action_logic 0.0.6 → 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CODE_OF_CONDUCT.md +50 -0
- data/Gemfile.lock +1 -1
- data/README.md +823 -83
- data/lib/action_logic/action_use_case.rb +2 -0
- data/lib/action_logic/errors.rb +12 -0
- data/lib/action_logic/version.rb +1 -1
- data/resources/action_coordinator_diagram.png +0 -0
- data/resources/action_task_diagram.png +0 -0
- data/resources/action_use_case_diagram.png +0 -0
- data/resources/diagrams.sketch +0 -0
- data/resources/overview_diagram.png +0 -0
- data/spec/action_logic/active_use_case_spec.rb +63 -0
- data/spec/fixtures/use_cases.rb +109 -15
- metadata +7 -1
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: c17b002948787563a10bc653bf44ee73f88e989d
|
4
|
+
data.tar.gz: cf88c4dfb9009e45b222acd03b7ec3f8eef9cc89
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 47dfecfa70d4193a4c69b418f747db4bd98e17178fa5c22e3a42ee37db724c8e9d9c8bb82e66360ba4d1f02ab2ecfff0c2b2c13a53d38bd9482e8a9dda502508
|
7
|
+
data.tar.gz: d834c1c3f475660f8cb03342aae902c85b8d3c5c8675bf4574d8df0b50aa71b4a929828bb8ca5fd21739c75b46d3c7412a78816e781f13dc2ecd52f2c15bf288
|
data/CODE_OF_CONDUCT.md
ADDED
@@ -0,0 +1,50 @@
|
|
1
|
+
# Contributor Code of Conduct
|
2
|
+
|
3
|
+
As contributors and maintainers of this project, and in the interest of
|
4
|
+
fostering an open and welcoming community, we pledge to respect all people who
|
5
|
+
contribute through reporting issues, posting feature requests, updating
|
6
|
+
documentation, submitting pull requests or patches, and other activities.
|
7
|
+
|
8
|
+
We are committed to making participation in this project a harassment-free
|
9
|
+
experience for everyone, regardless of level of experience, gender, gender
|
10
|
+
identity and expression, sexual orientation, disability, personal appearance,
|
11
|
+
body size, race, ethnicity, age, religion, or nationality.
|
12
|
+
|
13
|
+
Examples of unacceptable behavior by participants include:
|
14
|
+
|
15
|
+
* The use of sexualized language or imagery
|
16
|
+
* Personal attacks
|
17
|
+
* Trolling or insulting/derogatory comments
|
18
|
+
* Public or private harassment
|
19
|
+
* Publishing other's private information, such as physical or electronic
|
20
|
+
addresses, without explicit permission
|
21
|
+
* Other unethical or unprofessional conduct
|
22
|
+
|
23
|
+
Project maintainers have the right and responsibility to remove, edit, or
|
24
|
+
reject comments, commits, code, wiki edits, issues, and other contributions
|
25
|
+
that are not aligned to this Code of Conduct, or to ban temporarily or
|
26
|
+
permanently any contributor for other behaviors that they deem inappropriate,
|
27
|
+
threatening, offensive, or harmful.
|
28
|
+
|
29
|
+
By adopting this Code of Conduct, project maintainers commit themselves to
|
30
|
+
fairly and consistently applying these principles to every aspect of managing
|
31
|
+
this project. Project maintainers who do not follow or enforce the Code of
|
32
|
+
Conduct may be permanently removed from the project team.
|
33
|
+
|
34
|
+
This code of conduct applies both within project spaces and in public spaces
|
35
|
+
when an individual is representing the project or its community.
|
36
|
+
|
37
|
+
Instances of abusive, harassing, or otherwise unacceptable behavior may be
|
38
|
+
reported by contacting a project maintainer at rick.winfrey@gmail.com. All
|
39
|
+
complaints will be reviewed and investigated and will result in a response that
|
40
|
+
is deemed necessary and appropriate to the circumstances. Maintainers are
|
41
|
+
obligated to maintain confidentiality with regard to the reporter of an
|
42
|
+
incident.
|
43
|
+
|
44
|
+
|
45
|
+
This Code of Conduct is adapted from the [Contributor Covenant][homepage],
|
46
|
+
version 1.3.0, available at
|
47
|
+
[http://contributor-covenant.org/version/1/3/0/][version]
|
48
|
+
|
49
|
+
[homepage]: http://contributor-covenant.org
|
50
|
+
[version]: http://contributor-covenant.org/version/1/3/0/
|
data/Gemfile.lock
CHANGED
data/README.md
CHANGED
@@ -7,165 +7,905 @@
|
|
7
7
|
|
8
8
|
### Introduction
|
9
9
|
|
10
|
-
This is a business logic abstraction gem that provides structure to the organization and composition of business logic in a Ruby or Rails application. `ActionLogic` is inspired by
|
11
|
-
|
12
|
-
Why another business logic abstraction gem? `ActionLogic`
|
10
|
+
This is a business logic abstraction gem that provides structure to the organization and composition of business logic in a Ruby or Rails application. `ActionLogic` is inspired by gems like [ActiveInteraction](https://github.com/orgsync/active_interaction), [DecentExposure](https://github.com/hashrocket/decent_exposure), [Interactor](https://github.com/collectiveidea/interactor), [Light-Service](https://github.com/adomokos/light-service), [Mutations](https://github.com/cypriss/mutations), [Surrounded](https://github.com/saturnflyer/surrounded), [Trailblazer](https://github.com/apotonick/trailblazer) and [Wisper](https://github.com/krisleech/wisper).
|
11
|
+
|
12
|
+
Why another business logic abstraction gem? `ActionLogic` provides teams of various experience levels with a minimal yet powerful set of abstractions that promote easy to write and easy to understand code. By using `ActionLogic`, teams can more quickly and easily write business logic that honors the SOLID principles, is easy to test and easy to reason about, and provides a flexible foundation from which teams can model and define their application's business domains by focusing on reusable units of work that can be composed and validated with one another.
|
13
|
+
|
14
|
+
### Contents
|
15
|
+
|
16
|
+
* [Backstory](#backstory)
|
17
|
+
* [Overview](#overview)
|
18
|
+
* [`ActionContext`](#actioncontext)
|
19
|
+
* [`ActionTask`](#actiontask)
|
20
|
+
* [`ActionUseCase`](#actionusecase)
|
21
|
+
* [`ActionCoordinator`](#actioncoordinator)
|
22
|
+
* [Succeeding an `ActionContext`](#succeeding-an-actioncontext)
|
23
|
+
* [Failing an `ActionContext`](#failing-an-actioncontext)
|
24
|
+
* [Halting an `ActionContext`](#halting-an-actioncontext)
|
25
|
+
* [Custom `ActionContext` Status](#custom-actioncontext)
|
26
|
+
* [Error Handling](#error-handling)
|
27
|
+
* [Attribute Validations](#attribute-validations)
|
28
|
+
* [Type Validations](#type-validations)
|
29
|
+
* [Custom Type Validations](#custom-type-validations)
|
30
|
+
* [Presence Validations](#presence-validations)
|
31
|
+
* [Custom Presence Validations](#custom-presence-validations)
|
32
|
+
* [Before Validations](#before-validations)
|
33
|
+
* [After Validations](#after-validations)
|
34
|
+
* [Around Validations](#around-validations)
|
35
|
+
|
36
|
+
### Backstory
|
37
|
+
|
38
|
+
Consider a traditional e-commerce Rails application. Users can shop online and add items to their shopping cart until they are ready to check out.
