rollups 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/CHANGELOG.md +3 -0
- data/LICENSE.txt +21 -0
- data/README.md +386 -0
- data/lib/generators/rollups_generator.rb +26 -0
- data/lib/generators/templates/dimensions.rb.tt +12 -0
- data/lib/generators/templates/standard.rb.tt +11 -0
- data/lib/rollup.rb +98 -0
- data/lib/rollup/aggregator.rb +178 -0
- data/lib/rollup/model.rb +10 -0
- data/lib/rollup/utils.rb +75 -0
- data/lib/rollup/version.rb +5 -0
- data/lib/rollups.rb +16 -0
- metadata +180 -0
checksums.yaml
ADDED
@@ -0,0 +1,7 @@
|
|
1
|
+
---
|
2
|
+
SHA256:
|
3
|
+
metadata.gz: ac85a23af0477ab8ce7abed983995b06a935f82c2836fc0a63ff1fe4b150b9d7
|
4
|
+
data.tar.gz: bb60e3c7c87e22e68a4ba94652e011031292ce879d60824f745d7f808055684c
|
5
|
+
SHA512:
|
6
|
+
metadata.gz: abee36f374c42202b8a13919ff863aef511af7c28258eb5f657b0b00de4d8f9d329d53d1f58768377372eae428c09b7f223fbcdeb2b7a43a87339fbe3d9f3e9e
|
7
|
+
data.tar.gz: 452951f5d6a8fda80288650b7061e0e17ab9bfd76a2638a7f529b4919ec2933fb725b07e6f85392665fd4e44b69790eed1d69b88341122aa8338640e4b45c868
|
data/CHANGELOG.md
ADDED
data/LICENSE.txt
ADDED
@@ -0,0 +1,21 @@
|
|
1
|
+
The MIT License (MIT)
|
2
|
+
|
3
|
+
Copyright (c) 2020 Andrew Kane
|
4
|
+
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
7
|
+
in the Software without restriction, including without limitation the rights
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
10
|
+
furnished to do so, subject to the following conditions:
|
11
|
+
|
12
|
+
The above copyright notice and this permission notice shall be included in
|
13
|
+
all copies or substantial portions of the Software.
|
14
|
+
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
|
21
|
+
THE SOFTWARE.
|
data/README.md
ADDED
@@ -0,0 +1,386 @@
|
|
1
|
+
# Rollup
|
2
|
+
|
3
|
+
:fire: Rollup time-series data in Rails
|
4
|
+
|
5
|
+
Works great with [Ahoy](https://github.com/ankane/ahoy) and [Searchjoy](https://github.com/ankane/searchjoy)
|
6
|
+
|
7
|
+
## Installation
|
8
|
+
|
9
|
+
Add this line to your application’s Gemfile:
|
10
|
+
|
11
|
+
```ruby
|
12
|
+
gem 'rollups'
|
13
|
+
```
|
14
|
+
|
15
|
+
For Rails < 6, also add:
|
16
|
+
|
17
|
+
```ruby
|
18
|
+
gem 'activerecord-import'
|
19
|
+
```
|
20
|
+
|
21
|
+
And run:
|
22
|
+
|
23
|
+
```sh
|
24
|
+
bundle install
|
25
|
+
rails generate rollups
|
26
|
+
rails db:migrate
|
27
|
+
```
|
28
|
+
|
29
|
+
## Contents
|
30
|
+
|
31
|
+
- [Getting Started](#getting-started)
|
32
|
+
- [Creating Rollups](#creating-rollups)
|
33
|
+
- [Querying Rollups](#querying-rollups)
|
34
|
+
- [Other Topics](#other-topics)
|
35
|
+
- [Examples](#examples)
|
36
|
+
|
37
|
+
## Getting Started
|
38
|
+
|
39
|
+
Store the number of users created by day in the `rollups` table
|
40
|
+
|
41
|
+
```ruby
|
42
|
+
User.rollup("New users")
|
43
|
+
```
|
44
|
+
|
45
|
+
Get the series
|
46
|
+
|
47
|
+
```ruby
|
48
|
+
Rollup.series("New users")
|
49
|
+
# {
|
50
|
+
# Sat, 24 May 2020 => 50,
|
51
|
+
# Sun, 25 May 2020 => 100,
|
52
|
+
# Mon, 26 May 2020 => 34
|
53
|
+
# }
|
54
|
+
```
|
55
|
+
|
56
|
+
Use a rake task or background job to create rollups on a regular basis. Don’t worry too much about naming - you can [rename](#naming) later if needed.
