workflow 0.3.0 → 0.4.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/.gitignore +7 -0
- data/README.markdown +433 -0
- data/Rakefile +26 -26
- data/VERSION +1 -0
- data/lib/workflow.rb +127 -31
- data/test/couchtiny_example.rb +46 -0
- data/test/main_test.rb +420 -0
- data/test/readme_example.rb +37 -0
- data/test/without_active_record_test.rb +54 -0
- data/workflow.rb +1 -0
- metadata +32 -14
- data/README.rdoc +0 -452
data/.gitignore
ADDED
data/README.markdown
ADDED
@@ -0,0 +1,433 @@
|
|
1
|
+
What is workflow?
|
2
|
+
-----------------
|
3
|
+
|
4
|
+
Workflow is a finite-state-machine-inspired API for modeling and
|
5
|
+
interacting with what we tend to refer to as 'workflow'.
|
6
|
+
|
7
|
+
A lot of business modeling tends to involve workflow-like concepts, and
|
8
|
+
the aim of this library is to make the expression of these concepts as
|
9
|
+
clear as possible, using similar terminology as found in state machine
|
10
|
+
theory.
|
11
|
+
|
12
|
+
So, a workflow has a state. It can only be in one state at a time. When
|
13
|
+
a workflow changes state, we call that a transition. Transitions occur
|
14
|
+
on an event, so events cause transitions to occur. Additionally, when an
|
15
|
+
event fires, other arbitrary code can be executed, we call those actions.
|
16
|
+
So any given state has a bunch of events, any event in a state causes a
|
17
|
+
transition to another state and potentially causes code to be executed
|
18
|
+
(an action). We can hook into states when they are entered, and exited
|
19
|
+
from, and we can cause transitions to fail (guards), and we can hook in
|
20
|
+
to every transition that occurs ever for whatever reason we can come up
|
21
|
+
with.
|
22
|
+
|
23
|
+
Now, all that's a mouthful, but we'll demonstrate the API bit by bit
|
24
|
+
with a real-ish world example.
|
25
|
+
|
26
|
+
Let's say we're modeling article submission from journalists. An article
|
27
|
+
is written, then submitted. When it's submitted, it's awaiting review.
|
28
|
+
Someone reviews the article, and then either accepts or rejects it.
|
29
|
+
Here is the expression of this workflow using the API:
|
30
|
+
|
31
|
+
class Article
|
32
|
+
include Workflow
|
33
|
+
workflow do
|
34
|
+
state :new do
|
35
|
+
event :submit, :transitions_to => :awaiting_review
|
36
|
+
end
|
37
|
+
state :awaiting_review do
|
38
|
+
event :review, :transitions_to => :being_reviewed
|
39
|
+
end
|
40
|
+
state :being_reviewed do
|
41
|
+
event :accept, :transitions_to => :accepted
|
42
|
+
event :reject, :transitions_to => :rejected
|
43
|
+
end
|
44
|
+
state :accepted
|
45
|
+
state :rejected
|
46
|
+
end
|
47
|
+
end
|
48
|
+
|
49
|
+
Nice, isn't it!
|
50
|
+
|
51
|
+
Let's create an article instance and check in which state it is:
|
52
|
+
|
53
|
+
article = Article.new
|
54
|
+
article.accepted? # => false
|
55
|
+
article.new? # => true
|
56
|
+
|
57
|
+
You can also access the whole `current_state` object including the list
|
58
|
+
of possible events and other meta information:
|
59
|
+
|
60
|
+
article.current_state
|
61
|
+
=> #<Workflow::State:0x7f1e3d6731f0 @events={
|
62
|
+
:submit=>#<Workflow::Event:0x7f1e3d6730d8 @action=nil,
|
63
|
+
@transitions_to=:awaiting_review, @name=:submit, @meta={}>},
|
64
|
+
name:new, meta{}
|
65
|
+
|
66
|
+
Now we can call the submit event, which transitions to the
|
67
|
+
<tt>:awaiting_review</tt> state:
|
68
|
+
|
69
|
+
article.submit!
|
70
|
+
article.awaiting_review? # => true
|
71
|
+
|
72
|
+
Events are actually instance methods on a workflow, and depending on the
|
73
|
+
state you're in, you'll have a different set of events used to
|
74
|
+
transition to other states.
|
75
|
+
|
76
|
+
|
77
|
+
Installation
|
78
|
+
------------
|
79
|
+
|
80
|
+
gem install workflow
|
81
|
+
|
82
|
+
Alternatively you can just download the lib/workflow.rb and put it in
|
83
|
+
the lib folder of your Rails or Ruby application.