|
39
|
+
The happy path scenario might go something like this: the user submits their order form, an orders controller action records the order in the database,
|
40
|
+
submits the order total to a payment processor, waits for a response from the payment processor, and upon a success response from the payment processor sends
|
41
|
+
an order confirmation email to the user, the order is sent internally to the warehouse for fulfillment which requires creating various records in the database,
|
42
|
+
and finally the server responds to the initial POST request with a rendered html page including a message indicating the order was successfully processed. In this
|
43
|
+
work flow there are at least 7 distinct steps or tasks that must be satisfied in order for the application's business logic to be considered correct according
|
44
|
+
to specifications.
|
45
|
+
|
46
|
+
Although this flow works well for most users, there are other users whose credit card information might be expired or users who might attempt to check out when
|
47
|
+
your application's payment processor service is down. Additional edge case scenarios start to pop up in error logs as exception emails fill up your inbox.
|
48
|
+
What happens when that user that is notorious for having 100 tabs open forgets to complete the checkout process and submits a two week old order form that
|
49
|
+
includes an item that your e-commerce store no longer stocks? What happens if an item is sold out? The edge cases and exception emails pile up, and as each one comes in
|
50
|
+
you add more and more logic to that controller action.
|
51
|
+
|
52
|
+
What once was a simple controller action designed with only the happy path of a successful checkout in mind has now become 100 lines long with 5 to 10 levels
|
53
|
+
of nested if statements. The voice of Uncle Bob starts ringing in your ears and you know there must be a better way. You think on it for awhile and consider not only
|
54
|
+
the technical challenges of refactoring this code, but you'd also like to make this code reusable and modular. You want this code to be easy to test and easy to maintain.
|
55
|
+
You want to honor the SOLID principles by writing classes that are singularly focused and easy to extend. You reason these new classes should only have to change if the
|
56
|
+
business logic they execute changes. You see that there are relationships between the entities and you see the possibility of abstractions that allow entities of similar types
|
57
|
+
to interact nicely with each other. You begin thinking about interfaces and the Liskov Substitution Principle, and eventually your mind turns towards domains and data modeling.
|
58
|
+
Where does it end you wonder?
|
59
|
+
|
60
|
+
But you remember your team. It's a team of people all wanting to do their best, and represent a variety of backgrounds and experiences. Each person has varying degress of familiarity
|
61
|
+
with different types of abstractions and approaches, and you wonder what abstractions might be as easy to work with for a new developer as they are for an experienced developer?
|
62
|
+
You consider DSL's you've used in the past and wonder what is that ideal balance between magic and straightforward OOP design?
|
63
|
+
|
64
|
+
As more and more questions pile up in the empty space of your preferred text editor, you receive another exception email for a new problem with the order flow. The questions about
|
65
|
+
how to refactor this code transform into asking questions about how can you edit the existing code to add the new fix? Add a new nested if statement? You do what you can given the
|
66
|
+
constraints you're faced with, and add another 5 lines and another nested if statement. You realize there is not enough time to make this refactor happen, and you've got to push the
|
67
|
+
fix out as soon as possible. Yet, as you merge your feature branch in master and deploy a hotfix, you think surely there must be a better way.
|
68
|
+
|
69
|
+
`ActionLogic` was born from many hours thinking about these questions and considering how it might be possible to achieve a generic set of abstractions to help guide
|
70
|
+
business logic that would promote the SOLID principles and be easy for new and experienced developers to understand and extend. It's not a perfect abstraction (as nothing is),
|
71
|
+
but *can* help simplify your application's business logic by encouraging you to consider the smallest units of work required for your business logic while offering features
|
72
|
+
like type and presence validation that help reduce or eliminate boiler plate, defensive code (nil checks anyone?). However, as with all general purpose libraries, your mileage
|
73
|
+
will vary.
|
13
74
|
|
14
75
|
### Overview
|
15
76
|
|
16
77
|
There are three levels of abstraction provided by `ActionLogic`:
|
17
78
|
|
18
|
-
* `ActionTask` (
|
19
|
-
* `ActionUseCase` (
|
20
|
-
* `ActionCoordinator` (
|
79
|
+
* [`ActionTask` (a concrete unit of work)](#action_task)
|
80
|
+
* [`ActionUseCase` (organizes two or more `ActionTasks`)](#action_use_case)
|
81
|
+
* [`ActionCoordinator` (coordinates two or more `ActionUseCases`)](#action_coordinator)
|
82
|
+
|
83
|
+
Each level of abstraction operates with a shared, mutable data structure referred to as a `context` and is an instance of `ActionContext`. This shared `context` is threaded
|
84
|
+
through each `ActionTask`, `ActionUseCase` and / or `ActionCoordinator` until all work is completed. The resulting `context` is returned to the original caller
|
85
|
+
(typically in a Rails application this will be a controller action). In the problem described above we might have an `ActionUseCase` for organizing the checkout order flow,
|
86
|
+
and each of the distinct steps would be represented by a separate `ActionTask`. However, overtime it may make more sense to split apart the singular `ActionUseCase` for the order
|
87
|
+
flow into smaller `ActionUseCases` that are isolated by their domain (users, payment processor, inventory / warehouse, email, etc.). Considering that we limit our `ActionUseCases` to
|
88
|
+
single domains, then the `ActionCoordinator` abstraction would allow us to coordinate communication between the `ActionUseCases` and their `ActionTasks` to fulfill the necessary
|
89
|
+
work required when a user submits a checkout order form.
|
90
|
+
|
91
|
+
The diagram below illustrates how the `ActionTask`, `ActionUseCase` and `ActionCoordinator` abstractions work together, and the role of `ActionContext` as the primary, single input:
|
92
|
+
|
93
|
+
<img src="https://raw.githubusercontent.com/rewinfrey/action_logic/master/resources/overview_diagram.png" />
|
94
|
+
|
95
|
+
### ActionContext
|
96
|
+
|
97
|
+
The glue that binds the three layers of abstraction provided in `ActionLogic` is `ActionContext`. Anytime an `ActionTask`, `ActionUseCase` or `ActionCoordinator` is invoked
|
98
|
+
an instance of `ActionContext` is created and passed as an input parameter to the receiving execution context. Because each of the three abstractions works in the same way
|
99
|
+
with `ActionContext`, it is intended to be a relatively simple "learn once understand everywhere" abstraction.
|
100
|
+
|
101
|
+
Instances of `ActionContext` are always referred to within the body of `call` methods defined in any `ActionTask`, `ActionUseCase` or `ActionCoordinator` as `context`. An
|
102
|
+
instance of `ActionContext` is a thin wrapper around Ruby's standard library [`OpenStruct`](http://ruby-doc.org/stdlib-2.0.0/libdoc/ostruct/rdoc/OpenStruct.html). This allows
|
103
|
+
instances of `ActionContext` to be maximally flexible. Arbitrary attributes can be defined on a `context` and their values can be of any type.
|
104
|
+
|
105
|
+
In addition to allowing arbitrary attributes and values to be defined on a `context`, instances of `ActionContext` also conform to a set of simple rules:
|
106
|
+
|
107
|
+
* Every `context` instance is instantiated with a default `status` of `:success`
|
108
|
+
* A `context` responds to `success?` which returns true if the `status` is `:success`
|
109
|
+
* A `context` responds to `fail!` which sets the `status` to `:failure`
|
110
|
+
* A `context` responds to `fail?` which returns true if the `status` is `:failure`
|
111
|
+
* A `context` rseponds to `halt!` which sets the `status` to `:halted`
|
112
|
+
* A `context` responds to `halted?` which returns true if the `status` is `:halted`
|
113
|
+
|
114
|
+
Enough with the words, let's look at some code! The following shows an instance of `ActionContext` and its various abilities:
|
115
|
+
|
116
|
+
```ruby
|
117
|
+
context = ActionLogic::ActionContext.new
|
118
|
+
|
119
|
+
context # => #<ActionLogic::ActionContext status=:success>
|
120
|
+
|
121
|
+
# default status is `:success`:
|
122
|
+
context.status # => :success
|
123
|
+
|
124
|
+
# defining a new attribute called `name` with the value `"Example"`:
|
125
|
+
context.name = "Example"
|
126
|
+
|
127
|
+
# retrieving the value of the `name` attribute:
|
128
|
+
context.name # => "Example"
|
129
|
+
|
130
|
+
# you can set attributes to anything, including Procs:
|
131
|
+
context.lambda_example = -> { "here" }
|
132
|
+
|
133
|
+
context.lambda_example # => #<Proc:0x007f8d6b0a9ba0@-:11 (lambda)>
|
134
|
+
|
135
|
+
context.lambda_example.call # => "here"
|
21
136
|
|
22
|
-
|
137
|
+
# contexts can be failed:
|
138
|
+
context.fail!
