# GNU MediaGoblin -- federated, autonomous media hosting
-# Copyright (C) 2011 MediaGoblin contributors. See AUTHORS.
+# Copyright (C) 2011, 2012 MediaGoblin contributors. See AUTHORS.
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published by
# along with this program. If not, see <http://www.gnu.org/licenses/>.
import Image
+import os
+import logging
-from celery.task import Task
-from celery import registry
-
-from mediagoblin.db.util import ObjectId
from mediagoblin import mg_globals as mgg
+from mediagoblin.decorators import get_workbench
+from mediagoblin.processing import BadMediaFail, \
+ create_pub_filepath, FilenameBuilder
+from mediagoblin.tools.exif import exif_fix_image_orientation, \
+ extract_exif, clean_exif, get_gps_data, get_useful, \
+ exif_image_needs_rotation
-from mediagoblin.util import lazy_pass_to_ugettext as _
-
-from mediagoblin.process_media.errors import *
-
-THUMB_SIZE = 180, 180
-MEDIUM_SIZE = 640, 640
-
+_log = logging.getLogger(__name__)
-def create_pub_filepath(entry, filename):
- return mgg.public_store.get_unique_filepath(
- ['media_entries',
- unicode(entry['_id']),
- filename])
-################################
-# Media processing initial steps
-################################
-
-class ProcessMedia(Task):
+def resize_image(entry, filename, new_path, exif_tags, workdir, new_size,
+ size_limits=(0, 0)):
"""
- Pass this entry off for processing.
+ Store a resized version of an image and return its pathname.
+
+ Arguments:
+ entry -- the entry for the image to resize
+ filename -- the filename of the original image being resized
+ new_path -- public file path for the new resized image
+ exif_tags -- EXIF data for the original image
+ workdir -- directory path for storing converted image files
+ new_size -- 2-tuple size for the resized image
"""
- def run(self, media_id):
- """
- Pass the media entry off to the appropriate processing function
- (for now just process_image...)
- """
- entry = mgg.database.MediaEntry.one(
- {'_id': ObjectId(media_id)})
-
- # Try to process, and handle expected errors.
- try:
- process_image(entry)
- except BaseProcessingFail, exc:
- mark_entry_failed(entry[u'_id'], exc)
- return
+ try:
+ resized = Image.open(filename)
+ except IOError:
+ raise BadMediaFail()
+ resized = exif_fix_image_orientation(resized, exif_tags) # Fix orientation
+ resized.thumbnail(new_size, Image.ANTIALIAS)
- entry['state'] = u'processed'
- entry.save()
+ # Copy the new file to the conversion subdir, then remotely.
+ tmp_resized_filename = os.path.join(workdir, new_path[-1])
+ with file(tmp_resized_filename, 'w') as resized_file:
+ resized.save(resized_file)
+ mgg.public_store.copy_local_to_storage(tmp_resized_filename, new_path)
- def on_failure(self, exc, task_id, args, kwargs, einfo):
- """
- If the processing failed we should mark that in the database.
- Assuming that the exception raised is a subclass of BaseProcessingFail,
- we can use that to get more information about the failure and store that
- for conveying information to users about the failure, etc.
- """
- entry_id = args[0]
- mark_entry_failed(entry_id, exc)
+SUPPORTED_FILETYPES = ['png', 'gif', 'jpg', 'jpeg']
-process_media = registry.tasks[ProcessMedia.name]
+def sniff_handler(media_file, **kw):
+ if kw.get('media') is not None: # That's a double negative!
+ name, ext = os.path.splitext(kw['media'].filename)
+ clean_ext = ext[1:].lower() # Strip the . from ext and make lowercase
+ if clean_ext in SUPPORTED_FILETYPES:
+ _log.info('Found file extension in supported filetypes')
+ return True
+ else:
+ _log.debug('Media present, extension not found in {0}'.format(
+ SUPPORTED_FILETYPES))
+ else:
+ _log.warning('Need additional information (keyword argument \'media\')'
+ ' to be able to handle sniffing')
-def mark_entry_failed(entry_id, exc):
- """
- Mark a media entry as having failed in its conversion.
+ return False
- Uses the exception that was raised to mark more information. If the
- exception is a derivative of BaseProcessingFail then we can store extra
- information that can be useful for users telling them why their media failed
- to process.
- Args:
- - entry_id: The id of the media entry
+@get_workbench
+def process_image(entry, workbench=None):
+ """Code to process an image. Will be run by celery.
+ A Workbench() represents a local tempory dir. It is automatically
+ cleaned up when this function exits.
"""
- # Was this a BaseProcessingFail? In other words, was this a
- # type of error that we know how to handle?
- if isinstance(exc, BaseProcessingFail):
- # Looks like yes, so record information about that failure and any
- # metadata the user might have supplied.
