/marty

Primary LanguageRubyMIT LicenseMIT

Build Status

Marty

Marty is a framework for viewing and reporting on versioned data. Marty provides its own scripting environment. Scripts can also be verioned. RESTful APIs are provided for accessing and executing vesioned scripts.

Marty also provides user authentication and session support. Currently, only LDAP-based authentication is supported. There's also support for role-based authorization.

Rake tasks

The Marty framework provides several rake tasks to manage its database tables and delorean scripts.

To create the correct migrations for a Marty-based application (see below for getting the internal dummy application to work):

rake marty:install:migrations

The Marty database needs to be seeded with specific configuration information. Before running your Marty application for the first time you will need to run:

rake marty:seed

If you are using Delorean scripts in your application you can load them with a rake task. By default these scripts will be picked up from app/delorean_scripts. To load scripts:

rake marty:load_scripts

You can override the default directory by setting the LOAD_DIR environment variable:

LOAD_DIR=<delorean script directory> marty:load_scripts

To delete scripts:

rake marty:delete_scripts

Scheduled Job

To use scheduled background jobs, add to config/application.rb:

config.active_job.queue_adapter = :delayed_job

Using Monit to keep background jobs alive

To have monit keep track of background jobs run a bash script to configure monit after deployment. An example script is provided in .application-scripts

Marty SqlServers

Marty comes with support for using the ActiveRecord SQL Server adapter. To use this functionality you will need:

  • A config/sql_servers.yml file that specifies your endpoint configurations, similar to config/database.yml
  • Create an initializer that:
    • requires the needed files with require 'marty/sql_servers'
    • connects on startup with Marty::SqlServers.generate_database_connections!

note The ActiveRecord SQL Server adapter will escape queries and stored procedure variables by default.

note The ActiveRecord SQL Server adapter enforces that all XML is sent as UTF-16 Little Endian. You will want to remove the instructions on XML or ensure that they are UTF-16 LE.

Dummy Application & Testing

Make sure that extjs is installed (or symbolically linked) in the dummy app at spec/dummy/public.

Docker doesn't support symlinks, so in order to run it in Docker you will have to copy extjs files.

cp -r PATH/TO/YOUR/EXTJS/FILES spec/dummy/public/extjs

You can run it with Docker:

make app-initialise-docker
make app-start

To run tests:

make app-bash
HEADLESS=true rspec

You can also use docker for running Postgres and Redis only:

cp spec/dummy/config/database.yml.db_in_docker.example spec/dummy/config/database.yml

make db-start
make db-stop

Postgres container would listen to 3030 port

To run without Docker:

Marty currently only runs with PostgreSQL. To be able to run the tests you will first need to create a database.yml file in spec/dummy/config. You can use the example file by doing:

cp spec/dummy/config/database.yml.example spec/dummy/config/database.yml

To initialize the dummy application for a demo run:

bundle install
bundle exec rake db:create db:migrate db:seed app:marty:load_scripts
cd spec/dummy
bundle exec rails s

Or, to run the application with the job workers running:

bundle install
bundle exec rake db:create db:migrate db:seed app:marty:load_scripts
cd spec/dummy
gem install foreman
foreman start

The Marty dummy app should now be accessible in your browser: localhost:3000

You can log in using marty as both user and password.

To create the test database in prepartion to run your tests:

RAILS_ENV=test bundle exec rake db:create

Then to run the tests:

bundle exec rspec

Test Isolation Support

To make it easier to run tests in CI pipelines, Marty comes with a functionality that can make recordings of external calls which can be replayed during tests.

To use this RequestRecording functionality you will need to do the following:

  • Include gem 'vcr' and gem 'webmock' in your gemfile
  • Require the desired files in your RSpec configuration
    • Marty::Gem.require_file!('spec/support/request_recording') for HTTP requests
    • Marty::Gem.require_file!('spec/support/request_recording/sql_server') for SQL Server requests
  • When running a test that connects to an external service, first run the test in an environment with connectivity and enable "Recording Mode" by setting the environment variable MARTY_RSPEC_RECORD=true
  • Should you need to regenerate a test recording, the same environment variable can be used to update test recordings.

History & Status

Marty was originally part of a much larger project internal to PennyMac. We have split Marty from its original home with the goal of making it generally available as a platform for working with versioned data and scripts. However, some rspec and cucumber tests are still in the parent and have yet to be ported. Also, documentation is sorely lacking. Hopefully, this will be rectified soon.

WIP Documentation

Reports in Delorean

Reports are the interface inside marty for users to submit data to processes as well as export data. To create a report you will need to create a Delorean script with the following conditions:

  • The script name must end in _report, e.g. example_report.dl
  • In a node inside the above script, define the following:
    • title: The display title for the report in the UI
    • form: A list of fields that will be mapped to any paramters in your node.
    • result: This is the value returned after running the report
    • optional
      • roles: provide a list of roles that have access to execute the report.
      • format: provide the format for the report result. This will default to json.

An exmaple node would be, which would belong in example_report.dl

ExmapleExtract:
    title   = "My Example Report"
    roles   = ['admin', 'user_manager']
    raw     = # some logic to do things goes here
    result  = Marty::Helper.to_csv(raw)
    form    = []
    format  = "csv"