|
84
|
+
|
85
|
+
|
86
|
+
Examples
|
87
|
+
--------
|
88
|
+
|
89
|
+
After installation or downloading of the library you can easily try out
|
90
|
+
all the example code from this README in irb.
|
91
|
+
|
92
|
+
$ irb
|
93
|
+
require 'rubygems'
|
94
|
+
require 'workflow'
|
95
|
+
|
96
|
+
Now just copy and paste the source code from the beginning of this README
|
97
|
+
file snippet by snippet and observe the output.
|
98
|
+
|
99
|
+
|
100
|
+
Transition event handler
|
101
|
+
------------------------
|
102
|
+
|
103
|
+
The best way is to use convention over configuration and to define a
|
104
|
+
method with the same name as the event. Then it is automatically invoked
|
105
|
+
when event is raised. For the Article workflow defined earlier it would
|
106
|
+
be:
|
107
|
+
|
108
|
+
class Article
|
109
|
+
def reject
|
110
|
+
puts 'sending email to the author explaining the reason...'
|
111
|
+
end
|
112
|
+
end
|
113
|
+
|
114
|
+
`article.review!; article.reject!` will cause a state transition, persist the new state
|
115
|
+
(if integrated with ActiveRecord) and invoke this user defined reject
|
116
|
+
method.
|
117
|
+
|
118
|
+
You can also define event handler accepting/requiring additional
|
119
|
+
arguments:
|
120
|
+
|
121
|
+
class Article
|
122
|
+
def review(reviewer = '')
|
123
|
+
puts "[#{reviewer}] is now reviewing the article"
|
124
|
+
end
|
125
|
+
end
|
126
|
+
|
127
|
+
article2 = Article.new
|
128
|
+
article2.submit!
|
129
|
+
article2.review!('Homer Simpson') # => [Homer Simpson] is now reviewing the article
|
130
|
+
|
131
|
+
|
132
|
+
### The old, deprecated way
|
133
|
+
|
134
|
+
The old way, using a block is still supported but deprecated:
|
135
|
+
|
136
|
+
event :review, :transitions_to => :being_reviewed do |reviewer|
|
137
|
+
# store the reviewer
|
138
|
+
end
|
139
|
+
|
140
|
+
We've noticed, that mixing the list of events and states with the blocks
|
141
|
+
invoked for particular transitions leads to a bumpy and poorly readable code
|
142
|
+
due to a deep nesting. We tried (and dismissed) lambdas for this. Eventually
|
143
|
+
we decided to invoke an optional user defined callback method with the same
|
144
|
+
name as the event (convention over configuration) as explained before.
|
145
|
+
|
146
|
+
|
147
|
+
Integration with ActiveRecord
|
148
|
+
-----------------------------
|
149
|
+
|
150
|
+
Workflow library can handle the state persistence fully automatically. You
|
151
|
+
only need to define a string field on the table called `workflow_state`
|
152
|
+
and include the workflow mixin in your model class as usual:
|
153
|
+
|
154
|
+
class Order < ActiveRecord::Base
|
155
|
+
include Workflow
|
156
|
+
workflow do
|
157
|
+
# list states and transitions here
|
158
|
+
end
|
159
|
+
end
|
160
|
+
|
161
|
+
On a database record loading all the state check methods e.g.
|
162
|
+
`article.state`, `article.awaiting_review?` are immediately available.
|
163
|
+
For new records or if the workflow_state field is not set the state
|
164
|
+
defaults to the first state declared in the workflow specification. In
|
165
|
+
our example it is `:new`, so `Article.new.new?` returns true and
|
166
|
+
`Article.new.approved?` returns false.
|
167
|
+
|
168
|
+
At the end of a successful state transition like `article.approve!` the
|
169
|
+
new state is immediately saved in the database.
|
170
|
+
|
171
|
+
You can change this behaviour by overriding `persist_workflow_state`
|
172
|
+
method.
|
173
|
+
|
174
|
+
|
175
|
+
### Custom workflow database column
|
176
|
+
|
177
|
+
[meuble](http://imeuble.info/) contributed a solution for using
|
178
|
+
custom persistence column easily, e.g. for a legacy database schema:
|
179
|
+
|
180
|
+
class LegacyOrder < ActiveRecord::Base
|
181
|
+
include Workflow
|
182
|
+
|
183
|
+
workflow_column :foo_bar # use this legacy database column for
|
184
|
+
# persistence
|
185
|
+
end
|
186
|
+
|
187
|
+
|
188
|
+
|
189
|
+
### Single table inheritance
|
190
|
+
|
191
|
+
Single table inheritance is also supported. Descendant classes can either
|
192
|
+
inherit the workflow definition from the parent or override with its own
|
193
|
+
definition.
