return urllib2.build_opener(*openers)
+def transfer_stats(sent, received, **kwargs):
+ """
+ This function takes two arguments: sent is the number of bytes
+ sent so far, received is the number of bytes received. The
+ function returns a continuation that you can call later.
+
+ The continuation takes the same two arguments. It returns a tuple
+ of the number of bytes sent, the number of bytes received and the
+ time since the original function was invoked.
+ """
+ start_time = time.time()
+ start_sent = sent
+ start_received = received
+
+ def e(sent, received, **kwargs):
+ return (sent - start_sent,
+ received - start_received,
+ time.time() - start_time)
+
+ return e
+
# If not None, a subprocess.Popen object corresponding to a
# update_feeds.py process.
update_feed_process = None
cache = self.cache[table]
if time.time() - cache[None] > 60:
- self.cache[table].clear()
+ # logger.debug("%s: Cache too old: clearing" % (table,))
+ del self.cache[table]
+ cache = None
except KeyError:
cache = None
# The cache is empty or the caller wants a column that we
# don't cache.
if (table, column) in self.cached_columns:
+ # logger.debug("%s: Rebuilding cache" % (table,))
+
do_cache = True
self.cache[table] = cache = {}
try:
if id is not None:
- return cache[column][id]
+ value = cache[column][id]
+ # logger.debug("%s.%s:%s -> %s" % (table, column, id, value))
+ return value
else:
return cache[column].values()
except KeyError:
+ # logger.debug("%s.%s:%s -> Not found" % (table, column, id))
return None
class Feed(BaseObject):
except OSError:
pass
- raise exception
+ return None
else:
#open(filename,"a").close() # "Touch" the file
file = open(filename,"a")
time.sleep(1)
def _updateFeed(self, configdir, url, etag, modified, expiryTime=24, proxy=None, imageCache=False, postFeedUpdateFunc=None, *postFeedUpdateFuncArgs):
+ logger.debug("Updating %s" % url)
+
success = False
have_serial_execution_lock = False
try:
- download_start = time.time ()
+ update_start = time.time ()
progress_handler = HTTPProgressHandler(download_callback)
openers.append (proxy)
kwargs = {'handlers':openers}
+ feed_transfer_stats = transfer_stats(0, 0)
+
tmp=feedparser.parse(url, etag=etag, modified=modified, **kwargs)
- download_duration = time.time () - download_start
-
+ download_duration = time.time () - update_start
+
opener = downloader(progress_handler, proxy)
if JobManager().do_quit:
expiry = float(expiryTime) * 3600.
currentTime = 0
-
+
+ updated_objects = 0
+ new_objects = 0
+
def wc_success():
try:
wc().stream_register (self.key, "", 6 * 60 * 60)
|woodchuck.Indicator.StreamWide),
transferred_down=progress_handler.stats['received'],
transferred_up=progress_handler.stats['sent'],
- transfer_time=download_start,
+ transfer_time=update_start,
transfer_duration=download_duration,
- new_objects=len (tmp.entries),
- objects_inline=len (tmp.entries))
+ new_objects=new_objects,
+ updated_objects=updated_objects,
+ objects_inline=new_objects + updated_objects)
except KeyError:
logger.warn(
"Failed to register update of %s with woodchuck!"
logger.debug("%s: No changes to feed." % (self.key,))
mainthread.execute(wc_success, async=True)
success = True
- elif len(tmp["entries"])==0 and not tmp.version:
+ elif len(tmp["entries"])==0 and not tmp.get('version', None):
# An error occured fetching or parsing the feed. (Version
# will be either None if e.g. the connection timed our or
# '' if the data is not a proper feed)
logger.error(
"Error fetching %s: version is: %s: error: %s"
- % (url, str (tmp.version),
+ % (url, str (tmp.get('version', 'unset')),
str (tmp.get ('bozo_exception', 'Unknown error'))))
logger.debug(tmp)
def register_stream_update_failed(http_status):
#reversedEntries = self.getEntries()
#reversedEntries.reverse()
- ids = self.getIds()
-
tmp["entries"].reverse()
for entry in tmp["entries"]:
