redis-objects-legacy 1.6.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +12 -0
- data/.travis.yml +16 -0
- data/ATOMICITY.rdoc +154 -0
- data/CHANGELOG.rdoc +362 -0
- data/Gemfile +4 -0
- data/LICENSE +202 -0
- data/README.md +600 -0
- data/Rakefile +14 -0
- data/lib/redis/base_object.rb +62 -0
- data/lib/redis/counter.rb +145 -0
- data/lib/redis/enumerable_object.rb +28 -0
- data/lib/redis/hash_key.rb +163 -0
- data/lib/redis/helpers/core_commands.rb +89 -0
- data/lib/redis/list.rb +160 -0
- data/lib/redis/lock.rb +89 -0
- data/lib/redis/objects/connection_pool_proxy.rb +31 -0
- data/lib/redis/objects/counters.rb +155 -0
- data/lib/redis/objects/hashes.rb +60 -0
- data/lib/redis/objects/lists.rb +58 -0
- data/lib/redis/objects/locks.rb +73 -0
- data/lib/redis/objects/sets.rb +58 -0
- data/lib/redis/objects/sorted_sets.rb +49 -0
- data/lib/redis/objects/values.rb +64 -0
- data/lib/redis/objects/version.rb +5 -0
- data/lib/redis/objects.rb +199 -0
- data/lib/redis/set.rb +182 -0
- data/lib/redis/sorted_set.rb +325 -0
- data/lib/redis/value.rb +65 -0
- data/lib/redis-objects-legacy.rb +1 -0
- data/spec/redis_autoload_objects_spec.rb +46 -0
- data/spec/redis_namespace_compat_spec.rb +24 -0
- data/spec/redis_objects_active_record_spec.rb +162 -0
- data/spec/redis_objects_conn_spec.rb +276 -0
- data/spec/redis_objects_custom_serializer.rb +198 -0
- data/spec/redis_objects_instance_spec.rb +1666 -0
- data/spec/redis_objects_model_spec.rb +1097 -0
- data/spec/spec_helper.rb +92 -0
- metadata +214 -0
data/README.md
ADDED
@@ -0,0 +1,600 @@
|
|
1
|
+
Redis::Objects - Map Redis types directly to Ruby objects
|
2
|
+
=========================================================
|
3
|
+
|
4
|
+
[![Build Status](https://app.travis-ci.com/nateware/redis-objects.svg?branch=master)](https://travis-ci.com/nateware/redis-objects)
|
5
|
+
[![Code Coverage](https://codecov.io/gh/nateware/redis-objects/branch/master/graph/badge.svg)](https://codecov.io/gh/nateware/redis-objects)
|
6
|
+
[![Donate](https://www.paypalobjects.com/en_US/i/btn/btn_donate_SM.gif)](https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=MJF7JU5M7F8VL)
|
7
|
+
|
8
|
+
This is **not** an ORM. People that are wrapping ORM’s around Redis are missing the point.
|
9
|
+
|
10
|
+
The killer feature of Redis is that it allows you to perform _atomic_ operations
|
11
|
+
on _individual_ data structures, like counters, lists, and sets. The **atomic** part is HUGE.
|
12
|
+
Using an ORM wrapper that retrieves a "record", updates values, then sends those values back,
|
13
|
+
_removes_ the atomicity, and thus the major advantage of Redis. Just use MySQL, k?
|
14
|
+
|
15
|
+
This gem provides a Rubyish interface to Redis, by mapping [Redis data types](http://redis.io/commands)
|
16
|
+
to Ruby objects, via a thin layer over the `redis` gem. It offers several advantages
|
17
|
+
over the lower-level redis-rb API:
|
18
|
+
|
19
|
+
1. Easy to integrate directly with existing ORMs - ActiveRecord, DataMapper, etc. Add counters to your model!
|
20
|
+
2. Complex data structures are automatically Marshaled (if you set :marshal => true)
|
21
|
+
3. Integers are returned as integers, rather than '17'
|
22
|
+
4. Higher-level types are provided, such as Locks, that wrap multiple calls
|
23
|
+
|
24
|
+
This gem originally arose out of a need for high-concurrency atomic operations;
|
25
|
+
for a fun rant on the topic, see [An Atomic Rant](http://nateware.com/2010/02/18/an-atomic-rant),
|
26
|
+
or scroll down to [Atomic Counters and Locks](#atomicity) in this README.
|
27
|
+
|
28
|
+
There are two ways to use Redis::Objects, either as an include in a model class (to
|
29
|
+
tightly integrate with ORMs or other classes), or standalone by using classes such
|
30
|
+
as `Redis::List` and `Redis::SortedSet`.