|
194
|
+
|
195
|
+
Custom workflow state persistence
|
196
|
+
---------------------------------
|
197
|
+
|
198
|
+
If you do not use a relational database and ActiveRecord, you can still
|
199
|
+
integrate the workflow very easily. To implement persistence you just
|
200
|
+
need to override `load_workflow_state` and
|
201
|
+
`persist_workflow_state(new_value)` methods. Next section contains an example for
|
202
|
+
using CouchDB, a document oriented database.
|
203
|
+
|
204
|
+
[Tim Lossen](http://tim.lossen.de/) implemented support
|
205
|
+
for [remodel](http://github.com/tlossen/remodel) / [redis](http://github.com/antirez/redis)
|
206
|
+
key-value store.
|
207
|
+
|
208
|
+
Integration with CouchDB
|
209
|
+
------------------------
|
210
|
+
|
211
|
+
We are using the compact [couchtiny library](http://github.com/geekq/couchtiny)
|
212
|
+
here. But the implementation would look similar for the popular
|
213
|
+
couchrest library.
|
214
|
+
|
215
|
+
require 'couchtiny'
|
216
|
+
require 'couchtiny/document'
|
217
|
+
require 'workflow'
|
218
|
+
|
219
|
+
class User < CouchTiny::Document
|
220
|
+
include Workflow
|
221
|
+
workflow do
|
222
|
+
state :submitted do
|
223
|
+
event :activate_via_link, :transitions_to => :proved_email
|
224
|
+
end
|
225
|
+
state :proved_email
|
226
|
+
end
|
227
|
+
|
228
|
+
def load_workflow_state
|
229
|
+
self[:workflow_state]
|
230
|
+
end
|
231
|
+
|
232
|
+
def persist_workflow_state(new_value)
|
233
|
+
self[:workflow_state] = new_value
|
234
|
+
save!
|
235
|
+
end
|
236
|
+
end
|
237
|
+
|
238
|
+
Please also have a look at
|
239
|
+
[the full source code](http://github.com/geekq/workflow/blob/master/test/couchtiny_example.rb).
|
240
|
+
|
241
|
+
Accessing your workflow specification
|
242
|
+
-------------------------------------
|
243
|
+
|
244
|
+
You can easily reflect on workflow specification programmatically - for
|
245
|
+
the whole class or for the current object. Examples:
|
246
|
+
|
247
|
+
article2.current_state.events # lists possible events from here
|
248
|
+
article2.current_state.events[:reject].transitions_to # => :rejected
|
249
|
+
|
250
|
+
Article.workflow_spec.states.keys
|
251
|
+
#=> [:rejected, :awaiting_review, :being_reviewed, :accepted, :new]
|
252
|
+
|
253
|
+
Article.workflow_spec.state_names
|
254
|
+
#=> [:rejected, :awaiting_review, :being_reviewed, :accepted, :new]
|
255
|
+
|
256
|
+
# list all events for all states
|
257
|
+
Article.workflow_spec.states.values.collect &:events
|
258
|
+
|
259
|
+
|
260
|
+
You can also store and later retrieve additional meta data for every
|
261
|
+
state and every event:
|
262
|
+
|
263
|
+
class MyProcess
|
264
|
+
include Workflow
|
265
|
+
workflow do
|
266
|
+
state :main, :meta => {:importance => 8}
|
267
|
+
state :supplemental, :meta => {:importance => 1}
|
268
|
+
end
|
269
|
+
end
|
270
|
+
puts MyProcess.workflow_spec.states[:supplemental].meta[:importance] # => 1
|
271
|
+
|
272
|
+
The workflow library itself uses this feature to tweak the graphical
|
273
|
+
representation of the workflow. See below.
|
274
|
+
|
275
|
+
|
276
|
+
Advanced transition hooks
|
277
|
+
-------------------------
|
278
|
+
|
279
|
+
### on_entry/on_exit
|
280
|
+
|
281
|
+
We already had a look at the declaring callbacks for particular workflow
|
282
|
+
events. If you would like to react to all transitions to/from the same state
|
283
|
+
in the same way you can use the on_entry/on_exit hooks. You can either define it
|
284
|
+
with a block inside the workflow definition or through naming
|
285
|
+
convention, e.g. for the state :pending just define the method
|
286
|
+
`on_pending_exit(new_state, event, *args)` somewhere in your class.