|
139
|
+
|
140
|
+
context.status # => :failure
|
141
|
+
|
142
|
+
context.failure? # => true
|
143
|
+
|
144
|
+
# contexts can also be halted:
|
145
|
+
context.halt!
|
146
|
+
|
147
|
+
context.status # => :halted
|
148
|
+
|
149
|
+
context.halted? # => true
|
150
|
+
```
|
151
|
+
|
152
|
+
Now that we have seen what `ActionContext` can do, let's take a look at the lowest level of absraction in `ActionLogic` that consumes instances of `ActionContext`, the `ActionTask`
|
153
|
+
abstraction.
|
23
154
|
|
24
155
|
### ActionTask
|
25
156
|
|
26
|
-
At the core of every `ActionLogic` work flow is an `ActionTask`. These
|
157
|
+
At the core of every `ActionLogic` work flow is an `ActionTask`. These classes are the lowest level of abstraction in `ActionLogic` and are where concrete work is performed. All `ActionTasks` conform to the same structure and incorporate all features of `ActionLogic` including validations and error handling.
|
27
158
|
|
28
|
-
|
159
|
+
To implement an `ActionTask` class you must define a `call` method. You can also specify any before, after or around validations or an error handler. The following code example demonstrates an `ActionTask` class that includes before and after validations, and also demonstrates how an `ActionTask` is invoked :
|
29
160
|
|
30
161
|
```ruby
|
31
162
|
class ActionTaskExample
|
32
163
|
include ActionLogic::ActionTask
|
33
|
-
|
164
|
+
|
165
|
+
validates_before :expected_attribute1 => { :type => :string },
|
166
|
+
:expected_attribute2 => { :type => :integer, :presence => true }
|
167
|
+
validates_after :example_attribute1 => { :type => :string, :presence => ->(example_attribute1) { !example_attribute1.empty? } }
|
168
|
+
|
34
169
|
def call
|
35
|
-
|
36
|
-
context.
|
170
|
+
# adds `example_attribute1` to the shared `context` with the value "Example value"
|
171
|
+
context.example_attribute1 = "New value from context attributes: #{context.expected_attribute1} #{context.expected_attribute2}"
|
37
172
|
end
|
38
173
|
end
|
39
|
-
```
|
40
174
|
|
41
|
-
|
175
|
+
# ActionTasks are invoked by calling an `execute` static method directly on the class with an optional hash of key value pairs:
|
176
|
+
result = ActionTaskExample.execute(:expected_attribute1 => "example", :expected_attribute2 => 123)
|
42
177
|
|
43
|
-
|
44
|
-
result =
|
45
|
-
result # => <ActionContext :success=true, :example_attribute1="Example value", :example_attribute2=123, :message = "">
|
178
|
+
# The result object is the shared context object (an instance of ActionContext):
|
179
|
+
result # => #<ActionLogic::ActionContext expected_attribute1="example", expected_attribute2=123, status=:success, example_attribute1="New value from context attributes: example 123">
|
46
180
|
```
|
47
181
|
|
48
|
-
|
182
|
+
The `ActionTaskExample` is invoked using the static method `execute` which takes an optional hash of attributes that is converted into an `ActionContext`. Assuming the before validations are satisfied, the `call` method is invoked. In the body of the `call` method the `ActionTask` can access the shared `ActionContext` instance via a `context` object. This shared `context` object allows for getting and setting attributes as needed. When the `call` method returns, the `context` is validated against any defined after validations, and the `context` is then returned to the caller.
|
183
|
+
|
184
|
+
The diagram below is a visual representation of how an `ActionTask` is evaluted when its `execute` method is invoked from a caller:
|
185
|
+
|
186
|
+
<img src="https://raw.githubusercontent.com/rewinfrey/action_logic/master/resources/action_task_diagram.png" />
|
187
|
+
|
188
|
+
Although this example is for the `ActionTask` abstraction, `ActionUseCase` and `ActionCoordinator` follow the same pattern. The difference is that `ActionUseCase` is designed to organize multiple `ActionTasks`, and `ActionCoordinator` is designed to organize many `ActionUseCases`.
|
49
189
|
|
50
190
|
### ActionUseCase
|
51
191
|
|
52
|
-
|
192
|
+
As business logic grows in complexity the number of steps or tasks required to fulfill that business logic tends to increase. Managing this complexity is a problem every team must face. Abstractions can help teams of varying experience levels work together and promote code that remains modular and simple to understand and extend. `ActionUseCase` represents a layer of abstraction that organizes multiple `ActionTasks` and executes each `ActionTask` in the order they are defined. Each task receives the same shared `context` so tasks can be composed together.
|
193
|
+
|
194
|
+
To implement an `ActionUseCase` class you must define a `call` method and a `tasks` method. You also can specify any before, after or around validations or an error handler. The following is an example showcasing how an `ActionUseCase` class organizes the execution of multiple `ActionTasks` and defines before and after validations on the shared `context`:
|
53
195
|
|
54
196
|
```ruby
|
55
197
|
class ActionUseCaseExample
|
56
198
|
include ActionLogic::ActionUseCase
|
57
|
-
|
199
|
+
|
200
|
+
validates_before :expected_attribute1 => { :type => :string },
|
201
|
+
:expected_attribute2 => { :type => :integer, :presence => true }
|
202
|
+
validates_after :example_task1 => { :type => :boolean, :presence => true },
|
203
|
+
:example_task2 => { :type => :boolean, :presence => true },
|
204
|
+
:example_task3 => { :type => :boolean, :presence => true },
|
205
|
+
:example_usecase1 => { :type => :boolean, :presence => true }
|
206
|
+
|
58
207
|
# The `call` method is invoked prior to invoking any of the ActionTasks defined by the `tasks` method.
|
59
208
|
# The purpose of the `call` method allows us to prepare the shared `context` prior to invoking the ActionTasks.
|
60
209
|
def call
|
61
|
-
context
|
62
|
-
context.
