QueryBigly

A quick and dirty Rails <> BigQuery integration. This gem is in beta

Installation

Add this line to your application's Gemfile:

gem 'query_bigly'

And then execute:

$ bundle

Or install it yourself as:

$ gem install query_bigly

It's prudent to note that since the QueryBigly defaults are dependent upon environment variables, these will need to be available to the gem upon instantiation. For troubleshooting, please see the following regarding Dotenv load order

Usage

Do These Things First

Copy the environment variables out of .env, and include them in your application. Change all values as pertains to your project
THIS IS IMPORTANT: create a dataset in your BigQuery project, and set DEFAULT_DATASET to that value. Make sure it's unique to your project, and don't overwrite or corrupt other datasets. Set your BIGQUERY_KEYFILE_JSON environment variable to the credentials gifted to you via devops, it should be surrounded in quotes as follows: '{type:"type", project_id: "project_id"}'.

Basic BigQuery actions

The QueryBigly::Client class is an abstracted way to consistently interface with the Google::Cloud::Bigquery class from the google-cloud-bigquery gem.

QueryBigly::Client.new() will instansiate a new client with your default attributes.
QueryBigly::Client.new(override_dataset, override_project_id) will enable you to override the defaults set in environment variables

Querying

Run a Query

statement = SELECT * FROM dogs;
QueryBigly::Client.new.run_query(statement)

Run an Asynchronous Query

statement = SELECT * FROM dogs;
QueryBigly::Client.new.run_async_query(statement)

PROTIP - make sure to squish your statements to remove whitespace.

Bulk Insertion

Currently this is a pretty ungraceful way to cram data into BigQuery. Be warned, this is a destructive action.

QueryBigly::Client.new.bulk_insert_model(klass, data_as_an_array_of_json, table_name=nil, custom_fields={})

This will dump the table of a given class into BigQuery. Be sure to properly set your DEFAULT_DATASET or override the default dataset upon creation of your QueryBigly::Client.new(override_dataset)

Streaming

Stream a Record

table_id = 'your_cool_table'
row_data = {'column_name'=>'your_cool_data', 'column_name_2'=>123}

QueryBigly::Client.new.stream_record_to_bigquery(table_id, row_data)

Table Creation

Stream an ActiveRecord::Model

As it stands, this will be the starting point for streaming our models to BigQuery -- more to come on this. Create a job via sidekiq/resque that follow this pattern:

class PushRecordToBigQueryWorker
  include Sidekiq::Worker

  def perform(klass, pk, custom_fields={}, partition_by=nil)
    QueryBigly::StreamModel.stream_model(klass, pk, custom_fields, partition_by)
  end
end

I elected to not include this job in the gem because of the various configurations of jobs that we have across our applications. It's up to you to build this piece of logic.

In your model, create an after_commit callback to push_your_model_to_bigquery
In that callback, instantiate the PushRecordToBigQueryWorker with the appropriate arguements

Custom Fields: Some records are going to require custom fields, whether your purpose is to flatten a JSON column (JSON data types are NOT supported in BigQuery) or to leave out erroneous information. The custom_fields pattern will look something like the following:

# example from Interaction in CXM
def custom_fields
    {
      "id"=>:integer,
      "occurred_at"=>:datetime,
      "created_at"=>:datetime,
      "updated_at"=>:datetime,
      "person_location_id"=>:integer,
      "payload"=>:json,
      "lead_uid"=>:string,
      "source"=>:string,
      "marketing_source_urn"=>:string,
      "location_urn"=>:string,
      "client_urn"=>:string,
      "payload_type"=>:string,
      "full_name"=>:string,
      "location_name"=>:string,
      "payload->'type' AS nested_payload_type"=>:string,
      "payload->'ga_client_id' AS ga_client_id"=>:string,
      "payload->'system' AS user_system"=>:string,
      "payload->'category' AS category"=>:string,
      "payload->'normalized'->'customer'->'first_name' AS first_name"=>:string,
      "payload->'normalized'->'customer'->'last_name' AS last_name"=>:string,
      "payload->'normalized'->'customer'->'name' AS name"=>:string,
      "payload->'normalized'->'customer'->'telephone' AS telephone"=>:string,
      "payload->'normalized'->'customer'->'email' AS email"=>:string,
      "payload->'normalized'->'customer'->'existing_customer' AS existing_customer"=>:string,
      "payload->'normalized'->'customer'->'lead_type' AS lead_type"=>:string,
      "payload->'normalized'->'call'->'called_number' AS dialed_number"=>:string,
      "payload->'normalized'->'call'->'duration' AS call_duration"=>:string,
      "payload->'normalized'->'call'->'mp3_url' AS mp3_url"=>:string
    }
  end

Partition By: Since asynchronous inserts are not supported in this version, partitioning our tables by a datetime column is mandatory. The default is created_at, but you can overwrite that with any given datetime/timestamp/date column.

Development

After checking out the repo, run bin/setup to install dependencies. Then, run rake spec to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.

Contributing

Bug reports and pull requests are welcome on GitHub at https://github.com/g5search/query_bigly. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the Contributor Covenant code of conduct.

License

The gem is available as open source under the terms of the MIT License.

Code of Conduct

Everyone interacting in the QueryBigly project’s codebases, issue trackers, chat rooms and mailing lists is expected to follow the code of conduct.

g5search / query_bigly