|
287
|
+
|
288
|
+
### on_transition
|
289
|
+
|
290
|
+
If you want to be informed about everything happening everywhere, e.g. for
|
291
|
+
logging then you can use the universal `on_transition` hook:
|
292
|
+
|
293
|
+
workflow do
|
294
|
+
state :one do
|
295
|
+
event :increment, :transitions_to => :two
|
296
|
+
end
|
297
|
+
state :two
|
298
|
+
on_transition do |from, to, triggering_event, *event_args|
|
299
|
+
Log.info "#{from} -> #{to}"
|
300
|
+
end
|
301
|
+
end
|
302
|
+
|
303
|
+
|
304
|
+
### Guards
|
305
|
+
|
306
|
+
If you want to halt the transition conditionally, you can just raise an
|
307
|
+
exception. There is a helper called `halt!`, which raises the
|
308
|
+
Workflow::TransitionHalted exception. You can provide an additional
|
309
|
+
`halted_because` parameter.
|
310
|
+
|
311
|
+
def reject(reason)
|
312
|
+
halt! 'We do not reject articles unless the reason is important' \
|
313
|
+
unless reason =~ /important/i
|
314
|
+
end
|
315
|
+
|
316
|
+
The traditional `halt` (without the exclamation mark) is still supported
|
317
|
+
too. This just prevents the state change without raising an
|
318
|
+
exception.
|
319
|
+
|
320
|
+
### Hook order
|
321
|
+
|
322
|
+
The whole event sequence is as follows:
|
323
|
+
|
324
|
+
* event specific action
|
325
|
+
* on_transition (if action did not halt)
|
326
|
+
* on_exit
|
327
|
+
* PERSIST WORKFLOW STATE, i.e. transition
|
328
|
+
* on_entry
|
329
|
+
|
330
|
+
|
331
|
+
Documenting with diagrams
|
332
|
+
-------------------------
|
333
|
+
|
334
|
+
You can generate a graphical representation of your workflow for
|
335
|
+
documentation purposes. S. Workflow::create_workflow_diagram.
|
336
|
+
|
337
|
+
|
338
|
+
Earlier versions
|
339
|
+
----------------
|
340
|
+
|
341
|
+
The `workflow` library was originally written by Ryan Allen.
|
342
|
+
|
343
|
+
The version 0.3 was almost completely (including ActiveRecord
|
344
|
+
integration, API for accessing workflow specification,
|
345
|
+
method_missing free implementation) rewritten by Vladimir Dobriakov
|
346
|
+
keeping the original workflow DSL spirit.
|
347
|
+
|
348
|
+
|
349
|
+
Migration from the original Ryan's library
|
350
|
+
------------------------------------------
|
351
|
+
|
352
|
+
Credit: Michael (rockrep)
|
353
|
+
|
354
|
+
Accessing workflow specification
|
355
|
+
|
356
|
+
my_instance.workflow # old
|
357
|
+
MyClass.workflow_spec # new
|
358
|
+
|
359
|
+
Accessing states, events, meta, e.g.
|
360
|
+
|
361
|
+
my_instance.workflow.states(:some_state).events(:some_event).meta[:some_meta_tag] # old
|
362
|
+
MyClass.workflow_spec.states[:some_state].events[:some_event].meta[:some_meta_tag] # new
|
363
|
+
|
364
|
+
Causing state transitions
|
365
|
+
|
366
|
+
my_instance.workflow.my_event # old
|
367
|
+
my_instance.my_event! # new
|
368
|
+
|
369
|
+
when using both a block and a callback method for an event, the block executes prior to the callback
|
370
|
+
|
371
|
+
|
372
|
+
Changelog
|
373
|
+
---------
|
374
|
+
|
375
|
+
### New in the version 0.4.0
|
376
|
+
|
377
|
+
* completely rewritten the documentation to match my branch
|
378
|
+
* switch to [jeweler][] for building gems
|
379
|
+
* every described feature is backed up by an automated test
|
380
|
+
|
381
|
+
### New in the version 0.3.0
|
382
|
+
|
383
|
+
Intermixing of transition graph definition (states, transitions)
|
384
|
+
on the one side and implementation of the actions on the other side
|
385
|
+
for a bigger state machine can introduce clutter.
|
386
|
+
|
387
|
+
To reduce this clutter it is now possible to use state entry- and
|
388
|
+
exit- hooks defined through a naming convention. For example, if there
|
389
|
+
is a state :pending, then instead of using a
|
390
|
+
block:
|
391
|
+
|
392
|
+
state :pending do
|
393
|
+
on_entry do
|
394
|
+
# your implementation here
|
395
|
+
end
|
396
|
+
end
|
397
|
+
|
398
|
+
you can hook in by defining method
|
399
|
+
|
400
|
+
def on_pending_exit(new_state, event, *args)
|
401
|
+
# your implementation here
|
402
|
+
end
|
403
|
+
|
404
|
+
anywhere in your class. You can also use a simpler function signature
|
405
|
+
like `def on_pending_exit(*args)` if your are not interested in
|
406
|
+
arguments. Please note: `def on_pending_exit()` with an empty list
|
407
|
+
would not work.