|
210
|
+
context # => #<ActionLogic::ActionContext expected_attribute1="example", expected_attribute2=123, status=:success>
|
211
|
+
context.example_usecase1 = true
|
63
212
|
end
|
64
|
-
|
213
|
+
|
65
214
|
def tasks
|
66
215
|
[ActionTaskExample1,
|
67
216
|
ActionTaskExample2,
|
68
217
|
ActionTaskExample3]
|
69
218
|
end
|
70
219
|
end
|
220
|
+
|
221
|
+
class ActionTaskExample1
|
222
|
+
include ActionLogic::ActionTask
|
223
|
+
validates_after :example_task1 => { :type => :boolean, :presence => true }
|
224
|
+
|
225
|
+
def call
|
226
|
+
context # => #<ActionLogic::ActionContext expected_attribute1="example", expected_attribute2=123, status=:success, example_usecase1=true>
|
227
|
+
context.example_task1 = true
|
228
|
+
end
|
229
|
+
end
|
230
|
+
|
231
|
+
class ActionTaskExample2
|
232
|
+
include ActionLogic::ActionTask
|
233
|
+
validates_after :example_task2 => { :type => :boolean, :presence => true }
|
234
|
+
|
235
|
+
def call
|
236
|
+
context # => #<ActionLogic::ActionContext expected_attribute1="example", expected_attribute2=123, status=:success, example_usecase1=true, example_task1=true>
|
237
|
+
context.example_task2 = true
|
238
|
+
end
|
239
|
+
end
|
240
|
+
|
241
|
+
class ActionTaskExample3
|
242
|
+
include ActionLogic::ActionTask
|
243
|
+
validates_after :example_task3 => { :type => :boolean, :presence => true }
|
244
|
+
|
245
|
+
def call
|
246
|
+
context # => #<ActionLogic::ActionContext expected_attribute1="example", expected_attribute2=123, status=:success, example_usecase1=true, example_task1=true, example_task2=true>
|
247
|
+
context.example_task3 = true
|
248
|
+
end
|
249
|
+
end
|
250
|
+
|
251
|
+
# To invoke the ActionUseCaseExample, we call its execute method with the required attributes:
|
252
|
+
result = ActionUseCaseExample.execute(:expected_attribute1 => "example", :expected_attribute2 => 123)
|
253
|
+
|
254
|
+
result # => #<ActionLogic::ActionContext expected_attribute1="example", expected_attribute2=123, status=:success, example_usecase1=true, example_task1=true, example_task2=true, example_task3=true>
|
255
|
+
```
|
256
|
+
|
257
|
+
By following the value of the shared `context` from the `ActionUseCaseExample` to each of the `ActionTask` classes, it is possible to see how the shared `context` is mutated to accomodate the various attributes and their values each execution context adds to the `context`. It also reveals the order in which the `ActionTasks` are evaluated, and indicates that the `call` method of the `ActionUseCaseExample` is invoked prior to any of the `ActionTasks` defined in the `tasks` method.
|
258
|
+
|
259
|
+
To help visualize the flow of execution when an `ActionUseCase` is invoked, this diagram aims to illustrate the relationship between `ActionUseCase` and `ActionTasks` and the order in which operations are performed:
|
260
|
+
|
261
|
+
<img src="https://raw.githubusercontent.com/rewinfrey/action_logic/master/resources/action_use_case_diagram.png" />
|
262
|
+
|
263
|
+
### ActionCoordinator
|
264
|
+
|
265
|
+
Sometimes the behavior we wish our Ruby or Rails application to provide requires us to coordinate work between various domains of our application's business logic. The `ActionCoordinator` abstraction is intended to help coordinate multiple `ActionUseCases` by allowing you to define a plan of which `ActionUseCases` to invoke depending on the outcome of each `ActionUseCase` execution. The `ActionCoordinator` abstraction is the highest level of abstraction in `ActionLogic`.
|
266
|
+
|
267
|
+
To implement an `ActionCoordinator` class, you must define a `call` method in addition to a `plan` method. The purpose of the `plan` method is to define a state transition map that links together the various `ActionUseCase` classes the `ActionCoordinator` is organizing, as well as allowing you to define error or halt scenarios based on the result of each `ActionUseCase`. The following code example demonstrates a simple `ActionCoordinator`:
|
268
|
+
|
269
|
+
```ruby
|
270
|
+
class ActionCoordinatorExample
|
271
|
+
include ActionLogic::ActionCoordinator
|
272
|
+
|
273
|
+
def call
|
274
|
+
context.required_attribute1 = "required attribute 1"
|
275
|
+
context.required_attribute2 = "required attribute 2"
|
276
|
+
end
|
277
|
+
|
278
|
+
def plan
|
279
|
+
{
|
280
|
+
ActionUseCaseExample1 => { :success => ActionUseCaseExample2,
|
281
|
+
:failure => ActionUseCaseFailureExample },
|
282
|
+
ActionUseCaseExample2 => { :success => nil,
|
283
|
+
:failure => ActionUseCaseFailureExample },
|
284
|
+
ActionUseCaseFailureExample => { :success => nil }
|
285
|
+
}
|
286
|
+
end
|
287
|
+
end
|
288
|
+
|
289
|
+
class ActionUseCaseExample1
|
290
|
+
include ActionLogic::ActionUseCase
|
291
|
+
|
292
|
+
validates_before :required_attribute1 => { :type => :string }
|
293
|
+
|
294
|
+
def call
|
295
|
+
context # => #<ActionLogic::ActionContext status=:success, required_attribute1="required attribute 1", required_attribute2="required attribute 2">
|
296
|
+
context.example_usecase1 = true
|
297
|
+
end
|
298
|
+
|
299
|
+
# Normally `tasks` would define multiple tasks, but in this example, I've used one ActionTask to keep the overall code example smaller
|
300
|
+
def tasks
|
301
|
+
[ActionTaskExample1]
|
302
|
+
end
|
303
|
+
end
|
304
|
+
|
305
|
+
class ActionUseCaseExample2
|
306
|
+
include ActionLogic::ActionUseCase
|
307
|
+
|
308
|
+
validates_before :required_attribute2 => { :type => :string }
|
309
|
+
|
310
|
+
def call
|
311
|
+
context # => #<ActionLogic::ActionContext status=:success, required_attribute1="required attribute 1", required_attribute2="required attribute 2", example_usecase1=true, example_task1=true>
|
312
|
+
context.example_usecase2 = true
|
313
|
+
end
|
314
|
+
|
315
|
+
# Normally `tasks` would define multiple tasks, but in this example, I've used one ActionTask to keep the overall code example smaller
|
316
|
+
def tasks
|
317
|
+
[ActionTaskExample2]
|
318
|
+
end
|
319
|
+
end
|
320
|
+
|
321
|
+
# In this example, we are not calling ActionUseCaseFailureExample, but is used to illustrate the purpose of the `plan` of our ActionCoordinator
|
322
|
+
# in the event of a failure in one of the consumed `ActionUseCases`
|
323
|
+
class ActionUseCaseFailureExample
|
324
|
+
include ActionLogic::ActionUseCase
|
325
|
+
|
326
|
+
def call
|
327
|
+
end
|
328
|
+
|
329
|
+
def tasks
|
330
|
+
[ActionTaskLogFailure,
|
331
|
+
ActionTaskEmailFailure]
|
332
|
+
end
|
333
|
+
end
|
334
|
+
|
335
|
+
class ActionTaskExample1
|
336
|
+
include ActionLogic::ActionTask
|
337
|
+
validates_after :example_task1 => { :type => :boolean, :presence => true }
|
338
|
+
|
339
|
+
def call
|
340
|
+
context # => #<ActionLogic::ActionContext status=:success, required_attribute1="required attribute 1", required_attribute2="required attribute 2", example_usecase1=true>
|
341
|
+
context.example_task1 = true
|
342
|
+
end
|
343
|
+
end
|
344
|
+
|
345
|
+
class ActionTaskExample2
|
346
|
+
include ActionLogic::ActionTask
|
347
|
+
validates_after :example_task2 => { :type => :boolean, :presence => true }
|
348
|
+
|
349
|
+
def call
|
350
|
+
context # => #<ActionLogic::ActionContext status=:success, required_attribute1="required attribute 1", required_attribute2="required attribute 2", example_usecase1=true, example_task1=true, example_usecase2=true>
|
351
|
+
context.example_task2 = true
|
352
|
+
end
|
353
|
+
end
|
354
|
+
|
355
|
+
result = ActionCoordinatorExample.execute
|
356
|
+
|
357
|
+
result # => #<ActionLogic::ActionContext status=:success, required_attribute1="required attribute 1", required_attribute2="required attribute 2", example_usecase1=true, example_task1=true, example_usecase2=true, example_task2=true>
|
71
358
|
```
|
72
359
|
|
73
|
-
|
360
|
+
<img src="https://raw.githubusercontent.com/rewinfrey/action_logic/master/resources/action_coordinator_diagram.png" />
|
74
361
|
|
75
|
-
|
362
|
+
### Succeeding an `ActionContext`
|
363
|
+
By default, the value of the `status` attribute of instances of `ActionContext` is `:success`. Normally this is useful information for the caller of an `ActionTask`, `ActionUseCase` or `ActionCoordinator`
|
364
|
+
because it informs the caller that the various execution context(s) were successful. In other words, a `:success` status indicates that none of the execution contexts had a failure
|
365
|
+
or halted execution.
