pravj / kayak

A customizable Twitter client to collect tweets matching certain criteria

kayak

A customizable Twitter client to collect tweets matching certain criteria

Setting up kayak

Twitter Application Credentials

• Create a new application using Twitter Application Management console.
• Note down the Consumer Key (API Key) and Consumer Secret (API Secret) credentials.

[Optional] Application's Access Level should be Read only.

Store Credentials as Environment Variables

• Save the consumer key and consumer secret credentials as envrionment variables titled KAYAK_CONSUMER_KEY and KAYAK_CONSUMER_SECRET respectively.
• One possible way to add environment variables is to add them to your ~/.bashrc file and source it.

KAYAK_CONSUMER_KEY="xxxxxxxx"; export KAYAK_CONSUMER_KEY
KAYAK_CONSUMER_SECRET="xxxxxxxxxxxxxxxx"; export KAYAK_CONSUMER_SECRET

Installation

For now the package isn't up there on PyPI, hence you'll have to install it manually.

$ git clone [email protected]:pravj/kayak.git
$ cd kayak
$ pip install .

It will make the package available in your local environment.

--

Using kayak

The package exports KayakClient class to use, it defines the following method.

get_tweets(search_query, minimum_retweet=1, older_tweets=True)
    Returns an iterator for Twitter API responses (tweet entities i.e. KayakTweet).

    :param (str) search_query: Search query operator to use.

    :param (int) minimum_retweet: Number of retweets (Default 1)
        a tweet should have at least.

    :param (bool) older_tweets:
        If True (default), it will return the iterator with older tweets.
        If False, it will return newer tweets on each iteration.

The class KayakTweet provides following variables of a tweet.

• text : Tweet text
• id : ID of the Tweet
• retweets : Number of retweets

class KayakTweet(object):
    """
    Class to change the representation of a Tweet object.
    Currently you can access the 'text', 'id', and 'retweet count' attributes.
    """

--

Create a client instance.

from kayak import KayakClient

client = KayakClient()

Refer to the API documentation Query operators to create your custom search query.

To collect old tweets (that have been created in past)

search_query = '#custserv'

# Similarly, if you want tweets mentioning Twitter account '@potus'
# search_query = '@potus'

old_tweets_iterator = client.get_tweets(search_query, minimum_retweet=2)

for tweets in old_tweets_iterator:
	for tweet in tweets:
		print tweet.id, tweet.retweets

		# same value (str) as 'tweet.text' (using __repr__ data method)
		print tweet

To collect new tweets (as they are being created continuously)

# Ideally, to avoid raising Exception
# one should wait in between successive iterations.
# So that we don't end up having no new tweets.
import time

new_tweets_iterator = client.get_tweets('#custserv', older_tweets=False)

while True:
	new_tweets = new_tweets_iterator.next()

	for tweet in new_tweets:
		print tweet.id, tweet.retweets
		print tweet.text

	time.sleep(60)

Tweet Filteration

Custom Filter Function

External Filteration using Decorators : under exploration

--

LICENSE

Pravendra Singh • GPLv3