# Yield so as to make the main thread a bit more
# responsive.
time.sleep(0)
+ entry_transfer_stats = transfer_stats(
+ *feed_transfer_stats(**progress_handler.stats)[0:2])
+
if JobManager().do_quit:
raise KeyboardInterrupt
- received_base = progress_handler.stats['received']
- sent_base = progress_handler.stats['sent']
object_size = 0
date = self.extractDate(entry)
entry["id"] = None
content = self.extractContent(entry)
object_size = len (content)
- received_base -= len (content)
tmpEntry = {"title":entry["title"], "content":content,
"date":date, "link":entry["link"], "author":entry["author"], "id":entry["id"]}
id = self.generateUniqueId(tmpEntry)
+ current_version = self.db.execute(
+ 'select date, ROWID from feed where id=?',
+ (id,)).fetchone()
+ if (current_version is not None
+ and current_version[0] == date):
+ logger.debug("ALREADY DOWNLOADED %s (%s)"
+ % (entry["title"], entry["link"]))
+ continue
+
+ if current_version is not None:
+ # The version was updated. Mark it as unread.
+ logger.debug("UPDATED: %s (%s)"
+ % (entry["title"], entry["link"]))
+ updated_objects += 1
+ else:
+ logger.debug("NEW: %s (%s)"
+ % (entry["title"], entry["link"]))
+ new_objects += 1
+
#articleTime = time.mktime(self.entries[id]["dateTuple"])
soup = BeautifulSoup(self.getArticle(tmpEntry)) #tmpEntry["content"])
images = soup('img')
baseurl = tmpEntry["link"]
- #if not id in ids:
if imageCache and len(images) > 0:
self.serial_execution_lock.release ()
have_serial_execution_lock = False
for img in images:
+ if not img.has_key('src'):
+ continue
+
filename = self.addImage(
configdir, self.key, baseurl, img['src'],
opener=opener)
file = open(tmpEntry["contentLink"], "w")
file.write(soup.prettify())
file.close()
- if id in ids:
- self.db.execute("UPDATE feed SET updated=? WHERE id=?;", (currentTime, id) )
- self.db.commit()
- else:
- values = (id, tmpEntry["title"], tmpEntry["contentLink"], tmpEntry["date"], currentTime, tmpEntry["link"], 0)
- self.db.execute("INSERT INTO feed (id, title, contentLink, date, updated, link, read) VALUES (?, ?, ?, ?, ?, ?, ?);", values)
- self.db.commit()
-# else:
-# try:
-# self.db.execute("UPDATE feed SET updated=? WHERE id=?;", (currentTime, id) )
-# self.db.commit()
-# filename = configdir+self.key+".d/"+id+".html"
-# file = open(filename,"a")
-# utime(filename, None)
-# file.close()
-# images = self.db.execute("SELECT imagePath FROM images where id=?;", (id, )).fetchall()
-# for image in images:
-# file = open(image[0],"a")
-# utime(image[0], None)
-# file.close()
-# except:
-# pass
-
+
+ values = {'id': id,
+ 'title': tmpEntry["title"],
+ 'contentLink': tmpEntry["contentLink"],
+ 'date': tmpEntry["date"],
+ 'updated': currentTime,
+ 'link': tmpEntry["link"],
+ 'read': 0}
+
+ if current_version is not None:
+ # This is an update. Ensure that the existing
+ # entry is replaced.
