README.md
# Pragma
> ## This project is not being maintained anymore.
>
> Unfortunately, I'm not doing much work on REST APIs (or APIs in general) anymore, which means
> it's been a while since I last used Pragma in a real project or improved the ecosystem. If anyone
> wants to take over the project and keep it alive, feel free to reach out to me at
> hello@alessandro.codes.
[](https://travis-ci.org/pragmarb/pragma)
[](https://coveralls.io/github/pragmarb/pragma?branch=master)
[](https://codeclimate.com/github/pragmarb/pragma/maintainability)
Welcome to Pragma, an expressive, opinionated ecosystem for building beautiful RESTful APIs with
Ruby.
You can think of this as a meta-gem that pulls in the following pieces:
- [Pragma::Operation](https://github.com/pragmarb/pragma-operation);
- [Pragma::Policy](https://github.com/pragmarb/pragma-policy);
- [Pragma::Decorator](https://github.com/pragmarb/pragma-decorator);
- [Pragma::Contract](https://github.com/pragmarb/pragma-contract).
Additionally, it also provides default CRUD operations that leverage all of the Pragma components
and will make creating new resources in your API a breeze.
Looking for a Rails integration? Check out [pragma-rails](https://github.com/pragmarb/pragma-rails)!
## Philosophy
Pragma was created with a very specific goal in mind: to make the development of JSON APIs a matter
of hours, not days. In other words, Pragma is for JSON APIs what Rails is for web applications.
Here are the ground rules:
1. **Pragma is opinionated.** With Pragma, you don't get to make a lot of choices and that's
_exactly_ why people are using it: they want to focus on the business logic of their API rather
than the useless details. We understand this approach will not work in some cases and that's
alright. If you need more personalization, only use a subset of Pragma (see item 2) or something
else.
2. **Pragma is modular.** Pragma is built as a set of gems (currently 6), plus some standalone
tools. You can pick one or more modules and use them in your application as you see fit. Even
though they are completely independent from each other, they nicely integrate and work best when
used together, creating an ecosystem that will dramatically speed up your design and development
process.
3. **Pragma is designed to be Rails-free.** Just as what happens with Trailblazer, our Rails
integration is decoupled from the rest of the ecosystem and all of the gems can be used without
Rails. This is just a byproduct of the project's design: Pragma is built with pure Ruby.
[pragma-rails](https://github.com/pragmarb/pragma-rails) is the only available framework
integration at the moment, but more will come!
### Why not Trailblazer?
[Trailblazer](https://github.com/trailblazer/trailblazer) and all of its companion projects are
awesome. They are so awesome that Pragma is built on top of them: even though we're not using
the Trailblazer gem itself yet, many of the Pragma gems are simply extensions of their Trailblazer
counterparts:
- decorators are [ROAR representers](https://github.com/apotonick/roar);
- contracts are [Reform forms](https://github.com/apotonick/reform);
- operations are [Trailblazer operations](https://github.com/trailblazer/trailblazer-operation).
Trailblazer and Pragma have different (but similar) places in the Ruby world: Trailblazer is an
architecture for building all kinds of web applications in an intelligent, rational way, while
Pragma is an architecture for building JSON APIs. We have shamelessly taken all of the flexibility
and awesomeness from the Trailblazer project and restricted it to a narrow field of work, providing
tools, helpers and integrations that could never be part of Trailblazer due to their specificity.
Thank you, guys!
## Installation
Add this line to your application's Gemfile:
```ruby
gem 'pragma'
```
And then execute:
```console
$ bundle
```
Or install it yourself as:
```console
$ gem install pragma
```
## Usage
### Project Structure
This gem works best if you follow the recommended structure for organizing resources:
```
└── api
└── v1
└── article
├── contract
│ ├── create.rb
│ └── update.rb
├── operation
│ ├── create.rb
│ ├── destroy.rb
│ ├── index.rb
│ └── update.rb
└── decorator
| ├── collection.rb
| └── instance.rb
└── policy.rb
```
Your modules and classes would, of course, follow the same structure: `API::V1::Article::Policy` and
so on and so forth.