|
57
|
+
|
58
|
+
## Creating Rollups
|
59
|
+
|
60
|
+
### Time Column
|
61
|
+
|
62
|
+
Specify the time column - `created_at` by default
|
63
|
+
|
64
|
+
```ruby
|
65
|
+
User.rollup("New users", column: :joined_at)
|
66
|
+
```
|
67
|
+
|
68
|
+
Change the default column for a model
|
69
|
+
|
70
|
+
```ruby
|
71
|
+
class User < ApplicationRecord
|
72
|
+
self.rollup_column = :joined_at
|
73
|
+
end
|
74
|
+
```
|
75
|
+
|
76
|
+
### Time Intervals
|
77
|
+
|
78
|
+
Specify the interval - `day` by default
|
79
|
+
|
80
|
+
```ruby
|
81
|
+
User.rollup("New users", interval: "week")
|
82
|
+
```
|
83
|
+
|
84
|
+
And when querying
|
85
|
+
|
86
|
+
```ruby
|
87
|
+
Rollup.series("New users", interval: "week")
|
88
|
+
```
|
89
|
+
|
90
|
+
Supported intervals are:
|
91
|
+
|
92
|
+
- hour
|
93
|
+
- day
|
94
|
+
- week
|
95
|
+
- month
|
96
|
+
- quarter
|
97
|
+
- year
|
98
|
+
|
99
|
+
Or any number of minutes or seconds:
|
100
|
+
|
101
|
+
- 1m, 5m, 15m
|
102
|
+
- 1s, 30s, 90s
|
103
|
+
|
104
|
+
Weeks start on Sunday by default. Change this with:
|
105
|
+
|
106
|
+
```ruby
|
107
|
+
Rollup.week_start = :monday
|
108
|
+
```
|
109
|
+
|
110
|
+
### Time Zones
|
111
|
+
|
112
|
+
The default time zone is `Time.zone`. Change this with:
|
113
|
+
|
114
|
+
```ruby
|
115
|
+
Rollup.time_zone = "Pacific Time (US & Canada)"
|
116
|
+
```
|
117
|
+
|
118
|
+
or
|
119
|
+
|
120
|
+
```ruby
|
121
|
+
User.rollup("New users", time_zone: "Pacific Time (US & Canada)")
|
122
|
+
```
|
123
|
+
|
124
|
+
Time zone objects also work. To see a list of available time zones in Rails, run `rake time:zones:all`.
|
125
|
+
|
126
|
+
See [date storage](#date-storage) for how dates are stored.
|
127
|
+
|
128
|
+
### Calculations
|
129
|
+
|
130
|
+
Rollups use `count` by default. For other calculations, use:
|
131
|
+
|
132
|
+
```ruby
|
133
|
+
Order.rollup("Revenue") { |r| r.sum(:revenue) }
|
134
|
+
```
|
135
|
+
|
136
|
+
Works with `count`, `sum`, `minimum`, `maximum`, and `average`. For `median` and `percentile`, check out [ActiveMedian](https://github.com/ankane/active_median).