+ values['ROWID'] = current_version[1]
+
+ cols, values = zip(*values.items())
+ self.db.execute(
+ "INSERT OR REPLACE INTO feed (%s) VALUES (%s);"
+ % (','.join(cols), ','.join(('?',) * len(values))),
+ values)
+ self.db.commit()
+
# Register the object with Woodchuck and mark it as
# downloaded.
def register_object_transferred(
else:
publication_time = None
- sent = progress_handler.stats['sent'] - sent_base
- received = (progress_handler.stats['received']
- - received_base)
+ sent, received, _ \
+ = entry_transfer_stats(**progress_handler.stats)
+ # sent and received are for objects (in
+ # particular, images) associated with this
+ # item. We also want to attribute the data
+ # transferred for the item's content. This is
+ # a good first approximation.
+ received += len(content)
mainthread.execute(
register_object_transferred(
async=True)
self.db.commit()
+ sent, received, _ \
+ = feed_transfer_stats(**progress_handler.stats)
logger.debug (
"%s: Update successful: transferred: %d/%d; objects: %d)"
- % (self.key,
- progress_handler.stats['sent'],
- progress_handler.stats['received'],
- len (tmp.entries)))
+ % (url, sent, received, len (tmp.entries)))
mainthread.execute (wc_success, async=True)
success = True
# Check that Woodchuck's state is up to date with respect our
# state.
- updater = os.path.basename(sys.argv[0]) == 'update_feeds.py'
- wc_init (self, True if updater else False)
- if wc().available() and updater:
- # The list of known streams.
- streams = wc().streams_list ()
- stream_ids = [s.identifier for s in streams]
-
- # Register any unknown streams. Remove known streams from
- # STREAMS_IDS.
- for key in self.getListOfFeeds():
- title = self.getFeedTitle(key)
- # XXX: We should also check whether the list of
- # articles/objects in each feed/stream is up to date.
- if key not in stream_ids:
- logger.debug(
- "Registering previously unknown channel: %s (%s)"
- % (key, title,))
- # Use a default refresh interval of 6 hours.
- wc().stream_register (key, title, 6 * 60 * 60)
- else:
- # Make sure the human readable name is up to date.
- if wc()[key].human_readable_name != title:
- wc()[key].human_readable_name = title
- stream_ids.remove (key)
-
-
- # Unregister any streams that are no longer subscribed to.
- for id in stream_ids:
- logger.debug("Unregistering %s" % (id,))
- w.stream_unregister (id)
+ try:
+ updater = os.path.basename(sys.argv[0]) == 'update_feeds.py'
+ wc_init (self, True if updater else False)
+ if wc().available() and updater:
+ # The list of known streams.
+ streams = wc().streams_list ()
+ stream_ids = [s.identifier for s in streams]
+
+ # Register any unknown streams. Remove known streams from
+ # STREAMS_IDS.
+ for key in self.getListOfFeeds():
+ title = self.getFeedTitle(key)
+ # XXX: We should also check whether the list of
+ # articles/objects in each feed/stream is up to date.
+ if key not in stream_ids:
+ logger.debug(
+ "Registering previously unknown channel: %s (%s)"
+ % (key, title,))
+ # Use a default refresh interval of 6 hours.
+ wc().stream_register (key, title, 6 * 60 * 60)
+ else:
+ # Make sure the human readable name is up to date.
+ if wc()[key].human_readable_name != title:
+ wc()[key].human_readable_name = title
+ stream_ids.remove (key)
+
+
+ # Unregister any streams that are no longer subscribed to.
+ for id in stream_ids:
+ logger.debug("Unregistering %s" % (id,))
+ w.stream_unregister (id)
+ except Exception:
+ logger.exception("Registering streams with Woodchuck")
def importOldFormatFeeds(self):
"""This function loads feeds that are saved in an outdated format, and converts them to sqlite"""
in_progress = stats['jobs-in-progress']
queued = stats['jobs-queued']
- percent = (100 * ((completed + in_progress / 2.))
- / (completed + in_progress + queued))
+ try:
+ percent = (100 * ((completed + in_progress / 2.))
+ / (completed + in_progress + queued))
+ except ZeroDivisionError:
+ percent = 100
update_server_object().UpdateProgress(
percent, completed, in_progress, queued, 0, 0, 0, key)
projects / feedingit / blobdiff