tweetstream provides two classes, SampleStream and FollowStream, that can be used to get tweets from Twitter's streaming API. An instance of one of the classes can be used as an iterator. In addition to fetching tweets, the object keeps track of the number of tweets collected and the rate at which tweets are received.

SampleStream delivers a sample of all tweets. FilterStream delivers tweets that match one or more criteria. Note that it's not possible to get all tweets without access to the "firehose" stream, which is not currently avaliable to the public.

Twitter's documentation about the streaming API can be found here: http://dev.twitter.com/pages/streaming_api_methods .

Note that the API is blocking. If for some reason data is not immediatly available, calls will block until enough data is available to yield a tweet.

Printing incoming tweets:

>>> stream = tweetstream.SampleStream("username", "password") >>> for tweet in stream: ... print tweet

The stream object can also be used as a context, as in this example that prints the author for each tweet as well as the tweet count and rate:

>>> with tweetstream.SampleStream("username", "password") as stream ... for tweet in stream: ... print "Got tweet from %-16st( tweet %d, rate %.1f tweets/sec)" % ( ... tweet["user"]["screen_name"], stream.count, stream.rate )

Stream objects can raise ConnectionError or AuthenticationError exceptions:

>>> try: ... with tweetstream.TweetStream("username", "password") as stream ... for tweet in stream: ... print "Got tweet from %-16st( tweet %d, rate %.1f tweets/sec)" % ( ... tweet["user"]["screen_name"], stream.count, stream.rate ) ... except tweetstream.ConnectionError, e: ... print "Disconnected from twitter. Reason:", e.reason

To get tweets that match specific criteria, use the FilterStream. FilterStreams take three keyword arguments: "locations", "follow" and "track".

Locations are a list of bounding boxes in which geotagged tweets should originate. The argument should be an iterable of longitude/latitude pairs.

Track specifies keywords to track. The argument should be an iterable of strings.

Follow returns statuses that reference given users. Argument should be an iterable of twitter user IDs. The IDs are userid ints, not the screen names.

>>> words = ["opera", "firefox", "safari"] >>> people = [123,124,125] >>> locations = ["-122.75,36.8", "-121.75,37.8"] >>> with tweetstream.FilterStream("username", "password", track=words, ... follow=people, locations=locations) as stream ... for tweet in stream: ... print "Got interesting tweet:", tweet

tweetstream used to contain the classes TweetStream, FollowStream, TrackStream LocationStream and ReconnectingTweetStream. These were deprecated when twitter changed its API end points. The same functionality is now available in SampleStream and FilterStream. The deprecated methods will emit a warning when used, but will remain functional for a while longer.

See the CHANGELOG file

The author is Rune Halvorsen <runefh@gmail.com>. The project resides at http://bitbucket.org/runeh/tweetstream . If you find bugs, or have feature requests, please report them in the project site issue tracker. Patches are also very welcome.

• Rune Halvorsen
• Christopher Schierkolk
• Reid Priedhorsky
• Johannes

This software is licensed under the New BSD License. See the LICENCE file in the top distribution directory for the full license text.