|
31
|
+
|
32
|
+
Installation and Setup
|
33
|
+
----------------------
|
34
|
+
Add it to your Gemfile as:
|
35
|
+
|
36
|
+
~~~ruby
|
37
|
+
gem 'redis-objects'
|
38
|
+
~~~
|
39
|
+
|
40
|
+
Redis::Objects needs a handle created by `Redis.new` or a [ConnectionPool](https://github.com/mperham/connection_pool):
|
41
|
+
|
42
|
+
The recommended approach is to use a `ConnectionPool` since this guarantees that most timeouts in the `redis` client
|
43
|
+
do not pollute your existing connection. However, you need to make sure that both `:timeout` and `:size` are set appropriately
|
44
|
+
in a multithreaded environment.
|
45
|
+
~~~ruby
|
46
|
+
require 'connection_pool'
|
47
|
+
Redis::Objects.redis = ConnectionPool.new(size: 5, timeout: 5) { Redis.new(:host => '127.0.0.1', :port => 6379) }
|
48
|
+
~~~
|
49
|
+
|
50
|
+
Redis::Objects can also default to `Redis.current` if `Redis::Objects.redis` is not set.
|
51
|
+
~~~ruby
|
52
|
+
Redis.current = Redis.new(:host => '127.0.0.1', :port => 6379)
|
53
|
+
~~~
|
54
|
+
|
55
|
+
(If you're on Rails, `config/initializers/redis.rb` is a good place for this.)
|
56
|
+
Remember you can use Redis::Objects in any Ruby code. There are **no** dependencies
|
57
|
+
on Rails. Standalone, Sinatra, Resque - no problem.
|
58
|
+
|
59
|
+
Alternatively, you can set the `redis` handle directly:
|
60
|
+
|
61
|
+
~~~ruby
|
62
|
+
Redis::Objects.redis = Redis.new(...)
|
63
|
+
~~~
|
64
|
+
|
65
|
+
Finally, you can even set different handles for different classes:
|
66
|
+
|
67
|
+
~~~ruby
|
68
|
+
class User
|
69
|
+
include Redis::Objects
|
70
|
+
end
|
71
|
+
class Post
|
72
|
+
include Redis::Objects
|
73
|
+
end
|
74
|
+
|
75
|
+
# you can also use a ConnectionPool here as well
|
76
|
+
User.redis = Redis.new(:host => '1.2.3.4')
|
77
|
+
Post.redis = Redis.new(:host => '5.6.7.8')
|
78
|
+
~~~
|
79
|
+
|
80
|
+
As of `0.7.0`, `redis-objects` now autoloads the appropriate `Redis::Whatever`
|
81
|
+
classes on demand. Previous strategies of individually requiring `redis/list`
|
82
|
+
or `redis/set` are no longer required.
|
83
|
+
|
84
|
+
Option 1: Model Class Include
|
85
|
+
=============================
|
86
|
+
Including Redis::Objects in a model class makes it trivial to integrate Redis types
|
87
|
+
with an existing ActiveRecord, DataMapper, Mongoid, or similar class. **Redis::Objects
|
88
|
+
will work with _any_ class that provides an `id` method that returns a unique value.**
|
89
|
+
Redis::Objects automatically creates keys that are unique to each object, in the format:
|
90
|
+
|
91
|
+
model_name:id:field_name
|
92
|
+
|
93
|
+
For illustration purposes, consider this stub class:
|
94
|
+
|
95
|
+
~~~ruby
|
96
|
+
class User
|
97
|
+
include Redis::Objects
|
98
|
+
counter :my_posts
|
99
|
+
def id
|
100
|
+
1
|
101
|
+
end
|
102
|
+
end
|
103
|
+
|
104
|
+
user = User.new
|
105
|
+
user.id # 1
|
106
|
+
user.my_posts.increment
|
107
|
+
user.my_posts.increment
|
108
|
+
user.my_posts.increment
|
109
|
+
puts user.my_posts.value # 3
|
110
|
+
user.my_posts.reset
|
111
|
+
puts user.my_posts.value # 0
|
112
|
+
user.my_posts.reset 5
|
113
|
+
puts user.my_posts.value # 5
|
114
|
+
~~~
|
115
|
+
|
116
|
+
Here's an example that integrates several data types with an ActiveRecord model:
|
117
|
+
|
118
|
+
~~~ruby
|
119
|
+
class Team < ActiveRecord::Base
