Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
A tool for pulling and storing delicious, delicious Twitter data
Ruby

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.
config
generators
lib
test
.gitignore
README
Rakefile
VERSION
pork_sandwich.gemspec

README

--README--

Integrating pork_sandwich into your rails project: 


1. Drop this into config/environment.rb (after the Rails::Initializer.run do |config| line, but before the "end" statement of that block):

  config.gem "pork_sandwich", :version => ">=0.4.17"

3. Make sure pork_sandwich is installed by running: 

     rake gems:install 

4. create the pork_sandwich models: 

	script/generate pork_sandwich_models 

5. create the pork_sandwich db migration:
 	
	script/generate pork_sandwich_migration

6. run:

	rake db:migrate

Pork Classes and Database Objects can now be referenced throughout your project. Happy porking.



___________________________________________


Class:  Search
Used for collecting and storing tweets that match a given search term 

Search moves backwards through twitter's search API, asking twitter for tweets from most recent to oldest. Without other stop conditions set, an historical pull will continue until the method collects all that is currently available from Twitter. Twitter's search API makes available tweets created as long as ~6 days ago. 

Pork::Search.new(query, {options})

query: a string representing the search you want to run. One or more words separated by spaces will be treated as AND terms, not as a phrase. Putting a - before a word will search for tweets that do not contain that word. Terms are case insensitive. 

:desired_count => specifies the number of tweets to collect, written as an integer, provided that that many tweets available from Twitter. 
:since_id => specifies the tweet status_id, written as an integer, at which to stop collecting
:from_user => specifies the username (as string) or user_id (as integer) of a twitter user who's timeline you would like to search
:pulls_per_hour => specifies the rate at which to make API calls. Default = 1500, which is recommended for non-whitelisted IPs


Methods: 

.historical_pull

Executes a search based on the search parameters specified by Search instance. Returns true when complete. 

.current_count

returns the number of tweets collected thus far

.desired_count

read or set the desired count

.from_user

read or set the from user

.since_id

read or set the status id at which to stop collecting

.pulls_per_hour

read or set number of pulls per hour

.db_ids_created

returns an array of the database id numbers of the tweets that have been saved into the database by this search.

Examples:

# Pulls all tweets from twitter's search API that mention pork:
search = Pork::Search.new('pork')
search.historical_pull

# Pull all tweets with a status id higher than 950463847 from twitter's search API that mention smokey AND tender: 
search = Pork::Search.new('smokey+tender')
search.since_id = 950463847
search.historical_pull

# Pull up to 150 tweets created by user 'sam1vp' from twitter's search API that mention 'bbq': 
search = Pork::Search.new('bbq', {:from_user => 'sam1vp', :desired_count => 150})
search.historical_pull


___________________________________________

Class: TwitterUser
Used for collecting data about a given twitter user

TwitterUser.new({options})

:twitter_id => status id of the user you wish to gather info about
:twitter_screen_name => string of the screen name of the user you wish to gather info about
:desired_follower_count => max number of followers you want to pull and save
:desired_friend_count => max number of friends you want to pull and save
:since_tweet_id => tweet status id at which to stop collecting this user's tweets


Methods: 

.pull_tweets

pulls and stores tweets from a user's timeline. If no stop conditions have been set, this will pull as many tweets as twitter makes available, usually up to ~2000 tweets.

.pull_account_info

pulls and stores this user's account info only if it's not already in the database.

.update_account_info

pulls and stores this user's account info, overwriting any existing info in the database

.pull_friends

pulls and stores account info for 100 friends per call, cursoring through a user's friends until a desired count has been reached or all friends have been pulled. This method also stores a row in the twitter_relationships table for each friend. 

.pull_followers

pulls and stores account info for 100 followers per call, cursoring through a user's followers until a desired count has been reached or all followers have been pulled. This method also stores a row in the twitter_relationships table for each follower. 

.pull_friend_ids

pulls an array of all the ids of this user's friends in one call, creating 'shell' rows in the twitter_accounts table and rows in the twitter_relationships table. Ideal for quickly harvesting relationship data.  

.pull_follower_ids

pulls an array of all the ids of this user's followers in one call, creating 'shell' rows in the twitter_accounts table and rows in the twitter_relationships table. Ideal for quickly harvesting relationship data.  

___________________________________________

Class: Saver

Saver, which handles database insertions, is automatically instantiated when needed and referenced by $SAVER. You only need to interact with it if you want to tag tweets, account data, or relationship data via the acts_as_taggable activerecord extension. To tag data stored by a given Search or TwitterUser instance, after creating that instance, write: 

$Saver.rules = {'tags' => {'tag' => foo, 'tag' => bar}} 

___________________________________________

Class: Log

Search and TwitterUser are designed to output both standard and error handling messages to a log file to help document the data collection process. In order to take advantage of these logging features, you must create a globally accessible logger: 

$PORK_LOG = Pork::Log.new(log [, options])

log: the place to which logger should write. accepts STDOUT, STDERR, or a string specifying the relative path of a log file, which will be either created or updated. 

options
:researcher => appends a researcher name to the start of any log message
:project => appends a project name to the start of any log message


Methods: 

.write(string)

use this method on $PORK_LOG to add your own messages to the log

___________________________________________

Expected Changes: 

- A reaction processor, which can be used to search for and store tweets that either mention, reply to, retweet, or are mentioned by, replied by, or retweeted by a given user
- New database structures and methods for tracking search terms or user_timelines over time
- New database structures and methods for 'crawling,' i.e. traversing the twitter relationship graph. This will make it possible to do things like pull all tweets from a user, his followers, and his follower's followers. 
- Optional batch insertion, which will greatly speed up the rate at which data is stored. This will likely limit the gem to projects that user either MySQL or Postgres, and may preclude the use of various active_record association methods. 








Something went wrong with that request. Please try again.