|
137
|
+
|
138
|
+
### Dimensions
|
139
|
+
|
140
|
+
*PostgreSQL only*
|
141
|
+
|
142
|
+
Create rollups with dimensions
|
143
|
+
|
144
|
+
```ruby
|
145
|
+
Order.group(:platform).rollup("Orders by platform")
|
146
|
+
```
|
147
|
+
|
148
|
+
Works with multiple groups as well
|
149
|
+
|
150
|
+
```ruby
|
151
|
+
Order.group(:platform, :channel).rollup("Orders by platform and channel")
|
152
|
+
```
|
153
|
+
|
154
|
+
Dimension names are determined by the `group` clause. To set manually, use:
|
155
|
+
|
156
|
+
```ruby
|
157
|
+
Order.group(:channel).rollup("Orders by source", dimension_names: ["source"])
|
158
|
+
```
|
159
|
+
|
160
|
+
See how to [query dimensions](#multiple-series).
|
161
|
+
|
162
|
+
### Updating Data
|
163
|
+
|
164
|
+
When you run a rollup for the first time, the entire series is calculated. When you run it again, newer data is added.
|
165
|
+
|
166
|
+
By default, the latest interval stored for a series is recalculated, since it was likely calculated before the interval completed. Earlier intervals aren’t recalculated since the source rows may have been deleted (this also improves performance).
|
167
|
+
|
168
|
+
To recalculate the last few intervals, use:
|
169
|
+
|
170
|
+
```ruby
|
171
|
+
User.rollup("New users", last: 3)
|
172
|
+
```
|
173
|
+
|
174
|
+
To only store data for completed intervals, use:
|
175
|
+
|
176
|
+
```ruby
|
177
|
+
User.rollup("New users", current: false)
|
178
|
+
```
|
179
|
+
|
180
|
+
To clear and recalculate the entire series, use:
|
181
|
+
|
182
|
+
```ruby
|
183
|
+
User.rollup("New users", clear: true)
|
184
|
+
```
|
185
|
+
|
186
|
+
To delete a series, use:
|
187
|
+
|
188
|
+
```ruby
|
189
|
+
Rollup.where(name: "New users", interval: "day").delete_all
|
190
|
+
```
|
191
|
+
|
192
|
+
## Querying Rollups
|
193
|
+
|
194
|
+
### Single Series
|
195
|
+
|
196
|
+
Get a series
|
197
|
+
|
198
|
+
```ruby
|
199
|
+
Rollup.series("New users")
|
200
|
+
```
|
201
|
+
|
202
|
+
Specify the interval if it’s not day
|
203
|
+
|
204
|
+
```ruby
|
205
|
+
Rollup.series("New users", interval: "week")
|
206
|
+
```
|
207
|
+
|
208
|
+
If a series has dimensions, they must match exactly as well
|
209
|
+
|
210
|
+
```ruby
|
211
|
+
Rollup.series("Orders by platform and channel", dimensions: {platform: "Web", channel: "Search"})
|
212
|
+
```
|
213
|
+
|
214
|
+
### Multiple Series
|
215
|
+
|
216
|
+
*PostgreSQL only*
|
217
|
+
|
218
|
+
Get multiple series grouped by dimensions
|
219
|
+
|
220
|
+
```ruby
|
221
|
+
Rollup.multi_series("Orders by platform")
|
222
|
+
```
|
223
|
+
|
224
|
+
Specify the interval if it’s not day
|
225
|
+
|
226
|
+
```ruby
|
227
|
+
Rollup.multi_series("Orders by platform", interval: "week")
|
228
|
+
```
|
229
|
+
|
230
|
+
Filter by dimensions
|
231
|
+
|
232
|
+
```ruby
|
233
|
+
Rollup.where_dimensions(platform: "Web").multi_series("Orders by platform and channel")
|
234
|
+
```
|
235
|
+
|
236
|
+
### Raw Data
|
237
|
+
|
238
|
+
Uses the `Rollup` model to query the data directly
|
239
|
+
|
240
|
+
```ruby
|
241
|
+
Rollup.where(name: "New users", interval: "day")
|
242
|
+
```
|
243
|
+
|
244
|
+
### List
|
245
|
+
|
246
|
+
List names and intervals
|
247
|
+