|
120
|
+
include Redis::Objects
|
121
|
+
|
122
|
+
lock :trade_players, :expiration => 15 # sec
|
123
|
+
value :at_bat
|
124
|
+
counter :hits
|
125
|
+
counter :runs
|
126
|
+
counter :outs
|
127
|
+
counter :inning, :start => 1
|
128
|
+
list :on_base
|
129
|
+
list :coaches, :marshal => true
|
130
|
+
set :outfielders
|
131
|
+
hash_key :pitchers_faced # "hash" is taken by Ruby
|
132
|
+
sorted_set :rank, :global => true
|
133
|
+
end
|
134
|
+
~~~
|
135
|
+
|
136
|
+
Familiar Ruby array operations Just Work (TM):
|
137
|
+
|
138
|
+
~~~ruby
|
139
|
+
@team = Team.find_by_name('New York Yankees')
|
140
|
+
@team.on_base << 'player1'
|
141
|
+
@team.on_base << 'player2'
|
142
|
+
@team.on_base << 'player3'
|
143
|
+
@team.on_base # ['player1', 'player2', 'player3']
|
144
|
+
@team.on_base.pop
|
145
|
+
@team.on_base.shift
|
146
|
+
@team.on_base.length # 1
|
147
|
+
@team.on_base.delete('player2')
|
148
|
+
@team.on_base = ['player1', 'player2'] # ['player1', 'player2']
|
149
|
+
~~~
|
150
|
+
|
151
|
+
Sets work too:
|
152
|
+
|
153
|
+
~~~ruby
|
154
|
+
@team.outfielders << 'outfielder1'
|
155
|
+
@team.outfielders << 'outfielder2'
|
156
|
+
@team.outfielders << 'outfielder1' # dup ignored
|
157
|
+
@team.outfielders # ['outfielder1', 'outfielder2']
|
158
|
+
@team.outfielders.each do |player|
|
159
|
+
puts player
|
160
|
+
end
|
161
|
+
player = @team.outfielders.detect{|of| of == 'outfielder2'}
|
162
|
+
@team.outfielders = ['outfielder1', 'outfielder3'] # ['outfielder1', 'outfielder3']
|
163
|
+
~~~
|
164
|
+
|
165
|
+
Hashes work too:
|
166
|
+
|
167
|
+
~~~ruby
|
168
|
+
@team.pitchers_faced['player1'] = 'pitcher2'
|
169
|
+
@team.pitchers_faced['player2'] = 'pitcher1'
|
170
|
+
@team.pitchers_faced = { 'player1' => 'pitcher2', 'player2' => 'pitcher1' }
|
171
|
+
~~~
|
172
|
+
|
173
|
+
And you can do unions and intersections between objects (kinda cool):
|
174
|
+
|
175
|
+
~~~ruby
|
176
|
+
@team1.outfielders | @team2.outfielders # outfielders on both teams
|
177
|
+
@team1.outfielders & @team2.outfielders # in baseball, should be empty :-)
|
178
|
+
~~~
|
179
|
+
|
180
|
+
Counters can be atomically incremented/decremented (but not assigned):
|
181
|
+
|
182
|
+
~~~ruby
|
183
|
+
@team.hits.increment # or incr
|
184
|
+
@team.hits.decrement # or decr
|
185
|
+
@team.hits.incr(3) # add 3
|
186
|
+
@team.runs = 4 # exception
|
187
|
+
~~~
|
188
|
+
|
189
|
+
Defining a different method as the `id` field is easy
|
190
|
+
|
191
|
+
~~~ruby
|
192
|
+
class User
|
193
|
+
include Redis::Objects
|
194
|
+
redis_id_field :uid
|
195
|
+
counter :my_posts
|
196
|
+
end
|
197
|
+
|
198
|
+
user.uid # 195137a1bdea4473
|
199
|
+
user.my_posts.increment # 1
|
200
|
+
~~~
|
201
|
+
|
202
|
+
Finally, for free, you get a `redis` method that points directly to a Redis connection:
|
203
|
+
|
204
|
+
~~~ruby
|
205
|
+
Team.redis.get('somekey')
|
206
|
+
@team = Team.new
|
207
|
+
@team.redis.get('somekey')
|
208
|
+
@team.redis.smembers('someset')
|
209
|
+
~~~
|
210
|
+
|
211
|
+
You can use the `redis` handle to directly call any [Redis API command](http://redis.io/commands).
|
212
|
+
|
213
|
+
Option 2: Standalone Usage
|
214
|
+
===========================
|
215
|
+
There is a Ruby class that maps to each Redis type, with methods for each
|
216
|
+
[Redis API command](http://redis.io/commands).
