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/>.
24 from datetime
import datetime
25 from itertools
import chain
, islice
28 from django
.db
import transaction
29 from django
.conf
import settings
31 from mygpo
.podcasts
.models
import Podcast
, URL
, Slug
, Episode
32 from mygpo
.core
.slugs
import assign_missing_episode_slugs
, PodcastSlug
33 from mygpo
.podcasts
.models
import DEFAULT_UPDATE_INTERVAL
, \
34 MIN_UPDATE_INTERVAL
, MAX_UPDATE_INTERVAL
35 from feedservice
.parse
import parse_feed
, FetchFeedException
36 from feedservice
.parse
.text
import ConvertMarkdown
37 from feedservice
.parse
.models
import ParserException
38 from feedservice
.parse
.vimeo
import VimeoError
39 from mygpo
.utils
import file_hash
, to_maxlength
40 from mygpo
.web
.logo
import CoverArt
41 from mygpo
.data
.podcast
import subscribe_at_hub
42 from mygpo
.data
.tasks
import update_related_podcasts
43 from mygpo
.pubsub
.models
import SubscriptionError
44 from mygpo
.directory
.tags
import update_category
47 logger
= logging
.getLogger(__name__
)
49 MAX_EPISODES_UPDATE
=200
51 class NoPodcastCreated(Exception):
52 """ raised when no podcast obj was created for a new URL """
55 class NoEpisodesException(Exception):
56 """ raised when parsing something that doesn't contain any episodes """
59 class PodcastUpdater(object):
60 """ Updates a number of podcasts with data from their feeds """
62 def update_queue(self
, 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 self
.update(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(self
, podcast_url
):
80 """ Update the podcast for the supplied URL """
83 parsed
= self
._fetch
_feed
(podcast_url
)
84 self
._validate
_parsed
(parsed
)
86 except (ParserException
, FetchFeedException
, NoEpisodesException
,
87 VimeoError
, ValueError, socket
.error
, urllib2
.HTTPError
) as ex
:
88 #TODO: catch valueError (for invalid Ipv6 in feedservice)
90 if isinstance(ex
, VimeoError
):
91 logger
.exception('Problem when updating Vimeo feed %s',
94 # if we fail to parse the URL, we don't even create the
97 p
= Podcast
.objects
.get(urls__url
=podcast_url
)
98 # if it exists already, we mark it as outdated
99 self
._mark
_outdated
(p
, 'error while fetching feed: %s' %
103 except Podcast
.DoesNotExist
:
104 raise NoPodcastCreated(ex
)
106 assert parsed
, 'fetch_feed must return something'
107 p
= Podcast
.objects
.get_or_create_for_url(podcast_url
)
108 episodes
= self
._update
_episodes
(p
, parsed
.episodes
)
109 max_episode_order
= self
._order
_episodes
(p
)
110 self
._update
_podcast
(p
, parsed
, episodes
, max_episode_order
)
114 def verify_podcast_url(self
, podcast_url
):
115 parsed
= self
._fetch
_feed
(podcast_url
)
116 self
._validate
_parsed
(parsed
)
120 def _fetch_feed(self
, podcast_url
):
122 t
= socket
.getdefaulttimeout()
123 socket
.setdefaulttimeout(10)
124 return parse_feed(podcast_url
, text_processor
=ConvertMarkdown())
125 socket
.setdefaulttimeout(t
)
129 def _validate_parsed(self
, 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
.episodes
:
136 raise NoEpisodesException('no episodes found')
139 def _update_podcast(self
, 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
.title
or podcast
.title
146 podcast
.description
= parsed
.description
or podcast
.description
147 podcast
.subtitle
= parsed
.subtitle
or podcast
.subtitle
148 podcast
.link
= parsed
.link
or podcast
.link
149 podcast
.logo_url
= parsed
.logo
or podcast
.logo_url
150 podcast
.author
= to_maxlength(Podcast
, 'author', parsed
.author
or podcast
.author
)
151 podcast
.language
= to_maxlength(Podcast
, 'language', parsed
.language
or podcast
.language
)
152 podcast
.content_types
= ','.join(parsed
.content_types
) or podcast
.content_types
153 #podcast.tags['feed'] = parsed.tags or podcast.tags.get('feed', [])
154 podcast
.common_episode_title
= to_maxlength(Podcast
,
155 'common_episode_title',
156 parsed
.common_title
or podcast
.common_episode_title
)
157 podcast
.new_location
= parsed
.new_location
or podcast
.new_location
158 podcast
.flattr_url
= to_maxlength(Podcast
, 'flattr_url',
159 parsed
.flattr
or podcast
.flattr_url
)
160 podcast
.hub
= parsed
.hub
or podcast
.hub
161 podcast
.license
= parsed
.license
or podcast
.license
162 podcast
.max_episode_order
= max_episode_order
164 podcast
.add_missing_urls(parsed
.urls
)
166 if podcast
.new_location
:
168 new_podcast
= Podcast
.objects
.get(urls__url
=podcast
.new_location
)
169 if new_podcast
!= podcast
:
170 self
._mark
_outdated
(podcast
, 'redirected to different podcast')
172 except Podcast
.DoesNotExist
:
173 podcast
.set_url(podcast
.new_location
)
176 # latest episode timestamp
177 episodes
= Episode
.objects
.filter(podcast
=podcast
, released__isnull
=False).order_by('released')
179 podcast
.update_interval
= get_update_interval(episodes
)
181 latest_episode
= episodes
.last()
183 podcast
.latest_episode_timestamp
= latest_episode
.released
185 podcast
.episode_count
= Episode
.objects
.filter(podcast
=podcast
).count()
188 self
._update
_categories
(podcast
, prev_latest_episode_timestamp
)
190 # try to download the logo and reset logo_url to None on http errors
191 found
= self
._save
_podcast
_logo
(podcast
.logo_url
)
193 podcast
.logo_url
= None
195 # The podcast is always saved (not just when there are changes) because
196 # we need to record the last update
197 logger
.info('Saving podcast.')
