Sync your historical events to KISSmetrics via CSV.

http://support.kissmetrics.com/integrations/csv-import/recurring-import

Add this line to your application's Gemfile:

gem 'beso'

And then execute:

$ bundle

Or install it yourself as:

$ gem install beso

Next, create an initializer for beso. There, you can set up your S3 bucket information and define your serialization jobs:

# config/initializers/beso.rb
Beso.configure do |config|

  # First, set up your S3 credentials:

  config.access_key  = '[your AWS access key]'
  config.secret_key  = '[your AWS secret key]'
  config.bucket_name = 'beso' # recommended, but you can really call this anything

  # Then, define some jobs:

  config.job :message_delivered, :table => :messages do
    identity { |message| message.user.id }
    timestamp :created_at
    prop( :message_id ) { |message| message.id }
  end

  config.job :signed_up, :table => :users do
    identity { |user| user.id }
    timestamp :created_at
    prop( :age ){ |user| user.age }
  end
end

KISSmetrics events have three properties that must be defined:

• Identity
• Timestamp
• Event

The Identity field is some sort of identifier for your user. Even if your job is working on another table, you should probably have a way to tie the event back to the user who caused it. Here, you can provide one of three things:

• A proc that should receive the record and return the identity value
• A symbol that will get passed to record.send
• A literal (You'll probably want to do one of the other two options)

The Timestamp field is slightly different in that it should always be part of the table you are querying, not the user. This symbol will get sent to each record, but will also be used in determining the query for the job.

The Event name is inferred by the name of your job. It will be provided and formatted for you.

On top of this, you can specify up to ten custom properties. Like identity, you can pass either a proc, a symbol, or a literal:

config.job :signed_up, :table => :users do
  identity :id
  timestamp :created_at
  prop( :age ){ |user| user.age }
  prop( :new_user, true )
end

By requiring beso, you get the beso:run rake task. This task will do the following:

• Connect to your S3 bucket
• Pull down 'beso.yml' if it exists

beso.yml contains the timestamp of the last record queried for each job. If it doesn't exist, it will be created after the first run.

• Iterate over the jobs defined in the initializer you set up
• Create a CSV representation of all records newer than the timestamp found in beso.yml
• Upload each CSV to your S3 bucket with the event name and timestamp
• Update beso.yml with the latest timestamp for each job

The rake task is designed to be used via cron. For the moment, KISSmetrics will only process one CSV file per hour, so it makes sense that this task should be run at an interval of hours equal to the number of jobs you have defined. For example, if you have defined 4 jobs, this task should run once every 4 hours.

The rake task also accepts two options that you can set via environment variables.

BESO_PREFIX will change the prefix of the CSV filenames that get uploaded to S3. The default is 'beso', so it is recommended you use that when telling KISSmetrics what your filename pattern is. You can then adjust the prefix if you would like to upload CSV's that you don't want KISSmetrics to recognize.

BESO_ORIGIN will change the behavior of the task when there is no previous timestamp defined for a job in beso.yml.

By default, the task will use the last timestamp in your table (which effectively means the first run of this task will do nothing). This is because KISSmetrics charges you for every event you log through their system, so you probably don't want to upload 8 months worth of events straight away.

This option will accept two values to alter the behavior:

• now will set the first run timestamp to now, which will obviously not create any events.
• first will set the first run timestamp to the first timestamp in each table. Use this with BESO_PREFIX if you want to dump an entire table's worth of events to S3 without having KISSmetrics process them.

1. Fork it
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Added some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create new Pull Request