|
248
|
+
```ruby
|
249
|
+
Rollup.list
|
250
|
+
```
|
251
|
+
|
252
|
+
### Charts
|
253
|
+
|
254
|
+
Rollup works great with [Chartkick](https://github.com/ankane/chartkick)
|
255
|
+
|
256
|
+
```erb
|
257
|
+
<%= line_chart Rollup.series("New users") %>
|
258
|
+
```
|
259
|
+
|
260
|
+
For multiple series, set a `name` for each series before charting
|
261
|
+
|
262
|
+
```ruby
|
263
|
+
series = Rollup.multi_series("Orders by platform")
|
264
|
+
series.each do |s|
|
265
|
+
s[:name] = s[:dimensions]["platform"]
|
266
|
+
end
|
267
|
+
```
|
268
|
+
|
269
|
+
## Other Topics
|
270
|
+
|
271
|
+
### Naming
|
272
|
+
|
273
|
+
Use any naming convention you prefer. Some ideas are:
|
274
|
+
|
275
|
+
- Human - `New users`
|
276
|
+
- Underscore - `new_users`
|
277
|
+
- Dots - `new_users.count`
|
278
|
+
|
279
|
+
Rename with:
|
280
|
+
|
281
|
+
```ruby
|
282
|
+
Rollup.rename("Old name", "New name")
|
283
|
+
```
|
284
|
+
|
285
|
+
### Date Storage
|
286
|
+
|
287
|
+
Rollup stores both dates and times in the `time` column depending on the interval. For date intervals (day, week, etc), it stores `00:00:00` for the time part. Cast the `time` column to a date when querying in SQL to get the correct value.
|
288
|
+
|
289
|
+
- PostgreSQL: `time::date`
|
290
|
+
- MySQL: `CAST(time AS date)`
|
291
|
+
- SQLite: `date(time)`
|
292
|
+
|
293
|
+
## Examples
|
294
|
+
|
295
|
+
- [Ahoy](#ahoy)
|
296
|
+
- [Searchjoy](#searchjoy)
|
297
|
+
|
298
|
+
### Ahoy
|
299
|
+
|
300
|
+
Set the default rollup column for your models
|
301
|
+
|
302
|
+
```ruby
|
303
|
+
class Ahoy::Visit < ApplicationRecord
|
304
|
+
self.rollup_column = :started_at
|
305
|
+
end
|
306
|
+
```
|
307
|
+
|
308
|
+
and
|
309
|
+
|
310
|
+
```ruby
|
311
|
+
class Ahoy::Event < ApplicationRecord
|
312
|
+
self.rollup_column = :time
|
313
|
+
end
|
314
|
+
```
|
315
|
+
|
316
|
+
Hourly visits
|
317
|
+
|
318
|
+
```ruby
|
319
|
+
Ahoy::Visit.rollup("Visits", interval: "hour")
|
320
|
+
```
|
321
|
+
|
322
|
+
Visits by browser
|
323
|
+
|
324
|
+
```ruby
|
325
|
+
Ahoy::Visit.group(:browser).rollup("Visits by browser")
|
326
|
+
```
|
327
|
+
|
328
|
+
Unique homepage views
|
329
|
+
|
330
|
+
```ruby
|
331
|
+
Ahoy::Event.where(name: "Viewed homepage").joins(:visit).rollup("Homepage views") { |r| r.distinct.count(:visitor_token) }
|
332
|
+
```
|
333
|
+
|
334
|
+
Product views
|
335
|
+
|
336
|
+
```ruby
|
337
|
+
Ahoy::Event.where(name: "Viewed product").group_prop(:product_id).rollup("Product views")
|
338
|
+
```
|
339
|
+
|
340
|
+
### Searchjoy
|
341
|
+
|
342
|
+
Daily searches
|
343
|
+
|
344
|
+
```ruby
|
345
|
+
Searchjoy::Search.rollup("Searches")
|
346
|
+
```
|
347
|
+
|
348
|
+
Searches by query
|
349
|
+
|
350
|
+
```ruby
|
351
|
+
Searchjoy::Search.group(:normalized_query).rollup("Searches by query", dimension_names: ["query"])
|
352
|
+
```
|
353
|
+
|
354
|
+
Conversion rate
|
355
|
+
|
356
|
+
```ruby
|
357
|
+
Searchjoy::Search.rollup("Search conversion rate") { |r| r.average("(converted_at IS NOT NULL)::int") }
|
358
|
+
```
|
359
|
+
|
360
|
+
## History
|
361
|
+
|
362
|
+
View the [changelog](https://github.com/ankane/rollup/blob/master/CHANGELOG.md)
|
363
|
+
|
364
|
