2 # -*- coding: utf-8 -*-
4 # This file is part of my.gpodder.org.
6 # my.gpodder.org is free software: you can redistribute it and/or modify it
7 # under the terms of the GNU Affero General Public License as published by
8 # the Free Software Foundation, either version 3 of the License, or (at your
9 # option) any later version.
11 # my.gpodder.org is distributed in the hope that it will be useful, but
12 # WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
13 # or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public
14 # License for more details.
16 # You should have received a copy of the GNU Affero General Public License
17 # along with my.gpodder.org. If not, see <http://www.gnu.org/licenses/>.
22 from urlparse
import urljoin
25 from datetime
import datetime
, timedelta
26 from itertools
import chain
, islice
30 from django
.db
import transaction
31 from django
.conf
import settings
33 from mygpo
.podcasts
.models
import Podcast
, URL
, Slug
, Episode
34 from mygpo
.core
.slugs
import assign_missing_episode_slugs
, PodcastSlug
35 from mygpo
.podcasts
.models
import DEFAULT_UPDATE_INTERVAL
, \
36 MIN_UPDATE_INTERVAL
, MAX_UPDATE_INTERVAL
37 from mygpo
.utils
import file_hash
, to_maxlength
38 from mygpo
.web
.logo
import CoverArt
39 from mygpo
.data
.podcast
import subscribe_at_hub
40 from mygpo
.data
.tasks
import update_related_podcasts
41 from mygpo
.pubsub
.models
import SubscriptionError
42 from mygpo
.directory
.tags
import update_category
45 logger
= logging
.getLogger(__name__
)
47 MAX_EPISODES_UPDATE
= 200
50 class UpdatePodcastException(Exception):
54 class NoPodcastCreated(Exception):
55 """ raised when no podcast obj was created for a new URL """
58 class NoEpisodesException(Exception):
59 """ raised when parsing something that doesn't contain any episodes """
62 def update_podcasts(queue
):
63 """ Fetch data for the URLs supplied as the queue iterable """
65 for n
, podcast_url
in enumerate(queue
, 1):
66 logger
.info('Update %d - %s', n
, podcast_url
)
68 yield update_podcast(podcast_url
)
70 except NoPodcastCreated
as npc
:
71 logger
.info('No podcast created: %s', npc
)
74 logger
.exception('Error while updating podcast "%s"',
79 def update_podcast(podcast_url
):
80 """ Update the podcast for the supplied URL """
83 parsed
= _fetch_feed(podcast_url
)
84 _validate_parsed(parsed
)
86 except requests
.exceptions
.RequestException
as re
:
87 logging
.exception('Error while fetching response from feedservice')
90 except NoEpisodesException
as nee
:
91 logging
.warn('No episode found while parsing podcast')
93 # if we fail to parse the URL, we don't even create the
96 p
= Podcast
.objects
.get(urls__url
=podcast_url
)
97 # if it exists already, we mark it as outdated
98 _mark_outdated(p
, 'error while fetching feed: %s' % str(nee
))
101 except Podcast
.DoesNotExist
:
102 raise NoPodcastCreated(nee
)
104 assert parsed
, 'fetch_feed must return something'
105 p
= Podcast
.objects
.get_or_create_for_url(podcast_url
)
106 episodes
= _update_episodes(p
, parsed
.get('episodes', []))
107 max_episode_order
= _order_episodes(p
)
108 _update_podcast(p
, parsed
, episodes
, max_episode_order
)
112 def verify_podcast_url(podcast_url
):
113 parsed
= _fetch_feed(podcast_url
)
114 _validate_parsed(parsed
)
118 def _fetch_feed(podcast_url
):
119 params
= {'url': podcast_url
}
121 'Accept': 'application/json',
123 # markdown and other parameters?