|
366
|
+
|
367
|
+
### Failing an `ActionContext`
|
368
|
+
Using `context.fail!` does two important things: it immediately stops the execution of any proceeding business logic (prevents any additional `ActionTasks` from executing)
|
369
|
+
and also sets the status of the `context` as `:failure`. This status is most applicable to the caller or an `ActionCoordinator` that might have a plan specifically for a `:failure`
|
370
|
+
status of a resulting `ActionUseCase`.
|
371
|
+
|
372
|
+
The following is a simple example to show how a `context` is failed within a `call` method:
|
76
373
|
|
77
374
|
```ruby
|
78
|
-
|
375
|
+
class ActionTaskExample
|
376
|
+
include ActionLogic::ActionTask
|
377
|
+
|
378
|
+
def call
|
379
|
+
if failure_condition?
|
380
|
+
context.fail!
|
381
|
+
end
|
382
|
+
end
|
383
|
+
|
384
|
+
def failure_condition?
|
385
|
+
true
|
386
|
+
end
|
387
|
+
end
|
388
|
+
|
389
|
+
result = ActionTaskExample.execute
|
390
|
+
|
391
|
+
result # => #<ActionLogic::ActionContext status=:failure, message="">
|
79
392
|
```
|
80
393
|
|
81
|
-
|
394
|
+
When failing a `context` it is possible to also specify a message:
|
82
395
|
|
83
|
-
|
396
|
+
```ruby
|
397
|
+
class ActionTaskExample
|
398
|
+
include ActionLogic::ActionTask
|
399
|
+
|
400
|
+
def call
|
401
|
+
if failure_condition?
|
402
|
+
context.fail! "Something was invalid"
|
403
|
+
end
|
404
|
+
end
|
405
|
+
|
406
|
+
def failure_condition?
|
407
|
+
true
|
408
|
+
end
|
409
|
+
end
|
410
|
+
|
411
|
+
result = ActionTaskExample.execute
|
412
|
+
|
413
|
+
result # => #<ActionLogic::ActionContext status=:failure, message="Something was invalid">
|
414
|
+
|
415
|
+
result.message # => "Something was invalid"
|
416
|
+
```
|
417
|
+
|
418
|
+
From the above example we see how it is possible to `fail!` a `context` while also specifying a clarifying message about the failure condition. Later, we retrieve
|
419
|
+
that failure message via the `message` attribute defined on the returned `context`.
|
420
|
+
|
421
|
+
### Halting an `ActionContext`
|
422
|
+
Like, failing a context, Using `context.halt!` does two important things: it immediately halts the execution of any proceeding business logic (prevents any additional `ActionTasks`
|
423
|
+
from executing) and also sets the status of the `context` as `:halted`. The caller may use that information to define branching logic or an `ActionCoordinator` may use that
|
424
|
+
information as part of its `plan`.
|
425
|
+
|
426
|
+
However, unlike failing a `context`, halting is designed to indicate that no more processing is required, but otherwise execution was successful.
|
427
|
+
|
428
|
+
The following is a simple example to show how a `context` is halted within a `call` method:
|
429
|
+
|
430
|
+
```ruby
|
431
|
+
class ActionTaskExample
|
432
|
+
include ActionLogic::ActionTask
|
433
|
+
|
434
|
+
def call
|
435
|
+
if halt_condition?
|
436
|
+
context.halt!
|
437
|
+
end
|
438
|
+
end
|
439
|
+
|
440
|
+
def halt_condition?
|
441
|
+
true
|
442
|
+
end
|
443
|
+
end
|
444
|
+
|
445
|
+
result = ActionTaskExample.execute
|
446
|
+
|
447
|
+
result # => #<ActionLogic::ActionContext status=:halted, message="">
|
448
|
+
```
|
449
|
+
|
450
|
+
When failing a `context` it is possible to also specify a message:
|
451
|
+
|
452
|
+
```ruby
|
453
|
+
class ActionTaskExample
|
454
|
+
include ActionLogic::ActionTask
|
455
|
+
|
456
|
+
def call
|
457
|
+
if halt_condition?
|
458
|
+
context.halt! "Something required a halt"
|
459
|
+
end
|
460
|
+
end
|
461
|
+
|
462
|
+
def halt_condition?
|
463
|
+
true
|
464
|
+
end
|
465
|
+
end
|
466
|
+
|
467
|
+
result = ActionTaskExample.execute
|
468
|
+
|
469
|
+
result # => #<ActionLogic::ActionContext status=:halted, message="Something required a halt">
|
470
|
+
|
471
|
+
result.message # => "Something required a halt"
|
472
|
+
```
|
473
|
+
|
474
|
+
From the above example we see how it is possible to `halt!` a `context` while also specifying a clarifying message about the halt condition. Later, we retrieve
|
475
|
+
that halt message via the `message` attribute defined on the returned `context`.
|
476
|
+
|
477
|
+
### Custom `ActionContext` Status
|
478
|
+
It is worthwhile to point out that you should not feel limited to only using the three provided statuses of `:success`, `:failure` or `:halted`. It is easy to implement your
|
479
|
+
own system of statuses if you prefer. For example, consider a system that is used to defining various status codes or disposition codes to indicate the result of some business
|
480
|
+
logic. Instances of `ActionContext` can be leveraged to indicate these disposition codes by using the `status` attribute, or by defining custom attributes. You are encouraged
|
481
|
+
to expirement and play with the flexibility provided to you by `ActionContext` in determining what is optimal for your given code contexts and your team.
|
482
|
+
|
483
|
+
```ruby
|
484
|
+
class RailsControllerExample < ApplicationController
|
485
|
+
def create
|
486
|
+
case create_use_case.status
|
487
|
+
when :disposition_1 then ActionUseCaseSuccess1.execute(create_use_case)
|
488
|
+
when :disposition_2 then ActionUseCaseSuccess2.execute(create_use_case)
|
489
|
+
when :disposition_9 then ActionUseCaseFailure.execute(create_use_case)
|
490
|
+
else
|
491
|
+
ActionUseCaseDefault.execute(create_use_case)
|
492
|
+
end
|
493
|
+
end
|
494
|
+
|
495
|
+
private
|
496
|
+
|
497
|
+
def create_use_case
|
498
|
+
@create_use_case ||= ActionUseCaseExample.execute(params)
|
499
|
+
end
|
500
|
+
end
|
501
|
+
```
|
502
|
+
|
503
|
+
Although this contrived example would be ideal for an `ActionCoordinator` (because the result of `ActionUseCaseExample` drives the execution of the next `ActionUseCase`), this
|
504
|
+
example serves to show that `status` can be used with custom disposition codes to drive branching behavior.
