManager(ProcessingManager):
 366     def __init__(self):
 367         super(self.__class__, self).__init__()
 368         self.add_processor(InitialProcessor)
 369
 370
 371 if __name__ == '__main__':
 372     import sys
 373     import pprint
 374
 375     pp = pprint.PrettyPrinter()
 376
 377     result = extract_exif(sys.argv[1])
 378     gps = get_gps_data(result)
 379     clean = clean_exif(result)
 380     useful = get_useful(clean)
 381
 382     print pp.pprint(
 383         clean)