|
217
|
+
Note that calling `new` does not imply it's actually a "new" value - it just
|
218
|
+
creates a mapping between that Ruby object and the corresponding Redis data
|
219
|
+
structure, which may already exist on the `redis-server`.
|
220
|
+
|
221
|
+
Counters
|
222
|
+
--------
|
223
|
+
The `counter_name` is the key stored in Redis.
|
224
|
+
|
225
|
+
~~~ruby
|
226
|
+
@counter = Redis::Counter.new('counter_name')
|
227
|
+
@counter.increment # or incr
|
228
|
+
@counter.decrement # or decr
|
229
|
+
@counter.increment(3)
|
230
|
+
puts @counter.value
|
231
|
+
~~~
|
232
|
+
|
233
|
+
This gem provides a clean way to do atomic blocks as well:
|
234
|
+
|
235
|
+
~~~ruby
|
236
|
+
@counter.increment do |val|
|
237
|
+
raise "Full" if val > MAX_VAL # rewind counter
|
238
|
+
end
|
239
|
+
~~~
|
240
|
+
|
241
|
+
See the section on [Atomic Counters and Locks](#atomicity) for cool uses of atomic counter blocks.
|
242
|
+
|
243
|
+
Locks
|
244
|
+
-----
|
245
|
+
A convenience class that wraps the pattern of [using setnx to perform locking](http://redis.io/commands/setnx).
|
246
|
+
|
247
|
+
~~~ruby
|
248
|
+
@lock = Redis::Lock.new('serialize_stuff', :expiration => 15, :timeout => 0.1)
|
249
|
+
@lock.lock do
|
250
|
+
# do work
|
251
|
+
end
|
252
|
+
~~~
|
253
|
+
|
254
|
+
This can be especially useful if you're running batch jobs spread across multiple hosts.
|
255
|
+
|
256
|
+
Values
|
257
|
+
------
|
258
|
+
Simple values are easy as well:
|
259
|
+
|
260
|
+
~~~ruby
|
261
|
+
@value = Redis::Value.new('value_name')
|
262
|
+
@value.value = 'a'
|
263
|
+
@value.delete
|
264
|
+
~~~
|
265
|
+
|
266
|
+
Complex data is no problem with :marshal => true:
|
267
|
+
|
268
|
+
~~~ruby
|
269
|
+
@account = Account.create!(params[:account])
|
270
|
+
@newest = Redis::Value.new('newest_account', :marshal => true)
|
271
|
+
@newest.value = @account.attributes
|
272
|
+
puts @newest.value['username']
|
273
|
+
~~~
|
274
|
+
|
275
|
+
Compress data to save memory usage on Redis with :compress => true:
|
276
|
+
|
277
|
+
~~~ruby
|
278
|
+
@account = Account.create!(params[:account])
|
279
|
+
@marshaled_value = Redis::Value.new('marshaled', :marshal => true, :compress => true)
|
280
|
+
@marshaled_value.value = @account.attributes
|
281
|
+
@unmarshaled_value = Redis::Value.new('unmarshaled', :compress => true)
|
282
|
+
@unmarshaled_value = 'Really Long String'
|
283
|
+
puts @marshaled_value.value['username']
|
284
|
+
puts @unmarshaled_value.value
|
285
|
+
~~~
|
286
|
+
|
287
|
+
Lists
|
288
|
+
-----
|
289
|
+
Lists work just like Ruby arrays:
|
290
|
+
|
291
|
+
~~~ruby
|
292
|
+
@list = Redis::List.new('list_name')
|
293
|
+
@list << 'a'
|
294
|
+
@list << 'b'
|
295
|
+
@list.include? 'c' # false
|
296
|
+
@list.values # ['a','b']
|
297
|
+
@list << 'c'
|
298
|
+
@list.delete('c')
|
299
|
+
@list[0]
|
300
|
+
@list[0,1]
|
301
|
+
@list[0..1]
|
302
|
+
@list.shift
|
303
|
+
@list.pop
|
304
|
+
@list.clear
|
305
|
+
# etc
|
306
|
+
~~~
|
307
|
+
|
308
|
+
You can bound the size of the list to only hold N elements like so:
|
309
|
+
|
310
|
+
~~~ruby
|
311
|
+
# Only holds 10 elements, throws out old ones when you reach :maxlength.
|
312
|
+
@list = Redis::List.new('list_name', :maxlength => 10)
|
313
|
+
~~~
|
314
|
+
|
315
|
+
Complex data types are serialized with :marshal => true:
|
316
|
+
|
317
|
+
~~~ruby
|
318
|
+
@list = Redis::List.new('list_name', :marshal => true)
|
319
|
+
@list << {:name => "Nate", :city => "San Diego"}
|
320
|
+
@list << {:name => "Peter", :city => "Oceanside"}
|
321
|
+
@list.each do |el|
|
322
|
+
puts "#{el[:name]} lives in #{el[:city]}"
|
323
|
+
end
|
324
|
+
~~~
|
325
|
+
|
326
|
+
Note: If you run into issues, with Marshal errors, refer to the fix in [Issue #176](https://github.com/nateware/redis-objects/issues/176).
