Twitter Tap

Collect tweets to a mongoDB using the twitter search API.

About Twitter Tap

Twitter Tap is a python tool that connects to the Twitter API and issues calls to the search endpoint or the streaming API using a query that the user has entered.

Using the search API

If using the search API, the tool follows all the next_results links (with the corresponding max_id) so that all results are collected. When all the next_results links are exhausted the query is repeated using the since_id of the latest tweet from the results of the first query and follows all the next_results links again. The latest since_id is also stored in the database for each distinct query (query, geolocation, language), so that when the tool is restarted you will still only receive unique tweets.

Tweets are stored into a mongoDB, which has a unique index on the Tweet ID so that there is no duplication of data if more than 1 query is executed simultaneously.

There is an arbitrary wait time before each API call so that the rate limit is not reached. The default value of 2 seconds makes sure that there are no more than 450 requests per 15 minutes as is the rate limit of the search endpoint for authenticating with the app (not the user).

The tool can be run from the command line or be run as a daemon using supervisor (recommended). A sample supervisord.conf script is included with the tool.

Using the streaming API

You can also use the streaming API to store tweets. Storing tweets from the streaming API is much more straightforward, but you can only get the tweets you are currently streaming. Using the search API will also get you tweets from a couple of days back. Check out the documentation so that you can see what sorts of queries you can make with the search API vs the streaming API.

Installation

Install Twitter Tap using pip

pip install twitter-tap

Or, if you want the current code

git clone git://github.com/janezkranjc/twitter-tap.git
cd twitter-tap
python setup.py install

Before you start

Please follow this link https://apps.twitter.com/ and create a twitter app. You will need the consumer key and consumer secret to access the twitter API.

Using twitter tap

Run Twitter Tap in the command line like this.

tap

Show help text

tap -h

tap search -h

tap stream -h

Executing a query with the search API

To execute a query you must provide a query, the consumer secret and either the consumer key or the access token. Consumer key and secret can be obtained at the http://apps.twitter.com/ website, while the access token will be obtained when first connecting with the key and secret.

tap search --consumer-key CONSUMERKEY --consumer-secret CONSUMERSECRET -q "miley cyrus" -v DEBUG

Search options:

Executing a query with the streaming API

To execute a query you must provide a query, the consumer secret and either the consumer key or the access token. Consumer key and secret can be obtained at the http://apps.twitter.com/ website, while the access token will be obtained when first connecting with the key and secret.

tap stream --consumer-key CONSUMERKEY --consumer-secret CONSUMERSECRET --access-token ACCESSTOKEN --access-token-secret ACCESSTOKENSECRET --track "miley cyrus" -v DEBUG

Streaming options:

Where are the tweets stored

The tweets are stored in the mongoDB in a collection called tweets. This can be changed using the --tweets-collection option. There is also a collection for saving the highest since_id for queries, which is queries by default (can be changed using the --queries-collection option).

Running as a daemon

To run Tap as a daemon you are encouraged to use supervisor. (Doesn't work natively under windows. You should use cygwin.)

Here is a sample supervisord.conf file for running tap

; Sample supervisor config file for daemonizing the twitter search to mongodb software

[inet_http_server]          ; inet (TCP) server disabled by default
port=127.0.0.1:9001         ; (ip_address:port specifier, *:port for all iface)
username=manorastroman      ; (default is no username (open server))
password=kingofthedragonmen ; (default is no password (open server))

[supervisord]
stopsignal=INT
logfile=supervisord.log      ; (main log file;default $CWD/supervisord.log)
logfile_maxbytes=50MB        ; (max main logfile bytes b4 rotation;default 50MB)
logfile_backups=10           ; (num of main logfile rotation backups;default 10)
loglevel=info                ; (log level;default info; others: debug,warn,trace)
pidfile=supervisord.pid      ; (supervisord pidfile;default supervisord.pid)
nodaemon=false               ; (start in foreground if true;default false)
minfds=1024                  ; (min. avail startup file descriptors;default 1024)
minprocs=200                 ; (min. avail process descriptors;default 200)

[rpcinterface:supervisor]
supervisor.rpcinterface_factory = supervisor.rpcinterface:make_main_rpcinterface

[supervisorctl]
serverurl=http://127.0.0.1:9001 ; use an http:// url to specify an inet socket
username=manorastroman       ; should be same as http_username if set
password=kingofthedragonmen  ; should be same as http_password if set

[program:tapsearch]
command=tap search --consumer-key CONSUMERKEY --consumer-secret CONSUMERSECRET -q "miley cyrus" -v DEBUG
stdout_logfile=tap_search.log
stderr_logfile=tap_err.log
autostart=true
autorestart=true
startsecs=10
stopwaitsecs=10

[program:tapstream]
command=tap stream --consumer-key CONSUMERKEY --consumer-secret CONSUMERSECRET --access-token ACCESSTOKEN --access-token-secret ACCESSTOKENSECRET --track "miley cyrus" -v DEBUG
stdout_logfile=tap_stream.log
stderr_logfile=tap_stream_err.log
autostart=true
autorestart=true
startsecs=10
stopwaitsecs=10

Afterwards you can start the daemon like this (you must be in the same folder as supervisord.conf or your supervisord.conf must be /etc/)

supervisord

Open your browser to http://127.0.0.1:9001 to see the status of the daemon. By default the username is manorastroman and the password kingofthedragonmen.

Alternatively you can see the status like this

supervisorctl status

Or see the tail of the logs (log file locations can be setup in supervisord.conf)

supervisorctl tail tapsearch

supervisorctl tail tapstream

Whenever you feel like shutting it down

supervisorctl shutdown

Useful links

• MongoDB https://www.mongodb.org/
• Twitter developers https://dev.twitter.com/
• Supervisor http://supervisord.org/

Changes

v2.0.1:

• Fixed a bug where search results would fail to retreive tweets due to invalid since_id. The since_id is now manually computed.

v2.0.0:

• This version now uses two commands - search and stream, to use either with the search API or the streaming API (on version 1.1.0 you could only use the search API).

v1.1.0:

• Added two options for changing the default collection names for queries and tweets.

v1.0.0:

• This version no longer reuqires a separate settings.py file as all options can be entered as command line arguments

v0.1.0:

• Alpha release - this version needs a settings.py to enter credentials.