If you adhere to this structure, the gem will be able to locate all of your classes without any
explicit configuration. This will save you a lot of time and is highly recommended.
### Fantastic Five
Pragma comes with five built-in operations, often referred to as Fantastic Five (or "FF" for
brevity). They are, of course, Index, Show, Create, Update and Destroy.
These operations leverage the full power of the integrated Pragma ecosystem and require all four
components to be properly installed and configured in your application. You may reconfigure them
to skip some of the steps, but it is highly recommended to use them as they come.
You can find these operations under [lib/pragma/operation](https://github.com/pragmarb/pragma/tree/master/lib/pragma/operation).
To use them, simply create your own operations and inherit from ours. For instance:
```ruby
module API
module V1
module Article
module Operation
class Create < Pragma::Operation::Create
# This assumes that you have the following:
# 1) an Article model
# 2) a Policy (responding to #create?)
# 3) a Create contract
# 4) an Instance decorator
end
end
end
end
end
```
## Macros
The FF are implemented through their own set of macros, which take care of stuff like authorizing,
paginating, filtering etc.
If you want, you can use these macros in your own operations.
### Classes
**Used in:** Index, Show, Create, Update, Destroy
The `Classes` macro is responsible of tying together all the Pragma components: put it into an
operation and it will determine the class names of the related policy, model, decorators and
contract. You can override any of these classes when defining the operation or at runtime if you
wish.
Example usage:
```ruby
module API
module V1
module Article
module Operation
class Create < Pragma::Operation::Base
# Let the macro figure out class names.
step Pragma::Macro::Classes()
step :execute!
# But override the contract.
self['contract.default.class'] = Contract::CustomCreate
def execute!(options)
# `options` contains the following:
#
# `model.class`
# `policy.default.class`
# `policy.default.scope.class`
# `decorator.instance.class`
# `decorator.collection.class`
# `contract.default.class`
#
# These will be `nil` if the expected classes do not exist.
end
end
end
end
end
end
```
### Model
**Used in:** Index, Show, Create, Update, Destroy
The `Model` macro provides support for performing different operations with models. It can either
build a new instance of the model, if you are creating a new record, for instance, or it can find
an existing record by ID.
Example of building a new record:
```ruby
module API
module V1
module Article
module Operation
class Create < Pragma::Operation::Base
# This step can be done by Classes if you want.
self['model.class'] = ::Article
step Pragma::Macro::Model(:build)
step :save!
def save!(options)
# Here you'd usually validate and assign parameters before saving.
# ...
options['model'].save!
end
end
end
end
end
end
```
As we mentioned, `Model` can also be used to find a record by ID:
```ruby
module API
module V1
module Article
module Operation
class Show < Pragma::Operation::Base
# This step can be done by Classes if you want.
self['model.class'] = ::Article
step Pragma::Macro::Model(:find_by), fail_fast: true
step :respond!
def respond!(options)
options['result.response'] = Response::Ok.new(
entity: options['model']
)
end
end
end
end
end
end
```
In the example above, if the record is not found, the macro will respond with `404 Not Found` and a
descriptive error message for you. If you want to override the error handling logic, you can remove
the `fail_fast` option and instead implement your own `failure` step.
### Policy
**Used in:** Index, Show, Create, Update, Destroy
The `Policy` macro ensures that the current user can perform an operation on a given record.
Here's a usage example:
```ruby
module API
module V1
module Article
module Operation
class Show < Pragma::Operation::Base
# This step can be done by Classes if you want.
self['policy.default.class'] = Policy
step :model!
step Pragma::Macro::Policy(), fail_fast: true
# You can also specify a custom method to call on the policy:
# step Pragma::Macro::Policy(action: :custom_method), fail_fast: true
step :respond!
def model!(params:, **)
options['model'] = ::Article.find(params[:id])
end
end
end
end
end
end
```
If the user is not authorized to perform the operation (i.e. if the policy's `#show?` method returns
`false`), the macro will respond with `403 Forbidden` and a descriptive error message. If you want
to override the error handling logic, you can remove the `fail_fast` option and instead implement
your own `failure` step.
The macro accepts the following options, which can be defined on the operation or at runtime:
- `policy.context`: the context to use for the policy (optional, `current_user` is used if not
provided).