|
408
|
+
|
409
|
+
If both a function with a name according to naming convention and the
|
410
|
+
on_entry/on_exit block are given, then only on_entry/on_exit block is used.
|
411
|
+
|
412
|
+
|
413
|
+
Support
|
414
|
+
-------
|
415
|
+
|
416
|
+
### Reporting bugs
|
417
|
+
|
418
|
+
http://github.com/geekq/workflow/issues
|
419
|
+
|
420
|
+
|
421
|
+
About
|
422
|
+
-----
|
423
|
+
|
424
|
+
Author: Vladimir Dobriakov, http://www.innoq.com/blog/vd, http://blog.geekq.net/
|
425
|
+
|
426
|
+
Copyright (c) 2008-2009 Vodafone
|
427
|
+
|
428
|
+
Copyright (c) 2007-2008 Ryan Allen, FlashDen Pty Ltd
|
429
|
+
|
430
|
+
Based on the work of Ryan Allen and Scott Barron
|
431
|
+
|
432
|
+
Licensed under MIT license, see the MIT-LICENSE file.
|
433
|
+
|
data/Rakefile
CHANGED
@@ -11,34 +11,34 @@ Rake::TestTask.new do |t|
|
|
11
11
|
t.pattern = 'test/*_test.rb'
|
12
12
|
end
|
13
13
|
|
14
|
-
PKG_VERSION = "0.3.0"
|
15
|
-
PKG_FILES = FileList[
|
16
|
-
'MIT-LICENSE',
|
17
|
-
'README.rdoc',
|
18
|
-
'Rakefile',
|
19
|
-
'lib/**/*.rb',
|
20
|
-
'test/**/test_*.rb'
|
21
|
-
]
|
22
|
-
|
23
|
-
spec = Gem::Specification.new do |s|
|
24
|
-
s.name = "workflow"
|
25
|
-
s.version = PKG_VERSION
|
26
|
-
s.author = "Vladimir Dobriakov"
|
27
|
-
s.email = "vladimir@geekq.net"
|
28
|
-
s.homepage = "http://blog.geekQ.net/"
|
29
|
-
s.platform = Gem::Platform::RUBY
|
30
|
-
s.summary = "A replacement for acts_as_state_machine."
|
31
|
-
s.files = PKG_FILES.to_a
|
32
|
-
s.require_path = "lib"
|
33
|
-
end
|
34
|
-
|
35
14
|
Rake::RDocTask.new do |rdoc|
|
36
|
-
rdoc.
|
37
|
-
rdoc.rdoc_files.include("README.rdoc", "lib/**/*.rb")
|
15
|
+
rdoc.rdoc_files.include("lib/**/*.rb")
|
38
16
|
rdoc.options << "-S"
|
39
17
|
end
|
40
18
|
|
41
|
-
|
42
|
-
|
43
|
-
|
19
|
+
begin
|
20
|
+
require 'jeweler'
|
21
|
+
Jeweler::Tasks.new do |gemspec|
|
22
|
+
gemspec.name = "workflow"
|
23
|
+
gemspec.rubyforge_project = 'workflow'
|
24
|
+
gemspec.email = "vladimir@geekq.net"
|
25
|
+
gemspec.homepage = "http://blog.geekQ.net/"
|
26
|
+
gemspec.authors = ["Vladimir Dobriakov"]
|
27
|
+
gemspec.summary = "A replacement for acts_as_state_machine."
|
28
|
+
gemspec.description = <<-EOS
|
29
|
+
Workflow is a finite-state-machine-inspired API for modeling and interacting
|
30
|
+
with what we tend to refer to as 'workflow'.
|
31
|
+
|
32
|
+
* nice DSL to describe your states, events and transitions
|
33
|
+
* robust integration with ActiveRecord and non relational data stores
|
34
|
+
* various hooks for single transitions, entering state etc.
|
35
|
+
* convenient access to the workflow specification: list states, possible events
|
36
|
+
for particular state
|
37
|
+
EOS
|
38
|
+
|
39
|
+
Jeweler::GemcutterTasks.new
|
40
|
+
end
|
41
|
+
rescue LoadError
|
42
|
+
puts "Jeweler not available. Install it with: sudo gem install technicalpickles-jeweler -s http://gems.github.com"
|
44
43
|
end
|
44
|
+
|