def listen(self):
# slopppy way to keep it running and reconnecting if it hits a bump.
# @NOTE you MUST kill the process to get this to stop because of this hack. I know.... I suck, deal with it..
while True: # Keeps it going and reconnecting if it crashes. See 'sloppy' above.
try:
with tweetstream.FilterStream(self.username, self.password, track=self.track) as stream:
for tweet in stream:
try:
tuser = self.handle_user(tweet)
obj = Tweet(twitter_user=tuser)
obj.text = tweet["text"].encode('utf-8')
obj.tweet_id = tweet["id_str"]
obj.created_at = self.make_datetime(tweet["created_at"])
obj.save()
if TWAP_ARCHIVE:
self.archive_json(obj, tweet)
for tag in self.hashtags(tweet):
obj.tags.add(tag)
if tweet["geo"]:
self.handle_geo(obj, tweet["geo"])
if tweet["coordinates"]:
self.handle_coordinates(obj, tweet["coordinates"])
print "%s - %s" % (stream.count, self.hashtags(tweet))
except:
print "Error Reading Record, skipping."
except tweetstream.ConnectionError, e:
print "Twitter Connection Error:", e.reason
break
except:
def _handle_tweet(self, tweet_data):
"""
Bases raw tweet data into various django models.
:param tweet_data:
:return:
"""
try:
tweet = Tweet.objects.get(tweet_id=tweet_data['id'])
return None # Already exists, skip it.
except Tweet.DoesNotExist:
tweet = Tweet(
twitter_user = self._handle_user(tweet_data),
tweet_id = tweet_data['id'],
created_at = self._parse_datetime(tweet_data['created_at']),
text = tweet_data['text']
)
tweet.save()
hashtags = self._parse_hashtags(tweet_data['text'])
if hashtags:
tweet.tags.add(*hashtags)
return tweet
