Seed-O-Matic

Run repeatable seeds across a variety of environments

Seeds and db:fixture:load are great, but they're usually written in a way that they can only be used on an initial deploy. Seed-O-Matic gives you tools to specify seed data in a way that's repeatable across environments, allowing you to change seed data in a safe way between all of your environments.

Seed Files

Seed files are set up as JSON or YAML files. By default, Seed-O-Matic will look for seed files in config/seeds, although you can specify another directory when you run your seeds. Here's the structure of a typical seed file:

my_model:
  match_on: code
  tags: [initial_run, professional]
  seed_mode: once
  items:
    - name: My Model 1
      code: my_model
      other_model_lookup:
        code: om_1
    - name: My Model 2
      code: my_model_2
other_model:
  match_on: [code, category]
  seed_mode: always
  items:
    - name: Other Model 1
      code: om_1
      category: a
    - name: Other Model 2
      code: om_2
      category: b

The seed file starts with a model name. This should match the name of a model in your application.
The items array lists all the items you'd like to seed. Attribute names should match the name of your attributes.
You can specify a match_on attribute to prevent duplicate entries being created when seeds are run a second time. Seed-O-Matic will try to find an entry which matches your match_on fields, and update if one is found. Multiple items work as well.
You can tag your seed files as well, if you want to import a subset of seeds in particular circumstances.
seed_mode controls how SeedOMatic handles records that were matched based on the match_on fields. See 'Seed Mode' below for more details.

Running Seeds

In Ruby

Running seeds is as easy as calling:

SeedOMatic.run

This will look through the directory config/seeds, and import every file in that directory.

As a rake task

SeedOMatic provides the following rake task:

rake seedomatic:seed

This will run seeds in the config/seeds directory. You can provide SEED_DIRECTORY, TAGGED_WITH, and NOT_TAGGED_WITH arguments to the rake task to customize which seeds are run (See the sections below for valid values for these arguments).

Seed Mode

You can set a seed mode to control how SeedOMatic handles entries that already exist (i.e. models that matched the match_on fields). The following options are available:

once - Create new models, but don't update any existing models.
always - Always update, even if the model already exists.

The default seed mode is always.

Lookups

If your seeds depend on other seeded data, you can use lookups to locate other records. Just add _lookup to the name of any property, and then provide parameters to look up that data. Seed-O-Matic will search for a record that matches the criteria you specified. Seed-O-Matic will automatically determine the type of record to look up based on the association in your model.

Warning: Currently, lookups will only work correctly if the record exists at the time that the seed was run. If the item you're looking up is located in another seed file, ensure that it is imported first (by putting it in a higher alphabetical order).

Examples:

With key / value lookup parameters

category_lookup:
  code: lookup_code

This will set the category of the lookup to the first category with a code of lookup_code

With a string

category_lookup: name ilike '%code%'

This will look for category records matching the SQL fragment you passed in.

Importing specific directories, or specific files.

To load seed files from a specific directory, call:

SeedOMatic.run :dir => "your/directory"

To load a specific seed file, call:

SeedOMatic.run :file => 'config/seeds/my_model.yml'

Importing tags

You can load tags using the tagged_with, and not_tagged_with options:

SeedOMatic.run :tagged_with => 'my_tag'              # Will run all seeds tagged with my_tag
SeedOMatic.run :tagged_with => ['tag1', 'tag2']      # Will run if tagged with either tag1 or tag2
SeedOMatic.run :not_tagged_with => ['tag1', 'tag2']  # Will not run if either tag1 or tag2

TODO

Referencing lookups from other seed files
Selective updating of data

influitive/seedomatic