|
327
|
+
|
328
|
+
Hashes
|
329
|
+
------
|
330
|
+
Hashes work like a Ruby [Hash](http://ruby-doc.org/core/classes/Hash.html), with
|
331
|
+
a few Redis-specific additions. (The class name is "HashKey" not just "Hash", due to
|
332
|
+
conflicts with the Ruby core Hash class in other gems.)
|
333
|
+
|
334
|
+
~~~ruby
|
335
|
+
@hash = Redis::HashKey.new('hash_name')
|
336
|
+
@hash['a'] = 1
|
337
|
+
@hash['b'] = 2
|
338
|
+
@hash.each do |k,v|
|
339
|
+
puts "#{k} = #{v}"
|
340
|
+
end
|
341
|
+
@hash['c'] = 3
|
342
|
+
puts @hash.all # {"a"=>"1","b"=>"2","c"=>"3"}
|
343
|
+
@hash.clear
|
344
|
+
~~~
|
345
|
+
|
346
|
+
Redis also adds incrementing and bulk operations:
|
347
|
+
|
348
|
+
~~~ruby
|
349
|
+
@hash.incr('c', 6) # 9
|
350
|
+
@hash.bulk_set('d' => 5, 'e' => 6)
|
351
|
+
@hash.bulk_get('d','e') # "5", "6"
|
352
|
+
~~~
|
353
|
+
|
354
|
+
Remember that numbers become strings in Redis. Unlike with other Redis data types,
|
355
|
+
`redis-objects` can't guess at your data type in this situation, since you may
|
356
|
+
actually mean to store "1.5".
|
357
|
+
|
358
|
+
Sets
|
359
|
+
----
|
360
|
+
Sets work like the Ruby [Set](http://ruby-doc.org/core/classes/Set.html) class.
|
361
|
+
They are unordered, but guarantee uniqueness of members.
|
362
|
+
|
363
|
+
~~~ruby
|
364
|
+
@set = Redis::Set.new('set_name')
|
365
|
+
@set << 'a'
|
366
|
+
@set << 'b'
|
367
|
+
@set << 'a' # dup ignored
|
368
|
+
@set.member? 'c' # false
|
369
|
+
@set.members # ['a','b']
|
370
|
+
@set.members.reverse # ['b','a']
|
371
|
+
@set.each do |member|
|
372
|
+
puts member
|
373
|
+
end
|
374
|
+
@set.clear
|
375
|
+
# etc
|
376
|
+
~~~
|
377
|
+
|
378
|
+
You can perform Redis intersections/unions/diffs easily:
|
379
|
+
|
380
|
+
~~~ruby
|
381
|
+
@set1 = Redis::Set.new('set1')
|
382
|
+
@set2 = Redis::Set.new('set2')
|
383
|
+
@set3 = Redis::Set.new('set3')
|
384
|
+
members = @set1 & @set2 # intersection
|
385
|
+
members = @set1 | @set2 # union
|
386
|
+
members = @set1 + @set2 # union
|
387
|
+
members = @set1 ^ @set2 # difference
|
388
|
+
members = @set1 - @set2 # difference
|
389
|
+
members = @set1.intersection(@set2, @set3) # multiple
|
390
|
+
members = @set1.union(@set2, @set3) # multiple
|
391
|
+
members = @set1.difference(@set2, @set3) # multiple
|
392
|
+
~~~
|
393
|
+
|
394
|
+
Or store them in Redis:
|
395
|
+
|
396
|
+
~~~ruby
|
397
|
+
@set1.interstore('intername', @set2, @set3)
|
398
|
+
members = @set1.redis.get('intername')
|
399
|
+
@set1.unionstore('unionname', @set2, @set3)
|
400
|
+
members = @set1.redis.get('unionname')
|
401
|
+
@set1.diffstore('diffname', @set2, @set3)