124 url
= urljoin(settings
.FEEDSERVICE_URL
, 'parse')
125 r
= requests
.get(url
, params
=params
, headers
=headers
, timeout
=10)
129 def _validate_parsed(parsed
):
130 """ validates the parsed results and raises an exception if invalid
132 feedparser parses pretty much everything. We reject anything that
133 doesn't look like a feed"""
135 if not parsed
or not parsed
.get('episodes', []):
136 raise NoEpisodesException('no episodes found')
139 def _update_podcast(podcast
, parsed
, episodes
, max_episode_order
):
140 """ updates a podcast according to new parser results """
142 # we need that later to decide if we can "bump" a category
143 prev_latest_episode_timestamp
= podcast
.latest_episode_timestamp
145 podcast
.title
= parsed
.get('title') or podcast
.title
146 podcast
.description
= parsed
.get('description') or podcast
.description
147 podcast
.subtitle
= parsed
.get('subtitle') or podcast
.subtitle
148 podcast
.link
= parsed
.get('link') or podcast
.link
149 podcast
.logo_url
= parsed
.get('logo') or podcast
.logo_url
150 podcast
.author
= to_maxlength(Podcast
, 'author', parsed
.get('author') or
152 podcast
.language
= to_maxlength(Podcast
, 'language',
153 parsed
.get('language') or podcast
.language
)
154 podcast
.content_types
= ','.join(parsed
.get('content_types')) or \
155 podcast
.content_types
156 #podcast.tags['feed'] = parsed.tags or podcast.tags.get('feed', [])
157 podcast
.common_episode_title
= to_maxlength(
159 'common_episode_title',
160 parsed
.get('common_title') or podcast
.common_episode_title
)
161 podcast
.new_location
= parsed
.get('new_location') or podcast
.new_location
162 podcast
.flattr_url
= to_maxlength(Podcast
, 'flattr_url',
163 parsed
.get('flattr') or
165 podcast
.hub
= parsed
.get('hub') or podcast
.hub
166 podcast
.license
= parsed
.get('license') or podcast
.license
167 podcast
.max_episode_order
= max_episode_order
169 podcast
.add_missing_urls(parsed
.get('urls', []))
171 if podcast
.new_location
:
173 new_podcast
= Podcast
.objects
.get(urls__url
=podcast
.new_location
)
174 if new_podcast
!= podcast
:
175 _mark_outdated(podcast
, 'redirected to different podcast')
177 except Podcast
.DoesNotExist
:
178 podcast
.set_url(podcast
.new_location
)
180 # latest episode timestamp
181 episodes
= Episode
.objects
.filter(podcast
=podcast
,
182 released__isnull
=False)\
183 .order_by('released')
185 podcast
.update_interval
= get_update_interval(episodes
)
187 latest_episode
= episodes
.last()
189 podcast
.latest_episode_timestamp
= latest_episode
.released
191 # podcast.episode_count is not update here on purpose. It is, instead,
192 # continuously updated when creating new episodes in
193 # EpisodeManager.get_or_create_for_url
195 _update_categories(podcast
, prev_latest_episode_timestamp
)
197 # try to download the logo and reset logo_url to None on http errors
198 found
= _save_podcast_logo(podcast
.logo_url
)
200 podcast
.logo_url
= None
202 # The podcast is always saved (not just when there are changes) because
203 # we need to record the last update
204 logger
.info('Saving podcast.')
205 podcast
.last_update
= datetime
.utcnow()
209 subscribe_at_hub(podcast
)
210 except SubscriptionError
as se
:
211 logger
.warn('subscribing to hub failed: %s', str(se
))
214 slug
= PodcastSlug(podcast
).get_slug()
216 podcast
.add_slug(slug
)
218 assign_missing_episode_slugs(podcast
)
219 update_related_podcasts
.delay(podcast
)
222 def _update_categories(podcast
, prev_timestamp
):
223 """ checks some practical requirements and updates a category """
225 max_timestamp
= datetime
.utcnow() + timedelta(days
=1)
228 if not podcast
.latest_episode_timestamp
:
232 if prev_timestamp
and podcast
.latest_episode_timestamp
<= prev_timestamp
:
235 # too far in the future
236 if podcast
.latest_episode_timestamp
> max_timestamp
:
239 # not enough subscribers
240 if podcast
.subscriber_count() < settings
.MIN_SUBSCRIBERS_CATEGORY
:
243 update_category(podcast
)
246 def _update_episodes(podcast
, parsed_episodes
):
248 pid
= podcast
.get_id()
250 # list of (obj, fun) where fun is the function to update obj
251 updated_episodes
= []
252 episodes_to_update
= list(islice(parsed_episodes
, 0, MAX_EPISODES_UPDATE
))
253 logger
.info('Parsed %d (%d) episodes', len(parsed_episodes
),
254 len(episodes_to_update
))
256 logger
.info('Updating %d episodes', len(episodes_to_update
))
257 for n
, parsed
in enumerate(episodes_to_update
, 1):
259 url
= get_episode_url(parsed
)
261 logger
.info('Skipping episode %d for missing URL', n
)
264 logger
.info('Updating episode %d / %d', n
, len(parsed_episodes
))
266 episode
= Episode
.objects
.get_or_create_for_url(podcast
, url
)
268 update_episode(parsed
, episode
, podcast
)
269 updated_episodes
.append(episode
)
271 # and mark the remaining ones outdated
272 current_episodes
= Episode
.objects
.filter(podcast
=podcast
,
273 outdated
=False)[:500]
274 outdated_episodes
= set(current_episodes
) - set(updated_episodes
)
276 logger
.info('Marking %d episodes as outdated', len(outdated_episodes
))
277 for episode
in outdated_episodes