+
## Contributing
|
365
|
+
|
366
|
+
Everyone is encouraged to help improve this project. Here are a few ways you can help:
|
367
|
+
|
368
|
+
- [Report bugs](https://github.com/ankane/rollup/issues)
|
369
|
+
- Fix bugs and [submit pull requests](https://github.com/ankane/rollup/pulls)
|
370
|
+
- Write, clarify, or fix documentation
|
371
|
+
- Suggest or add new features
|
372
|
+
|
373
|
+
To get started with development:
|
374
|
+
|
375
|
+
```sh
|
376
|
+
git clone https://github.com/ankane/rollup.git
|
377
|
+
cd rollup
|
378
|
+
bundle install
|
379
|
+
|
380
|
+
# create databases
|
381
|
+
createdb rollup_test
|
382
|
+
mysqladmin create rollup_test
|
383
|
+
|
384
|
+
# run tests
|
385
|
+
bundle exec rake test
|
386
|
+
```
|
@@ -0,0 +1,26 @@
|
|
1
|
+
require "rails/generators/active_record"
|
2
|
+
|
3
|
+
# use rollups instead of rollup:install to avoid
|
4
|
+
# class Rollup < ActiveRecord::Base
|
5
|
+
# also works out nicely since it's the gem name
|
6
|
+
class RollupsGenerator < Rails::Generators::Base
|
7
|
+
include ActiveRecord::Generators::Migration
|
8
|
+
source_root File.join(__dir__, "templates")
|
9
|
+
|
10
|
+
def copy_templates
|
11
|
+
migration_template migration_source, "db/migrate/create_rollups.rb", migration_version: migration_version
|
12
|
+
end
|
13
|
+
|
14
|
+
def migration_source
|
15
|
+
case ActiveRecord::Base.connection_config[:adapter].to_s
|
16
|
+
when /postg/i
|
17
|
+
"dimensions.rb"
|
18
|
+
else
|
19
|
+
"standard.rb"
|
20
|
+
end
|
21
|
+
end
|
22
|
+
|
23
|
+
def migration_version
|
24
|
+
"[#{ActiveRecord::VERSION::MAJOR}.#{ActiveRecord::VERSION::MINOR}]"
|
25
|
+
end
|
26
|
+
end
|
@@ -0,0 +1,12 @@
|
|
1
|
+
class <%= migration_class_name %> < ActiveRecord::Migration<%= migration_version %>
|
2
|
+
def change
|
3
|
+
create_table :rollups do |t|
|
4
|
+
t.string :name, null: false
|
5
|
+
t.string :interval, null: false
|
6
|
+
t.datetime :time, null: false
|
7
|
+
t.jsonb :dimensions, null: false, default: {}
|
8
|
+
t.float :value
|
9
|
+
end
|
10
|
+
add_index :rollups, [:name, :interval, :time, :dimensions], unique: true
|
11
|
+
end
|
12
|
+
end
|
@@ -0,0 +1,11 @@
|
|
1
|
+
class <%= migration_class_name %> < ActiveRecord::Migration<%= migration_version %>
|
2
|
+
def change
|
3
|
+
create_table :rollups do |t|
|
4
|
+
t.string :name, null: false
|
5
|
+
t.string :interval, null: false
|
6
|
+
t.datetime :time, null: false
|
7
|
+
t.float :value
|
8
|
+
end
|
9
|
+
add_index :rollups, [:name, :interval, :time], unique: true
|
10
|
+
end
|
11
|
+
end
|
data/lib/rollup.rb
ADDED
@@ -0,0 +1,98 @@
|
|
1
|
+
class Rollup < ActiveRecord::Base
|
2
|
+
validates :name, presence: true
|
3
|
+
validates :interval, presence: true
|
4
|
+
validates :time, presence: true
|
5
|
+
|
6
|
+
class << self
|
7
|
+
attr_accessor :week_start
|
8
|
+
attr_writer :time_zone
|
9
|
+
end
|
10
|
+
self.week_start = :sunday
|
11
|
+
|
12
|
+
class << self
|
13
|
+
# do not memoize so Time.zone can change
|
14
|
+
def time_zone
|
15
|
+
(defined?(@time_zone) && @time_zone) || Time.zone || "Etc/UTC"
|
16
|
+
end
|
17
|
+
|
18
|
+
def series(name, interval: "day", dimensions: {})
|
19
|
+
Utils.check_dimensions if dimensions.any?