- mgg.database['media_entries'].update(
- {'_id': entry_id},
- {'$set': {u'state': u'failed',
- u'fail_error': exc.exception_path,
- u'fail_metadata': exc.metadata}})
- else:
- # Looks like no, so just mark it as failed and don't record a
- # failure_error (we'll assume it wasn't handled) and don't record
- # metadata (in fact overwrite it if somehow it had previous info
- # here)
- mgg.database['media_entries'].update(
- {'_id': entry_id},
- {'$set': {u'state': u'failed',
- u'fail_error': None,
- u'fail_metadata': {}}})
-
-
-def process_image(entry):
- """
- Code to process an image
- """
- workbench = mgg.workbench_manager.create_workbench()
-
- queued_filepath = entry['queued_media_file']
+ # Conversions subdirectory to avoid collisions
+ conversions_subdir = os.path.join(
+ workbench.dir, 'conversions')
+ os.mkdir(conversions_subdir)
+ queued_filepath = entry.queued_media_file
queued_filename = workbench.localized_file(
mgg.queue_store, queued_filepath,
'source')
+ name_builder = FilenameBuilder(queued_filename)
- try:
- thumb = Image.open(queued_filename)
- except IOError:
- raise BadMediaFail()
+ # EXIF extraction
+ exif_tags = extract_exif(queued_filename)
+ gps_data = get_gps_data(exif_tags)
- thumb.thumbnail(THUMB_SIZE, Image.ANTIALIAS)
- # ensure color mode is compatible with jpg
- if thumb.mode != "RGB":
- thumb = thumb.convert("RGB")
-
- thumb_filepath = create_pub_filepath(entry, 'thumbnail.jpg')
- thumb_file = mgg.public_store.get_file(thumb_filepath, 'w')
-
- with thumb_file:
- thumb.save(thumb_file, "JPEG", quality=90)
+ # Always create a small thumbnail
+ thumb_filepath = create_pub_filepath(
+ entry, name_builder.fill('{basename}.thumbnail{ext}'))
+ resize_image(entry, queued_filename, thumb_filepath,
+ exif_tags, conversions_subdir,
+ (mgg.global_config['media:thumb']['max_width'],
+ mgg.global_config['media:thumb']['max_height']))
# If the size of the original file exceeds the specified size of a `medium`
- # file, a `medium.jpg` files is created and later associated with the media
+ # file, a `.medium.jpg` files is created and later associated with the media
# entry.
medium = Image.open(queued_filename)
- medium_processed = False
+ if medium.size[0] > mgg.global_config['media:medium']['max_width'] \
+ or medium.size[1] > mgg.global_config['media:medium']['max_height'] \
+ or exif_image_needs_rotation(exif_tags):
+ medium_filepath = create_pub_filepath(
+ entry, name_builder.fill('{basename}.medium{ext}'))
+ resize_image(
+ entry, queued_filename, medium_filepath,
+ exif_tags, conversions_subdir,
+ (mgg.global_config['media:medium']['max_width'],
+ mgg.global_config['media:medium']['max_height']))
+ else:
+ medium_filepath = None
- if medium.size[0] > MEDIUM_SIZE[0] or medium.size[1] > MEDIUM_SIZE[1]:
- medium.thumbnail(MEDIUM_SIZE, Image.ANTIALIAS)
+ # Copy our queued local workbench to its final destination
+ original_filepath = create_pub_filepath(
+ entry, name_builder.fill('{basename}{ext}'))
+ mgg.public_store.copy_local_to_storage(queued_filename, original_filepath)
- if medium.mode != "RGB":
- medium = medium.convert("RGB")
+ # Remove queued media file from storage and database
+ mgg.queue_store.delete_file(queued_filepath)
+ entry.queued_media_file = []
- medium_filepath = create_pub_filepath(entry, 'medium.jpg')
- medium_file = mgg.public_store.get_file(medium_filepath, 'w')
+ # Insert media file information into database
+ media_files_dict = entry.setdefault('media_files', {})
+ media_files_dict[u'thumb'] = thumb_filepath
+ media_files_dict[u'original'] = original_filepath
+ if medium_filepath:
+ media_files_dict[u'medium'] = medium_filepath
- with medium_file:
- medium.save(medium_file, "JPEG", quality=90)
- medium_processed = True
+ # Insert exif data into database
+ exif_all = clean_exif(exif_tags)
- # we have to re-read because unlike PIL, not everything reads
- # things in string representation :)
- queued_file = file(queued_filename, 'rb')
+ if len(exif_all):
+ entry.media_data_init(exif_all=exif_all)
- with queued_file:
- original_filepath = create_pub_filepath(entry, queued_filepath[-1])
+ if len(gps_data):
+ for key in list(gps_data.keys()):
+ gps_data['gps_' + key] = gps_data.pop(key)
+ entry.media_data_init(**gps_data)
- with mgg.public_store.get_file(original_filepath, 'wb') as original_file:
- original_file.write(queued_file.read())
- mgg.queue_store.delete_file(queued_filepath)
- entry['queued_media_file'] = []
- media_files_dict = entry.setdefault('media_files', {})
- media_files_dict['thumb'] = thumb_filepath
- media_files_dict['original'] = original_filepath
- if medium_processed:
- media_files_dict['medium'] = medium_filepath
+if __name__ == '__main__':
+ import sys
+ import pprint
+
+ pp = pprint.PrettyPrinter()
+
+ result = extract_exif(sys.argv[1])
+ gps = get_gps_data(result)
+ clean = clean_exif(result)
+ useful = get_useful(clean)
- # clean up workbench
- workbench.destroy_self()
+ print pp.pprint(
+ clean)