|
505
|
+
|
506
|
+
### Error Handling
|
507
|
+
During execution of an `ActionTask`, `ActionUseCase` or `ActionCoordinator` you may wish to define custom behavior for handling errors. Within any of these classes
|
508
|
+
you can define an `error` method that receives as its input the error exception. Invoking an `error` method does not make any assumptions about the `status` of the
|
509
|
+
underlying `context`. Execution of the `ActionTask`, `ActionUseCase` or `ActionCoordinator` also stops after the `error` method returns, and execution of the work
|
510
|
+
flow continues as normal unless the `context` is failed or halted.
|
511
|
+
|
512
|
+
The following example is a simple illustration of how an `error` method is invoked for an `ActionTask`:
|
513
|
+
|
514
|
+
```ruby
|
515
|
+
class ActionTaskExample
|
516
|
+
include ActionLogic::ActionTask
|
517
|
+
|
518
|
+
def call
|
519
|
+
context.before_raise = true
|
520
|
+
raise "Something broke"
|
521
|
+
context.after_raise = true
|
522
|
+
end
|
523
|
+
|
524
|
+
def error(e)
|
525
|
+
context.error = "the error is passed in as an input parameter: #{e.class}"
|
526
|
+
end
|
527
|
+
end
|
528
|
+
|
529
|
+
result = ActionTaskExample.execute
|
84
530
|
|
85
|
-
|
531
|
+
# the status of the context is not mutated
|
532
|
+
result.status # => :success
|
86
533
|
|
87
|
-
|
534
|
+
result.error # => "the error is passed in as an input parameter: RuntimeError"
|
88
535
|
|
89
|
-
|
90
|
-
* Custom error handling defined as a callback
|
91
|
-
* Prematurely halt or fail a workflow
|
536
|
+
result.before_raise # => true
|
92
537
|
|
93
|
-
|
538
|
+
result.after_raise # => nil
|
539
|
+
```
|
540
|
+
|
541
|
+
It is important to note that defining an `error` method is **not** required. If at any point in the execution of an `ActionTask`, `ActionUseCase` or `ActionCoordinator`
|
542
|
+
an uncaught exception is thrown **and** an `error` method is **not** defined, the exception is raised to the caller.
|
543
|
+
|
544
|
+
### Attribute Validations
|
545
|
+
The most simple and basic type of validation offered by `ActionLogic` is attribute validation. To require that an attribute be defined on an instance of `ActionContext`, you
|
546
|
+
need only specify the name of the attribute and an empty hash with one of the three validation types (before, after or around):
|
547
|
+
|
548
|
+
```ruby
|
549
|
+
class ActionTaskExample
|
550
|
+
include ActionLogic::ActionTask
|
551
|
+
|
552
|
+
validates_before :required_attribute1 => {}
|
553
|
+
|
554
|
+
def call
|
555
|
+
end
|
556
|
+
end
|
557
|
+
|
558
|
+
result = ActionTaskExample.execute(:required_attribute1 => true)
|
559
|
+
|
560
|
+
result.status # => :success
|
561
|
+
|
562
|
+
result.required_attribute1 # => true
|
563
|
+
```
|
564
|
+
|
565
|
+
However, in the above example, if we were to invoke the `ActionTaskExample` without the `required_attribute1` parameter, the before validation would fail and raise
|
566
|
+
an `ActionLogic::MissingAttributeError` and also detail which attribute is missing:
|
567
|
+
|
568
|
+
```ruby
|
569
|
+
class ActionTaskExample
|
570
|
+
include ActionLogic::ActionTask
|
571
|
+
|
572
|
+
validates_before :required_attribute1 => {}
|
573
|
+
|
574
|
+
def call
|
575
|
+
end
|
576
|
+
end
|
577
|
+
|
578
|
+
ActionTaskExample.execute # ~> [:required_attribute1] (ActionLogic::MissingAttributeError)
|
579
|
+
```
|
580
|
+
|
581
|
+
Attribute validations are defined in the same way regardless of the timing of the validation ([before](#before-validations), [after](#after-validations) or
|
582
|
+
[around](#around-validations)). Please refer to the relevant sections for examples of their usage.
|
583
|
+
|
584
|
+
### Type Validations
|
585
|
+
In addition to attribute validations, `ActionLogic` also allows you to validate against the type of the value of the attribute you expect to be defined in an instance
|
586
|
+
of `ActionContext`. To understand the default types `ActionLogic` validates against, please see the following example:
|
587
|
+
|
588
|
+
```ruby
|
589
|
+
class ActionTaskExample
|
590
|
+
include ActionLogic::ActionTask
|
591
|
+
|
592
|
+
validates_after :integer_test => { :type => :integer },
|
593
|
+
:float_test => { :type => :float },
|
594
|
+
:string_test => { :type => :string },
|
595
|
+
:bool_test => { :type => :boolean },
|
596
|
+
:hash_test => { :type => :hash },
|
597
|
+
:array_test => { :type => :array },
|
598
|
+
:symbol_test => { :type => :symbol },
|
599
|
+
:nil_test => { :type => :nil }
|
600
|
+
|
601
|
+
def call
|
602
|
+
context.integer_test = 123
|
603
|
+
context.float_test = 1.0
|
604
|
+
context.string_test = "test"
|
605
|
+
context.bool_test = true
|
606
|
+
context.hash_test = {}
|
607
|
+
context.array_test = []
|
608
|
+
context.symbol_test = :symbol
|
609
|
+
context.nil_test = nil
|
610
|
+
end
|
611
|
+
end
|
612
|
+
|
613
|
+
result = ActionTaskExample.execute
|
614
|
+
|
615
|
+
result # => #<ActionLogic::ActionContext status=:success,
|
616
|
+
# integer_test=123,
|
617
|
+
# float_test=1.0,
|
618
|
+
# string_test="test",
|
619
|
+
# bool_test=true,
|
620
|
+
# hash_test={},
|
621
|
+
# array_test=[],
|
622
|
+
# symbol_test=:symbol,
|
623
|
+
# nil_test=nil>
|
624
|
+
```
|
625
|
+
|
626
|
+
It's important to point out that Ruby's `true` and `false` are not `Boolean` but `TrueClass` and `FalseClass` respectively. Additionally, `nil`'s type is `NilClass` in Ruby.
|
627
|
+
To simplify the way these validations work for `true` or `false`, type validations expect the symbol `:boolean` as the `:type`. `nil` is validated simply with the `:nil` `:type`.
|
628
|
+
|
629
|
+
As we saw with attribute validations, if an attribute's value does not conform to the type expected, `ActionLogic` will raise an `ActionLogic::AttributeTypeError`
|
630
|
+
with a detailed description about which attribute's value failed the validation:
|
631
|
+
|
632
|
+
```ruby
|
633
|
+
class ActionTaskExample
|
634
|
+
include ActionLogic::ActionTask
|
635
|
+
|
636
|
+
validates_after :integer_test => { :type => :integer }
|
637
|
+
|
638
|
+
def call
|
639
|
+
context.integer_test = 1.0
|
640
|
+
end
|
641
|
+
end
|
642
|
+
|
643
|
+
ActionTaskExample.execute # ~> ["Attribute: integer_test with value: 1.0 was expected to be of type integer but is float"] (ActionLogic::AttributeTypeError)
|
644
|
+
```
|
645
|
+
|
646
|
+
In addition to the above default types it is possible to also validate against user defined types.
|
647
|
+
|
648
|
+
### Custom Type Validations
|
649
|
+
If you would like to validate the type of attributes on a given `context` with your application's classes, `ActionLogic` is happy to provide that functionality.