|
402
|
+
members = @set1.redis.get('diffname')
|
403
|
+
~~~
|
404
|
+
|
405
|
+
And use complex data types too, with :marshal => true:
|
406
|
+
|
407
|
+
~~~ruby
|
408
|
+
@set1 = Redis::Set.new('set1', :marshal => true)
|
409
|
+
@set2 = Redis::Set.new('set2', :marshal => true)
|
410
|
+
@set1 << {:name => "Nate", :city => "San Diego"}
|
411
|
+
@set1 << {:name => "Peter", :city => "Oceanside"}
|
412
|
+
@set2 << {:name => "Nate", :city => "San Diego"}
|
413
|
+
@set2 << {:name => "Jeff", :city => "Del Mar"}
|
414
|
+
|
415
|
+
@set1 & @set2 # Nate
|
416
|
+
@set1 - @set2 # Peter
|
417
|
+
@set1 | @set2 # all 3 people
|
418
|
+
~~~
|
419
|
+
|
420
|
+
Sorted Sets
|
421
|
+
-----------
|
422
|
+
Due to their unique properties, Sorted Sets work like a hybrid between
|
423
|
+
a Hash and an Array. You assign like a Hash, but retrieve like an Array:
|
424
|
+
|
425
|
+
~~~ruby
|
426
|
+
@sorted_set = Redis::SortedSet.new('number_of_posts')
|
427
|
+
@sorted_set['Nate'] = 15
|
428
|
+
@sorted_set['Peter'] = 75
|
429
|
+
@sorted_set['Jeff'] = 24
|
430
|
+
|
431
|
+
# Array access to get sorted order
|
432
|
+
@sorted_set[0..2] # => ["Nate", "Jeff", "Peter"]
|
433
|
+
@sorted_set[0,2] # => ["Nate", "Jeff"]
|
434
|
+
|
435
|
+
@sorted_set['Peter'] # => 75
|
436
|
+
@sorted_set['Jeff'] # => 24
|
437
|
+
@sorted_set.score('Jeff') # same thing (24)
|
438
|
+
|
439
|
+
@sorted_set.rank('Peter') # => 2
|
440
|
+
@sorted_set.rank('Jeff') # => 1
|
441
|
+
|
442
|
+
@sorted_set.first # => "Nate"
|
443
|
+
@sorted_set.last # => "Peter"
|
444
|
+
@sorted_set.revrange(0,2) # => ["Peter", "Jeff", "Nate"]
|
445
|
+
|
446
|
+
@sorted_set['Newbie'] = 1
|
447
|
+
@sorted_set.members # => ["Newbie", "Nate", "Jeff", "Peter"]
|
448
|
+
@sorted_set.members.reverse # => ["Peter", "Jeff", "Nate", "Newbie"]
|
449
|
+
|
450
|
+
@sorted_set.rangebyscore(10, 100, :limit => 2) # => ["Nate", "Jeff"]
|
451
|
+
@sorted_set.members(:with_scores => true) # => [["Newbie", 1], ["Nate", 16], ["Jeff", 28], ["Peter", 76]]
|
452
|
+
|
453
|
+
# atomic increment
|
454
|
+
@sorted_set.increment('Nate')
|
455
|
+
@sorted_set.incr('Peter') # shorthand
|
456
|
+
@sorted_set.incr('Jeff', 4)
|
457
|
+
~~~
|
458
|
+
|
459
|
+
The other Redis Sorted Set commands are supported as well; see [Sorted Sets API](http://redis.io/commands#sorted_set).
|
460
|
+
|
461
|
+
<a name="atomicity"></a>
|
462
|
+
Atomic Counters and Locks
|
463
|
+
-------------------------
|
464
|
+
You are probably not handling atomicity correctly in your app. For a fun rant
|
465
|
+
on the topic, see [An Atomic Rant](http://nateware.com/an-atomic-rant.html).