198 podcast
.last_update
= datetime
.utcnow()
203 subscribe_at_hub(podcast
)
204 except SubscriptionError
as se
:
205 logger
.warn('subscribing to hub failed: %s', str(se
))
209 slug
= PodcastSlug(podcast
).get_slug()
211 podcast
.add_slug(slug
)
213 assign_missing_episode_slugs(podcast
)
214 update_related_podcasts
.delay(podcast
)
217 def _update_categories(self
, podcast
, prev_timestamp
):
218 """ checks some practical requirements and updates a category """
220 from datetime
import timedelta
222 max_timestamp
= datetime
.utcnow() + timedelta(days
=1)
225 if not podcast
.latest_episode_timestamp
:
229 if prev_timestamp
and podcast
.latest_episode_timestamp
<= prev_timestamp
:
232 # too far in the future
233 if podcast
.latest_episode_timestamp
> max_timestamp
:
236 # not enough subscribers
237 if podcast
.subscriber_count() < settings
.MIN_SUBSCRIBERS_CATEGORY
:
240 update_category(podcast
)
243 def _update_episodes(self
, podcast
, parsed_episodes
):
245 pid
= podcast
.get_id()
247 # list of (obj, fun) where fun is the function to update obj
248 updated_episodes
= []
249 episodes_to_update
= list(islice(parsed_episodes
, 0, MAX_EPISODES_UPDATE
))
250 logger
.info('Parsed %d (%d) episodes', len(parsed_episodes
),
251 len(episodes_to_update
))
253 logger
.info('Updating %d episodes', len(episodes_to_update
))
254 for n
, parsed
in enumerate(episodes_to_update
, 1):
256 url
= get_episode_url(parsed
)
258 logger
.info('Skipping episode %d for missing URL', n
)
261 logger
.info('Updating episode %d / %d', n
, len(parsed_episodes
))
263 episode
= Episode
.objects
.get_or_create_for_url(podcast
, url
)
265 update_episode(parsed
, episode
, podcast
)
266 updated_episodes
.append(episode
)
268 # and mark the remaining ones outdated
269 current_episodes
= Episode
.objects
.filter(podcast
=podcast
,
270 outdated
=False)[:500]
271 outdated_episodes
= set(current_episodes
) - set(updated_episodes
)
273 logger
.info('Marking %d episodes as outdated', len(outdated_episodes
))
274 for episode
in outdated_episodes
:
275 mark_outdated(episode
)
278 def _order_episodes(self
, podcast
):
279 """ Reorder the podcast's episode according to release timestamp
281 Returns the highest order value (corresponding to the most recent
284 num_episodes
= podcast
.episode_set
.count()
288 episodes
= podcast
.episode_set
.all().extra(select
={
289 'has_released': 'released IS NOT NULL',
291 .order_by('-has_released', '-released', 'pk')\
294 for n
, episode
in enumerate(episodes
.iterator(), 1):
295 # assign ``order`` from higher (most recent) to 0 (oldest)
296 # None means "unknown"
297 new_order
= num_episodes
- n
299 # optimize for new episodes that are newer than all existing
300 if episode
.order
== new_order
:
303 logger
.info('Updating order from {} to {}'.format(episode
.order
,
305 episode
.order
= new_order
308 return num_episodes
-1
310 def _save_podcast_logo(self
, cover_art
):
315 image_sha1
= hashlib
.sha1(cover_art
).hexdigest()
316 prefix
= CoverArt
.get_prefix(image_sha1
)
318 filename
= CoverArt
.get_original(prefix
, image_sha1
)
319 dirname
= CoverArt
.get_dir(filename
)
321 # get hash of existing file
322 if os
.path
.exists(filename
):
323 with
open(filename
) as f
:
324 old_hash
= file_hash(f
).digest()
328 logger
.info('Logo %s', cover_art
)
331 with
open(filename
, 'w') as fp
:
332 fp
.write(urllib2
.urlopen(cover_art
).read())
334 # get hash of new file
335 with
open(filename
) as f
:
336 new_hash
= file_hash(f
).digest()