|
650
|
+
|
651
|
+
Let's consider the following example:
|
652
|
+
|
653
|
+
```ruby
|
654
|
+
class ExampleClass
|
655
|
+
end
|
656
|
+
|
657
|
+
class ActionTaskExample
|
658
|
+
include ActionLogic::ActionTask
|
659
|
+
|
660
|
+
validates_after :example_attribute => { :type => :exampleclass }
|
661
|
+
|
662
|
+
def call
|
663
|
+
context.example_attribute = ExampleClass.new
|
664
|
+
end
|
665
|
+
end
|
666
|
+
|
667
|
+
result = ActionTaskExample.execute
|
668
|
+
|
669
|
+
result # => #<ActionLogic::ActionContext status=:success, example_attribute=#<ExampleClass:0x007f95d1922bd8>>
|
670
|
+
```
|
671
|
+
|
672
|
+
In the above example, a custom class `ExampleClass` is defined. In order to type validate against this class, the required format for the name of the class is simply
|
673
|
+
the lowercase version of the class as a symbol. `ExampleClass` becomes `:exampleclass`, `UserAttributes` becomes `:userattributes`,
|
674
|
+
`ReallyLongClassNameThatBreaks80ColumnsInVimRule` becomes `:reallylongclassnamethatbreaks80columnsinvimrule` and so on.
|
675
|
+
|
676
|
+
If a custom type validation fails, `ActionLogic` provides the same `ActionLogic::AttributeTypeError` with a detailed explanation about what attribute is in violation
|
677
|
+
of the type validation:
|
678
|
+
|
679
|
+
```ruby
|
680
|
+
class ExampleClass
|
681
|
+
end
|
682
|
+
|
683
|
+
class OtherClass
|
684
|
+
end
|
685
|
+
|
686
|
+
class ActionTaskExample
|
687
|
+
include ActionLogic::ActionTask
|
688
|
+
|
689
|
+
validates_after :example_attribute => { :type => :exampleclass }
|
690
|
+
|
691
|
+
def call
|
692
|
+
context.example_attribute = OtherClass.new
|
693
|
+
end
|
694
|
+
end
|
695
|
+
|
696
|
+
ActionTaskExample.execute # ~> ["Attribute: example_attribute with value: #<OtherClass:0x007fb5ca04edb8> was expected to be of type exampleclass but is otherclass"] (ActionLogic::AttributeTypeError)
|
697
|
+
```
|
698
|
+
|
699
|
+
Attribute and type validations are very helpful, but in some situations this is not enough. Additionally, `ActionLogic` provides presence validation so you can also verify that
|
700
|
+
a given attribute on a context not only has the correct type, but also has a value that is considered `present`.
|
701
|
+
|
702
|
+
### Presence Validations
|
703
|
+
|
704
|
+
`ActionLogic` also allows for presence validation for any attribute on an instance of `ActionContext`. Like other validations, presence validations can be specified in before, after or
|
705
|
+
around validations.
|
706
|
+
|
707
|
+
By default, presence validations simply check to determine if an attribute's value is not `nil` or is not `false`. To define a presence validation, you need only specify `:presence => true`
|
708
|
+
for the attribute you wish to validate against:
|
709
|
+
|
710
|
+
```ruby
|
711
|
+
class ActionTaskExample
|
712
|
+
include ActionLogic::ActionTask
|
713
|
+
|
714
|
+
validates_before :example_attribute => { :presence => true }
|
715
|
+
|
716
|
+
def call
|
717
|
+
end
|
718
|
+
end
|
94
719
|
|
95
|
-
|
720
|
+
result = ActionTaskExample.execute(:example_attribute => 123)
|
96
721
|
|
97
|
-
|
98
|
-
|
99
|
-
|
722
|
+
result # => #<ActionLogic::ActionContext example_attribute=123, status=:success>
|
723
|
+
```
|
724
|
+
|
725
|
+
However, if a presence validation fails, `ActionLogic` will raise an `ActionLogic::PresenceError` with a detailed description about the attribute failing the presence validation
|
726
|
+
and why:
|
727
|
+
|
728
|
+
```ruby
|
729
|
+
class ActionTaskExample
|
730
|
+
include ActionLogic::ActionTask
|
731
|
+
|
732
|
+
validates_before :example_attribute => { :presence => true }
|
733
|
+
|
734
|
+
def call
|
735
|
+
end
|
736
|
+
end
|
100
737
|
|
101
|
-
|
738
|
+
ActionTaskExample.execute(:example_attribute => nil) # ~> ["Attribute: example_attribute is missing value in context but presence validation was specified"] (ActionLogic::PresenceError)
|
739
|
+
```
|
102
740
|
|
103
|
-
|
104
|
-
* After validations are invoked after the execution context is invoked
|
105
|
-
* Aroud validations are invoked before and after the execution context is invoked
|
741
|
+
### Custom Presence Validations
|
106
742
|
|
107
|
-
|
743
|
+
Sometimes when wanting to validate presence of an attribute with an aggregate type (like `Array` or `Hash`), we may want to validate that such a type is not empty. If
|
744
|
+
you wish to validate presence for a type that requires inspecting the value of the attribute, `ActionLogic` allows you the ability to define a custom `Proc` to validate
|
745
|
+
an attribute's value against.
|
108
746
|
|
109
747
|
```ruby
|
110
|
-
class
|
748
|
+
class ActionTaskExample
|
111
749
|
include ActionLogic::ActionTask
|
112
|
-
|
113
|
-
validates_before :
|
114
|
-
|
115
|
-
|
116
|
-
validates_after :attribute3 => { :type => :boolean, :presence => true },
|
117
|
-
:attribute4 => { :type => :string, :presence => true }
|
118
|
-
|
119
|
-
validates_around :ids => { :type => :array, :presence => ->(ids) { !ids.empty? } }
|
120
|
-
|
750
|
+
|
751
|
+
validates_before :example_attribute => { :presence => ->(attribute) { attribute.any? } }
|
752
|
+
|
121
753
|
def call
|
122
|
-
# set attribute3 on the shared context to satisfy the `validates_after` validations
|
123
|
-
context.attribute3 = true
|
124
|
-
|
125
|
-
# set attribute4 on the shared context to satisfy the `validates_after` validations
|
126
|
-
context.attribute4 = "an example string value"
|
127
754
|
end
|
128
755
|
end
|
129
756
|
|
130
|
-
|
131
|
-
|
132
|
-
|
133
|
-
|
134
|
-
|
135
|
-
|
757
|
+
result = ActionTaskExample.execute(:example_attribute => ["element1", "element2", "element3"])
|
758
|
+
|
759
|
+
result # => #<ActionLogic::ActionContext example_attribute=["element1", "element2", "element3"], status=:success>
|
760
|
+
```
|
761
|
+
|
762
|
+
In the example above, we define a lambda that accepts as input the value of the attribute on the `context`. In this case, we are interested in verifying that
|
763
|
+
`example_attribute` is not an empty `Array` or an empty `Hash`. This passes our before validation because `ActionTaskExample` is invoked with an `example_attribute`
|
764
|
+
whose value is an array consisting of three elements.
|
765
|
+
|
766
|
+
However, if a custom presence validation fails, `ActionLogic` will raise an `ActionLogic::PresenceError` with a detailed description about the attribute failing
|
767
|
+
the custom presence validation:
|
768
|
+
|
769
|
+
```ruby
|
770
|
+
class ActionTaskExample
|
771
|
+
include ActionLogic::ActionTask
|
136
772
|
|
137
|
-
|
138
|
-
# attribute and value that will satisfy the `validates_around` validation:
|
139
|
-
params[:ids] = [1, 2, 3, 4]
|
773
|
+
validates_before :example_attribute => { :presence => ->(attribute) { attribute.any? } }
|
140
774
|
|
141
|
-
|
775
|
+
def call
|
776
|
+
end
|
777
|
+
end
|
778
|
+
|
779
|
+
ActionTaskExample.execute(:example_attribute => []) # ~> ["Attribute: example_attribute is missing value in context but custom presence validation was specified"] (ActionLogic::PresenceError)
|
142
780
|
```
|
143
781
|
|
144
|
-
|
782
|
+
In the above example, we have failed to pass the presence validation for `example_attribute` because the value of `example_attribute` is an empty array. When
|
783
|
+
the custom presence validation lambda is called, the lambda returns `false` and the `ActionLogic::PresenceError` is thrown, with an error message indicating
|
784
|
+
the attribute that failed the presence validation while also indicating that a custom presence validation was specified.
