script/lib/tweetstream/README
author Samuel Huron <samuel.huron@centrepompidou.fr>
Tue, 26 Apr 2011 10:53:35 +0200
changeset 121 2b794b7901d6
parent 12 4daf47fcf792
child 13 79b6e132e3d7
permissions -rw-r--r--
add .htaccess

.. -*- restructuredtext -*-

##########################################
tweetstream - Simple twitter streaming API
##########################################

Introduction
------------

tweetstream provides a class, TweetStream, that can be used to get
tweets from Twitter's streaming API. An instance of the class 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.

Subclasses are available for accessing the "track" and "follow" streams
as well.

There's also a ReconnectingTweetStream class that handles automatic
reconnecting.

Twitter's documentation about the streaming API can be found here:
http://apiwiki.twitter.com/Streaming-API-Documentation .

**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.

Examples
--------

Printing all incomming tweets:

>>> stream = tweetstream.TweetStream("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.TweetStream("username", "password") as stream
...     for tweet in stream:
...         print "Got tweet from %-16s\t( 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 %-16s\t( 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 relate to specific terms, use the TrackStream:

>>> words = ["opera", "firefox", "safari"]
>>> with tweetstream.TrackStream("username", "password", words) as stream
...     for tweet in stream:
...         print "Got interesting tweet:", tweet

To get only tweets from a set of users, use the FollowStream. The following
would get tweets for user 1, 42 and 8675309

>>> users = [1, 42, 8675309]
>>> with tweetstream.FollowStream("username", "password", users) as stream
...     for tweet in stream:
...         print "Got  tweet from:", tweet["user"]["screen_name"]


Simple tweet fetcher that sends tweets to an AMQP message server using carrot:

>>> from carrot.messaging import Publisher
>>> from carrot.connection import AMQPConnection
>>> from tweetstream import TweetStream
>>> amqpconn = AMQPConnection(hostname="localhost", port=5672,
...                           userid="test", password="test",
...                           vhost="test")
>>> publisher = Publisher(connection=amqpconn,
...                       exchange="tweets", routing_key="stream")
>>> with TweetStream("username", "password") as stream:
...    for tweet in stream:
...        publisher.send(tweet)
>>> publisher.close()


Changelog
---------

See the CHANGELOG file

Contact
-------

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.

License
-------

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