### Ordering
**Used in:** Index
As the name suggests, the `Ordering` macro allows you to easily implement default and user-defined
ordering.
Here's an example:
```ruby
module API
module V1
module Article
module Operation
class Index < Pragma::Operation::Base
# This step can be done by Classes if you want.
self['model.class'] = ::Article
self['ordering.default_column'] = :published_at
self['ordering.default_direction'] = :desc
self['ordering.columns'] = %i[title published_at updated_at]
step :model!
# This will override `model` with the ordered relation.
step Pragma::Macro::Ordering(), fail_fast: true
step :respond!
def model!(options)
options['model'] = options['model.class'].all
end
def respond!(options)
options['result.response'] = Response::Ok.new(
entity: options['model']
)
end
end
end
end
end
end
```
If the user provides an invalid order column or direction, the macro will respond with `422 Unprocessable Entity`
and a descriptive error message. If you wish to implement your own error handling logic, you can
remove the `fail_fast` option and implement your own `failure` step.
The macro accepts the following options, which can be defined on the operation or at runtime:
- `ordering.columns`: an array of columns the user can order by.
- `ordering.default_column`: the default column to order by (default: `created_at`).
- `ordering.default_direction`: the default direction to order by (default: `desc`).
- `ordering.column_param`: the name of the parameter which will contain the order column.
- `ordering.direction_param`: the name of the parameter which will contain the order direction.
### Pagination
**Used in:** Index
The `Pagination` macro is responsible for paginating collections of records through
[will_paginate](https://github.com/mislav/will_paginate). It also allows your users to set the
number of records per page.
```ruby
module API
module V1
module Article
module Operation
class Index < Pragma::Operation::Base
# This step can be done by Classes if you want.
self['model.class'] = ::Article
step :model!
# This will override `model` with the paginated relation.
step Pragma::Macro::Pagination(), fail_fast: true
step :respond!
def model!(options)
options['model'] = options['model.class'].all
end
def respond!(options)
options['result.response'] = Response::Ok.new(
entity: options['model']
)
end
end
end
end
end
end
```
In the example above, if the page or per-page number fail validation, the macro will respond with
`422 Unprocessable Entity` and a descriptive error message. If you wish to implement your own error
handling logic, you can remove the `fail_fast` option and implement your own `failure` step.
The macro accepts the following options, which can be defined on the operation or at runtime:
- `pagination.page_param`: the parameter that will contain the page number.
- `pagination.per_page_param`: the parameter that will contain the number of items to include in each page.
- `pagination.default_per_page`: the default number of items per page.
- `pagination.max_per_page`: the max number of items per page.
This macro is best used in conjunction with the [Collection](https://github.com/pragmarb/pragma-decorator#collection)
and [Pagination](https://github.com/pragmarb/pragma-decorator#pagination) modules of
[Pragma::Decorator](https://github.com/pragmarb/pragma-decorator), which will expose all the
pagination metadata.
### Decorator
**Used in:** Index, Show, Create, Update
The `Decorator` macro uses one of your decorators to decorate the model. If you are using
[expansion](https://github.com/pragmarb/pragma-decorator#associations), it will also make sure that
the expansion parameter is valid.
Example usage:
```ruby
module API
module V1
module Article
module Operation
class Show < Pragma::Operation::Base
# This step can be done by Classes if you want.
self['decorator.instance.class'] = Decorator::Instance
step :model!
step Pragma::Macro::Decorator(), fail_fast: true
step :respond!
def model!(params:, **)
options['model'] = ::Article.find(params[:id])
end
def respond!(options)
# Pragma does this for you in the default operations.
options['result.response'] = Response::Ok.new(
entity: options['result.decorator.instance']
)
end
end
end
end
end
end
```
The macro accepts the following options, which can be defined on the operation or at runtime:
- `expand.enabled`: whether associations can be expanded.
- `expand.limit`: how many associations can be expanded at once.
## Contributing
Bug reports and pull requests are welcome on GitHub at https://github.com/pragmarb/pragma.
## License
The gem is available as open source under the terms of the [MIT License](http://opensource.org/licenses/MIT).