:
278 mark_outdated(episode
)
282 def _order_episodes(podcast
):
283 """ Reorder the podcast's episode according to release timestamp
285 Returns the highest order value (corresponding to the most recent
288 num_episodes
= podcast
.episode_count
292 episodes
= podcast
.episode_set
.all().extra(select
={
293 'has_released': 'released IS NOT NULL',
295 .order_by('-has_released', '-released', 'pk')\
298 for n
, episode
in enumerate(episodes
.iterator(), 1):
299 # assign ``order`` from higher (most recent) to 0 (oldest)
300 # None means "unknown"
301 new_order
= num_episodes
- n
303 # optimize for new episodes that are newer than all existing
304 if episode
.order
== new_order
:
307 logger
.info('Updating order from {} to {}'.format(episode
.order
,
309 episode
.order
= new_order
312 return num_episodes
- 1
315 def _save_podcast_logo(cover_art
):
320 image_sha1
= hashlib
.sha1(cover_art
).hexdigest()
321 prefix
= CoverArt
.get_prefix(image_sha1
)
323 filename
= CoverArt
.get_original(prefix
, image_sha1
)
324 dirname
= CoverArt
.get_dir(filename
)
326 # get hash of existing file
327 if os
.path
.exists(filename
):
328 with
open(filename
) as f
:
329 old_hash
= file_hash(f
).digest()
333 logger
.info('Logo %s', cover_art
)
336 with
open(filename
, 'w') as fp
:
337 fp
.write(urllib2
.urlopen(cover_art
).read())
339 # get hash of new file
340 with
open(filename
) as f
:
341 new_hash
= file_hash(f
).digest()
343 # remove thumbnails if cover changed
344 if old_hash
!= new_hash
:
345 thumbnails
= CoverArt
.get_existing_thumbnails(prefix
, filename
)
346 logger
.info('Removing %d thumbnails', len(thumbnails
))
352 except (urllib2
.HTTPError
, urllib2
.URLError
, ValueError,
353 httplib
.BadStatusLine
, socket
.error
, IOError) as e
:
354 logger
.warn('Exception while updating podcast logo: %s', str(e
))
357 def _mark_outdated(podcast
, msg
=''):
358 logger
.info('marking podcast outdated: %s', msg
)
359 podcast
.outdated
= True
360 podcast
.last_update
= datetime
.utcnow()
362 _update_episodes(podcast
, [])
365 def get_episode_url(parsed_episode
):
366 """ returns the URL of a parsed episode """
367 for f
in parsed_episode
.get('files', []):
368 if f
.get('urls', []):
373 def update_episode(parsed_episode
, episode
, podcast
):
374 """ updates "episode" with the data from "parsed_episode" """
376 # TODO: check if there have been any changes, to avoid unnecessary updates
377 episode
.guid
= to_maxlength(Episode
, 'guid', parsed_episode
.get('guid') or
379 episode
.description
= parsed_episode
.get('description') or \
381 episode
.subtitle
= parsed_episode
.get('subtitle') or episode
.subtitle
382 episode
.content
= parsed_episode
.get('content') or \
383 parsed_episode
.get('description') or episode
.content
384 episode
.link
= to_maxlength(Episode
, 'link',
385 parsed_episode
.get('link') or episode
.link
)
386 episode
.released
= datetime
.utcfromtimestamp(
387 parsed_episode
.get('released')) if parsed_episode
.get('released') \
388 else episode
.released
389 episode
.author
= to_maxlength(Episode
, 'author',
390 parsed_episode
.get('author') or
392 episode
.duration
= parsed_episode
.get('duration') or episode
.duration
393 episode
.filesize
= parsed_episode
['files'][0]['filesize']
394 episode
.language
= parsed_episode
.get('language') or \
395 episode
.language
or podcast
.language
396 episode
.mimetypes
= ','.join(list(set(
397 filter(None, [f
['mimetype'] for f
in parsed_episode
.get('files', [])])
399 episode
.flattr_url
= to_maxlength(Episode
, 'flattr_url',
400 parsed_episode
.get('flattr') or
402 episode
.license
= parsed_episode
.get('license') or episode
.license
404 episode
.title
= to_maxlength(Episode
, 'title',
405 parsed_episode
.get('title') or
407 file_basename_no_extension(episode
.url
))
409 episode
.last_update
= datetime
.utcnow()
412 parsed_urls
= list(chain
.from_iterable(
413 f
.get('urls', []) for f
in parsed_episode
.get('files', [])))
414 episode
.add_missing_urls(parsed_urls
)
417 def mark_outdated(obj
):
418 """ marks obj outdated if its not already """
423 obj
.last_update
= datetime
.utcnow()
427 def get_update_interval(episodes
):
428 """ calculates the avg interval between new episodes """
430 count
= len(episodes
)
432 logger
.info('no episodes, using default interval of %dh',
433 DEFAULT_UPDATE_INTERVAL
)
434 return DEFAULT_UPDATE_INTERVAL
436 earliest
= episodes
[0]
437 now
= datetime
.utcnow()
439 timespan_s
= (now
- earliest
.released
).total_seconds()
440 timespan_h
= timespan_s
/ 60 / 60
442 interval
= int(timespan_h
/ count
)
443 logger
.info('%d episodes in %d days => %dh interval', count
,
444 timespan_h
/ 24, interval
)
446 # place interval between {MIN,MAX}_UPDATE_INTERVAL
447 interval
= max(interval
, MIN_UPDATE_INTERVAL
)
448 interval
= min(interval
, MAX_UPDATE_INTERVAL
)
453 def file_basename_no_extension(filename
):
454 """ Returns filename without extension
456 >>> file_basename_no_extension('/home/me/file.txt')
459 >>> file_basename_no_extension('file')
462 base
= os
.path
.basename(filename
)
463 name
, extension
= os
.path
.splitext(base
)