|
20
|
+
|
21
|
+
relation = where(name: name, interval: interval)
|
22
|
+
relation = relation.where(dimensions: dimensions) if Utils.dimensions_supported?
|
23
|
+
|
24
|
+
# use select_all due to incorrect casting with pluck
|
25
|
+
sql = relation.order(:time).select(Utils.time_sql(interval), :value).to_sql
|
26
|
+
result = connection.select_all(sql).rows
|
27
|
+
|
28
|
+
Utils.make_series(result, interval)
|
29
|
+
end
|
30
|
+
|
31
|
+
def multi_series(name, interval: "day")
|
32
|
+
Utils.check_dimensions
|
33
|
+
|
34
|
+
relation = where(name: name, interval: interval)
|
35
|
+
|
36
|
+
# use select_all to reduce allocations
|
37
|
+
sql = relation.order(:time).select(Utils.time_sql(interval), :value, :dimensions).to_sql
|
38
|
+
result = connection.select_all(sql).rows
|
39
|
+
|
40
|
+
result.group_by { |r| JSON.parse(r[2]) }.map do |dimensions, rollups|
|
41
|
+
{dimensions: dimensions, data: Utils.make_series(rollups, interval)}
|
42
|
+
end
|
43
|
+
end
|
44
|
+
|
45
|
+
def where_dimensions(dimensions)
|
46
|
+
Utils.check_dimensions
|
47
|
+
|
48
|
+
relation = self
|
49
|
+
dimensions.each do |k, v|
|
50
|
+
k = k.to_s
|
51
|
+
relation =
|
52
|
+
if v.nil?
|
53
|
+
relation.where("dimensions ->> ? IS NULL", k)
|
54
|
+
elsif v.is_a?(Array)
|
55
|
+
relation.where("dimensions ->> ? IN (?)", k, v.map { |vi| vi.as_json.to_s })
|
56
|
+
else
|
57
|
+
relation.where("dimensions ->> ? = ?", k, v.as_json.to_s)
|
58
|
+
end
|
59
|
+
end
|
60
|
+
relation
|
61
|
+
end
|
62
|
+
|
63
|
+
def list
|
64
|
+
select(:name, :interval).distinct.order(:name, :interval).map do |r|
|
65
|
+
{name: r.name, interval: r.interval}
|
66
|
+
end
|
67
|
+
end
|
68
|
+
|
69
|
+
# TODO maybe use in_batches
|
70
|
+
def rename(old_name, new_name)
|
71
|
+
where(name: old_name).update_all(name: new_name)
|
72
|
+
end
|
73
|
+
end
|
74
|
+
|
75
|
+
# feels cleaner than overriding _read_attribute
|
76
|
+
def inspect
|
77
|
+
if Utils.date_interval?(interval)
|
78
|
+
previous_time = time_before_type_cast
|
79
|
+
previous_time = previous_time.to_s(:db) if previous_time.is_a?(Time)
|
80
|
+
previous_time = previous_time + " 00:00:00" if previous_time.length == 10
|
81
|
+
super.sub("time: \"#{previous_time}\"", "time: \"#{time.to_s(:db)}\"")
|
82
|
+
else
|
83
|
+
super
|
84
|
+
end
|
85
|
+
end
|
86
|
+
|
87
|
+
def time
|
88
|
+
if Utils.date_interval?(interval) && !time_before_type_cast.nil?