|
145
785
|
|
146
|
-
|
786
|
+
### Before Validations
|
147
787
|
|
148
|
-
|
149
|
-
|
150
|
-
*
|
151
|
-
* :integer (rather than FixNum)
|
152
|
-
* :array
|
153
|
-
* :hash
|
154
|
-
* :nil (rather than NilClass)
|
788
|
+
If you combine Rails ActionController's `before_filter` and ActiveModel's `validates` then you have approximately what `ActionLogic` defines as `validates_before`.
|
789
|
+
Before validations can be defined in any of the `ActionLogic` abstractions (`ActionTask`, `ActionUseCase` and `ActionCoordinator`). In each abstraction a `validates_before`
|
790
|
+
operation is performed *before* invoking the `call` method.
|
155
791
|
|
156
|
-
|
792
|
+
Before validations allow you to specify a required attribute and optionally its type and / or presence. The following example illustrates how to specify a before
|
793
|
+
validation on a single attribute:
|
157
794
|
|
158
795
|
```ruby
|
796
|
+
class ActionTaskExample
|
797
|
+
include ActionLogic::ActionTask
|
798
|
+
|
799
|
+
validates_before :example_attribute => { :type => :array, :presence => ->(attribute) { attribute.any? } }
|
159
800
|
|
160
|
-
|
801
|
+
def call
|
802
|
+
end
|
161
803
|
end
|
162
804
|
|
163
|
-
|
805
|
+
result = ActionTaskExample.execute(:example_attribute => [1, 2, 3])
|
806
|
+
|
807
|
+
result # => #<ActionLogic::ActionContext example_attribute=[1, 2, 3], status=:success>
|
808
|
+
```
|
809
|
+
|
810
|
+
The following example illustrates how to specify a before validation for multiple attributes:
|
811
|
+
|
812
|
+
```ruby
|
813
|
+
class ActionTaskExample
|
164
814
|
include ActionLogic::ActionTask
|
165
|
-
|
166
|
-
|
167
|
-
|
815
|
+
|
816
|
+
validates_before :example_attribute => { :type => :array, :presence => ->(attribute) { attribute.any? } },
|
817
|
+
:example_attribute2 => { :type => :integer }
|
818
|
+
|
168
819
|
def call
|
169
820
|
end
|
170
821
|
end
|
822
|
+
|
823
|
+
result = ActionTaskExample.execute(:example_attribute => [1, 2, 3], :example_attribute2 => 1)
|
824
|
+
|
825
|
+
result # => #<ActionLogic::ActionContext example_attribute=[1, 2, 3], example_attribute2=1, status=:success>
|
826
|
+
```
|
827
|
+
|
828
|
+
### After Validations
|
829
|
+
|
830
|
+
If you combine Rails ActionController's `after_filter` and ActiveModel's `validates` then you have approximately what `ActionLogic` defines as `validates_after`.
|
831
|
+
After validations can be defined in any of the `ActionLogic` abstractions (`ActionTask`, `ActionUseCase` and `ActionCoordinator`). In each abstraction a `validates_after`
|
832
|
+
operation is performed *after* invoking the `call` method.
|
833
|
+
|
834
|
+
After validations allow you to specify a required attribute and optionally its type and / or presence. The following example illustrates how to specify an after
|
835
|
+
validation on a single attribute:
|
836
|
+
|
837
|
+
```ruby
|
838
|
+
class ActionTaskExample
|
839
|
+
include ActionLogic::ActionTask
|
840
|
+
|
841
|
+
validates_after :example_attribute => { :type => :array, :presence => ->(attribute) { attribute.any? } }
|
842
|
+
|
843
|
+
def call
|
844
|
+
context.example_attribute = [1, 2, 3]
|
845
|
+
end
|
846
|
+
end
|
847
|
+
|
848
|
+
result = ActionTaskExample.execute
|
849
|
+
|
850
|
+
result # => #<ActionLogic::ActionContext example_attribute=[1, 2, 3], status=:success>
|
851
|
+
```
|
852
|
+
The following example illustrates how to specify an after validation for multiple attributes:
|
853
|
+
|
854
|
+
```ruby
|
855
|
+
class ActionTaskExample
|
856
|
+
include ActionLogic::ActionTask
|
857
|
+
|
858
|
+
validates_after :example_attribute => { :type => :array, :presence => ->(attribute) { attribute.any? } },
|
859
|
+
:example_attribute2 => { :type => :integer }
|
860
|
+
|
861
|
+
def call
|
862
|
+
context.example_attribute = [1, 2, 3]
|
863
|
+
context.example_attribute2 = 1
|
864
|
+
end
|
865
|
+
end
|
866
|
+
|
867
|
+
result = ActionTaskExample.execute
|
868
|
+
|
869
|
+
result # => #<ActionLogic::ActionContext example_attribute=[1, 2, 3], example_attribute2=1, status=:success>
|
870
|
+
```
|
871
|
+
|
872
|
+
### Around Validations
|
873
|
+
|
874
|
+
If you combine Rails ActionController's `around_filter` and ActiveModel's `validates` then you have approximately what `ActionLogic` defines as `validates_around`.
|
875
|
+
Around validations can be defined in any of the `ActionLogic` abstractions (`ActionTask`, `ActionUseCase` and `ActionCoordinator`). In each abstraction a `validates_around`
|
876
|
+
operation is performed *before* and *after* invoking the `call` method.
|
877
|
+
|
878
|
+
Around validations allow you to specify a required attribute and optionally its type and / or presence. The following example illustrates how to specify an around
|
879
|
+
validation on a single attribute:
|
880
|
+
|
881
|
+
```ruby
|
882
|
+
class ActionTaskExample
|
883
|
+
include ActionLogic::ActionTask
|
884
|
+
|
885
|
+
validates_around :example_attribute => { :type => :array, :presence => ->(attribute) { attribute.any? } }
|
886
|
+
|
887
|
+
def call
|
888
|
+
end
|
889
|
+
end
|
890
|
+
|
891
|
+
result = ActionTaskExample.execute(:example_attribute => [1, 2, 3])
|
892
|
+
|
893
|
+
result # => #<ActionLogic::ActionContext example_attribute=[1, 2, 3], status=:success>
|
894
|
+
```
|
895
|
+
The following example illustrates how to specify an around validation for multiple attributes:
|
896
|
+
|
897
|
+
```ruby
|
898
|
+
class ActionTaskExample
|
899
|
+
include ActionLogic::ActionTask
|
900
|
+
|
901
|
+
validates_around :example_attribute => { :type => :array, :presence => ->(attribute) { attribute.any? } },
|
902
|
+
:example_attribute2 => { :type => :integer }
|
903
|
+
|
904
|
+
def call
|
905
|
+
end
|
906
|
+
end
|
907
|
+
|
908
|
+
result = ActionTaskExample.execute(:example_attribute => [1, 2, 3], :example_attribute2 => 1)
|
909
|
+
|
910
|
+
result # => #<ActionLogic::ActionContext example_attribute=[1, 2, 3], example_attribute2=1, status=:success>
|
171
911
|
```
|