|
466
|
+
|
467
|
+
Atomic counters are a good way to handle concurrency:
|
468
|
+
|
469
|
+
~~~ruby
|
470
|
+
@team = Team.find(1)
|
471
|
+
if @team.drafted_players.increment <= @team.max_players
|
472
|
+
# do stuff
|
473
|
+
@team.team_players.create!(:player_id => 221)
|
474
|
+
@team.active_players.increment
|
475
|
+
else
|
476
|
+
# reset counter state
|
477
|
+
@team.drafted_players.decrement
|
478
|
+
end
|
479
|
+
~~~
|
480
|
+
|
481
|
+
An _atomic block_ gives you a cleaner way to do the above. Exceptions or returning nil
|
482
|
+
will rewind the counter back to its previous state:
|
483
|
+
|
484
|
+
~~~ruby
|
485
|
+
@team.drafted_players.increment do |val|
|
486
|
+
raise Team::TeamFullError if val > @team.max_players # rewind
|
487
|
+
@team.team_players.create!(:player_id => 221)
|
488
|
+
@team.active_players.increment
|
489
|
+
end
|
490
|
+
~~~
|
491
|
+
|
492
|
+
Here's a similar approach, using an if block (failure rewinds counter):
|
493
|
+
|
494
|
+
~~~ruby
|
495
|
+
@team.drafted_players.increment do |val|
|
496
|
+
if val <= @team.max_players
|
497
|
+
@team.team_players.create!(:player_id => 221)
|
498
|
+
@team.active_players.increment
|
499
|
+
end
|
500
|
+
end
|
501
|
+
~~~
|
502
|
+
|
503
|
+
Class methods work too, using the familiar ActiveRecord counter syntax:
|
504
|
+
|
505
|
+
~~~ruby
|
506
|
+
Team.increment_counter :drafted_players, team_id
|
507
|
+
Team.decrement_counter :drafted_players, team_id, 2
|
508
|
+
Team.increment_counter :total_online_players # no ID on global counter
|
509
|
+
~~~
|
510
|
+
|
511
|
+
Class-level atomic blocks can also be used. This may save a DB fetch, if you have
|
512
|
+
a record ID and don't need any other attributes from the DB table:
|
513
|
+
|
514
|
+
~~~ruby
|
515
|
+
Team.increment_counter(:drafted_players, team_id) do |val|
|
516
|
+
TeamPitcher.create!(:team_id => team_id, :pitcher_id => 181)
|
517
|
+
Team.increment_counter(:active_players, team_id)
|
518
|
+
end
|
519
|
+
~~~
|
520
|
+
|
521
|
+
### Locks ###
|
522
|
+
|
523
|
+
Locks work similarly. On completion or exception the lock is released:
|
524
|
+
|
525
|
+
~~~ruby
|
526
|
+
class Team < ActiveRecord::Base
|
527
|
+
lock :reorder # declare a lock
|
528
|
+
end
|
529
|
+
|
530
|
+
@team.reorder_lock.lock do
|
531
|
+
@team.reorder_all_players
|
532
|
+
end
|
533
|
+
~~~
|
534
|
+
|
535
|
+
Class-level lock (same concept)
|
536
|
+
|
537
|
+
~~~ruby
|
538
|
+
Team.obtain_lock(:reorder, team_id) do
|
539
|
+
Team.reorder_all_players(team_id)
|
540
|
+
end
|
541
|
+
~~~
|
542
|
+
|
543
|
+
Lock expiration. Sometimes you want to make sure your locks are cleaned up should
|
544
|
+
the unthinkable happen (server failure). You can set lock expirations to handle
|
545
|
+
this. Expired locks are released by the next process to attempt lock. Just
|
546
|
+
make sure you expiration value is sufficiently large compared to your expected
|
547
|
+
lock time.
|
548
|
+
|
549
|
+
~~~ruby
|
550
|
+
class Team < ActiveRecord::Base
|
551
|
+
lock :reorder, :expiration => 15.minutes
|
552
|
+
end
|
553
|
+
~~~
|
554
|
+
|
555
|
+
Keep in mind that true locks serialize your entire application at that point. As
|
556
|
+
such, atomic counters are strongly preferred.
|
557
|
+
|
558
|
+
### Expiration ###
|
559
|
+
|
560
|
+
Use :expiration and :expireat options to set default expiration.
|
561
|
+
|
562
|
+
~~~ruby
|
563
|
+
value :value_with_expiration, :expiration => 1.hour
|
564
|
+
value :value_with_expireat, :expireat => lambda { Time.now + 1.hour }
|
565
|
+
~~~
|
566
|
+
|
567
|
+
:warning: In the above example, `expiration` is evaluated at class load time.
|
568
|
+
In this example, it will be one hour after loading the class, not after one hour
|
569
|
+
after setting a value. If you want to expire one hour after setting the value,
|
570
|
+
please use `:expireat` with `lambda`.
|
571
|
+
|
572
|
+
Custom serialization
|
573
|
+
--------------------
|
574
|
+
You can customize how values are serialized by setting `serializer: CustomSerializer`.
|
575
|
+
The default is `Marshal` from the standard lib, but it can be anything that responds to `dump` and
|
576
|
+
`load`. `JSON` and `YAML` are popular options.
|
577
|
+
|
578
|
+
If you need to pass extra arguments to `dump` or `load`, you can set
|
579
|
+
`marshal_dump_args: { foo: 'bar' }` and `marshal_load_args: { foo: 'bar' }` respectively.