338 # remove thumbnails if cover changed
339 if old_hash
!= new_hash
:
340 thumbnails
= CoverArt
.get_existing_thumbnails(prefix
, filename
)
341 logger
.info('Removing %d thumbnails', len(thumbnails
))
347 except (urllib2
.HTTPError
, urllib2
.URLError
, ValueError,
348 httplib
.BadStatusLine
, socket
.error
, IOError) as e
:
349 logger
.warn('Exception while updating podcast logo: %s', str(e
))
352 def _mark_outdated(self
, podcast
, msg
=''):
353 logger
.info('marking podcast outdated: %s', msg
)
354 podcast
.outdated
= True
355 podcast
.last_update
= datetime
.utcnow()
357 self
._update
_episodes
(podcast
, [])
360 def get_episode_url(parsed_episode
):
361 """ returns the URL of a parsed episode """
362 for f
in parsed_episode
.files
:
368 def update_episode(parsed_episode
, episode
, podcast
):
369 """ updates "episode" with the data from "parsed_episode" """
371 # TODO: check if there have been any changes, to avoid unnecessary updates
372 episode
.guid
= to_maxlength(Episode
, 'guid', parsed_episode
.guid
or episode
.guid
)
373 episode
.description
= parsed_episode
.description
or episode
.description
374 episode
.subtitle
= parsed_episode
.subtitle
or episode
.subtitle
375 episode
.content
= parsed_episode
.content
or parsed_episode
.description
or episode
.content
376 episode
.link
= to_maxlength(Episode
, 'link',
377 parsed_episode
.link
or episode
.link
)
378 episode
.released
= datetime
.utcfromtimestamp(parsed_episode
.released
) if parsed_episode
.released
else episode
.released
379 episode
.author
= to_maxlength(Episode
, 'author', parsed_episode
.author
or episode
.author
)
380 episode
.duration
= parsed_episode
.duration
or episode
.duration
381 episode
.filesize
= parsed_episode
.files
[0].filesize
382 episode
.language
= parsed_episode
.language
or episode
.language
or \
384 episode
.mimetypes
= ','.join(list(set(filter(None, [f
.mimetype
for f
in parsed_episode
.files
]))))
385 episode
.flattr_url
= to_maxlength(Episode
, 'flattr_url',
386 parsed_episode
.flattr
or
388 episode
.license
= parsed_episode
.license
or episode
.license
390 episode
.title
= to_maxlength(Episode
, 'title',
391 parsed_episode
.title
or episode
.title
or
392 file_basename_no_extension(episode
.url
))
394 episode
.last_update
= datetime
.utcnow()
397 parsed_urls
= list(chain
.from_iterable(f
.urls
for f
in parsed_episode
.files
))
398 episode
.add_missing_urls(parsed_urls
)
401 def mark_outdated(obj
):
402 """ marks obj outdated if its not already """
407 obj
.last_update
= datetime
.utcnow()
411 def get_update_interval(episodes
):
412 """ calculates the avg interval between new episodes """
414 count
= len(episodes
)
416 logger
.info('no episodes, using default interval of %dh',
417 DEFAULT_UPDATE_INTERVAL
)
418 return DEFAULT_UPDATE_INTERVAL
420 earliest
= episodes
[0]
421 now
= datetime
.utcnow()
423 timespan_s
= (now
- earliest
.released
).total_seconds()
424 timespan_h
= timespan_s
/ 60 / 60
426 interval
= int(timespan_h
/ count
)
427 logger
.info('%d episodes in %d days => %dh interval', count
,
428 timespan_h
/ 24, interval
)
430 # place interval between {MIN,MAX}_UPDATE_INTERVAL
431 interval
= max(interval
, MIN_UPDATE_INTERVAL
)
432 interval
= min(interval
, MAX_UPDATE_INTERVAL
)
437 def file_basename_no_extension(filename
):
438 """ Returns filename without extension
440 >>> file_basename_no_extension('/home/me/file.txt')
443 >>> file_basename_no_extension('file')
446 base
= os
.path
.basename(filename
)
447 name
, extension
= os
.path
.splitext(base
)