|
580
|
+
|
581
|
+
~~~ruby
|
582
|
+
class CustomSerializer
|
583
|
+
def self.dump(value)
|
584
|
+
# custom code for serializing
|
585
|
+
end
|
586
|
+
|
587
|
+
def self.load(value)
|
588
|
+
# custom code for deserializing
|
589
|
+
end
|
590
|
+
end
|
591
|
+
|
592
|
+
@account = Account.create!(params[:account])
|
593
|
+
@newest = Redis::Value.new('custom_serializer', marshal: true, serializer: CustomSerializer)
|
594
|
+
@newest.value = @account.attributes
|
595
|
+
~~~
|
596
|
+
|
597
|
+
Author
|
598
|
+
=======
|
599
|
+
Copyright (c) 2009-2019 [Nate Wiger](http://nateware.com). All Rights Reserved.
|
600
|
+
Released under the [Artistic License](http://www.opensource.org/licenses/artistic-license-2.0.php).
|
data/Rakefile
ADDED
@@ -0,0 +1,14 @@
|
|
1
|
+
require "bundler/gem_tasks"
|
2
|
+
|
3
|
+
desc "run all the specs"
|
4
|
+
task :test do
|
5
|
+
sh "bacon spec/*_spec.rb"
|
6
|
+
end
|
7
|
+
task :default => :test
|
8
|
+
task :spec => :test
|
9
|
+
|
10
|
+
desc "show changelog"
|
11
|
+
task :changelog do
|
12
|
+
latest = `git tag |tail -1`.chomp
|
13
|
+
sh "git log --pretty=format:'* %s %b [%an]' #{latest}..HEAD"
|
14
|
+
end
|
@@ -0,0 +1,62 @@
|
|
1
|
+
require 'redis/helpers/core_commands'
|
2
|
+
|
3
|
+
class Redis
|
4
|
+
# Defines base functionality for all redis-objects.
|
5
|
+
class BaseObject
|
6
|
+
include Redis::Helpers::CoreCommands
|
7
|
+
|
8
|
+
attr_reader :key, :options
|
9
|
+
|
10
|
+
def initialize(key, *args)
|
11
|
+
@key = key.is_a?(Array) ? key.flatten.join(':') : key
|
12
|
+
@options = args.last.is_a?(Hash) ? args.pop : {}
|
13
|
+
@myredis = Objects::ConnectionPoolProxy.proxy_if_needed(args.first)
|
14
|
+
end
|
15
|
+
|
16
|
+
# Dynamically query the handle to enable resetting midstream
|
17
|
+
def redis
|
18
|
+
@myredis || ::Redis::Objects.redis
|
19
|
+
end
|
20
|
+
|
21
|
+
alias :inspect :to_s # Ruby 1.9.2
|
22
|
+
|
23
|
+
def set_expiration
|
24
|
+
if !@options[:expiration].nil?
|
25
|
+
redis.expire(@key, @options[:expiration]) if redis.ttl(@key) < 0
|
26
|
+
elsif !@options[:expireat].nil?
|
27
|
+
expireat = @options[:expireat]
|
28
|
+
at = expireat.respond_to?(:call) ? expireat.call.to_i : expireat.to_i
|
29
|
+
redis.expireat(@key, at) if redis.ttl(@key) < 0
|
30
|
+
end
|
31
|
+
end
|
32
|
+
|
33
|
+
def allow_expiration(&block)
|
34
|
+
result = block.call
|
35
|
+
set_expiration
|
36
|
+
result
|
37
|
+
end
|
38
|
+
|
39
|
+
def to_json(*args)
|
40
|
+
to_hash.to_json(*args)
|
41
|
+
rescue NoMethodError => e
|
42
|
+
raise e.class, "The current runtime does not provide a `to_json` implementation. Require 'json' or another JSON library and try again."
|
43
|
+
end
|
44
|
+
|
45
|
+
def as_json(*)
|
46
|
+
to_hash
|
47
|
+
end
|
48
|
+
|
49
|
+
def to_hash
|
50
|
+
{ "key" => @key, "options" => @options, "value" => value }
|
51
|
+
end
|
52
|
+
|
53
|
+
# Math ops - delegate to value method
|
54
|
+
%w(== < > <= >=).each do |m|
|
55
|
+
class_eval <<-EndOverload
|
56
|
+
def #{m}(what)
|
57
|
+
value #{m} what
|
58
|
+
end
|
59
|
+
EndOverload
|
60
|
+
end
|
61
|
+
end
|
62
|
+
end
|