|
89
|
+
if time_before_type_cast.is_a?(Time)
|
90
|
+
time_before_type_cast.utc.to_date
|
91
|
+
else
|
92
|
+
Date.parse(time_before_type_cast.to_s)
|
93
|
+
end
|
94
|
+
else
|
95
|
+
super
|
96
|
+
end
|
97
|
+
end
|
98
|
+
end
|
@@ -0,0 +1,178 @@
|
|
1
|
+
class Rollup
|
2
|
+
class Aggregator
|
3
|
+
def initialize(klass)
|
4
|
+
@klass = klass # or relation
|
5
|
+
end
|
6
|
+
|
7
|
+
def rollup(name, column: nil, interval: "day", dimension_names: nil, time_zone: nil, current: true, last: nil, clear: false, &block)
|
8
|
+
raise "Name can't be blank" if name.blank?
|
9
|
+
|
10
|
+
column ||= @klass.rollup_column || :created_at
|
11
|
+
validate_column(column)
|
12
|
+
|
13
|
+
relation = perform_group(name, column: column, interval: interval, time_zone: time_zone, current: current, last: last, clear: clear)
|
14
|
+
result = perform_calculation(relation, &block)
|
15
|
+
|
16
|
+
dimension_names = set_dimension_names(dimension_names, relation)
|
17
|
+
records = prepare_result(result, name, dimension_names, interval)
|
18
|
+
|
19
|
+
maybe_clear(clear, name, interval) do
|
20
|
+
save_records(records) if records.any?
|
21
|
+
end
|
22
|
+
end
|
23
|
+
|
24
|
+
# basic version of Active Record disallow_raw_sql!
|
25
|
+
# symbol = column (safe), Arel node = SQL (safe), other = untrusted
|
26
|
+
# no need to quote/resolve column here, as Groupdate handles it
|
27
|
+
# TODO remove this method when gem depends on Groupdate 6+
|
28
|
+
def validate_column(column)
|
29
|
+
# matches table.column and column
|
30
|
+
unless column.is_a?(Symbol) || column.is_a?(Arel::Nodes::SqlLiteral) || /\A\w+(\.\w+)?\z/i.match(column.to_s)
|
31
|
+
raise "Non-attribute argument: #{column}. Use Arel.sql() for known-safe values"
|
32
|
+
end
|
33
|
+
end
|
34
|
+
|
35
|
+
def perform_group(name, column:, interval:, time_zone:, current:, last:, clear:)
|
36
|
+
raise ArgumentError, "Cannot use last and clear together" if last && clear
|
37
|
+
|
38
|
+
gd_options = {
|
39
|
+
current: current
|
40
|
+
}
|
41
|
+
|
42
|
+
# make sure Groupdate global options aren't applied
|
43
|
+
gd_options[:time_zone] = time_zone || Rollup.time_zone
|
44
|
+
gd_options[:week_start] = Rollup.week_start if interval.to_s == "week"
|
45
|
+
gd_options[:day_start] = 0 if Utils.date_interval?(interval)
|
46
|
+
|
47
|
+
if last
|
48
|
+
gd_options[:last] = last
|
49
|
+
elsif !clear
|
50
|
+
# if no rollups, compute all intervals
|
51
|
+
# if rollups, recompute last interval
|
52
|
+
max_time = Rollup.where(name: name, interval: interval).maximum(Utils.time_sql(interval))
|
53
|
+
if max_time
|
54
|
+
# aligns perfectly if time zone doesn't change
|
55
|
+
# if time zone does change, there are other problems besides this
|
56
|
+
gd_options[:range] = max_time..
|
57
|
+
end
|
58
|
+
end
|
59
|
+
|
60
|
+
# intervals are stored as given
|
61
|
+
# we don't normalize intervals (i.e. change 60s -> 1m)
|
62
|
+
case interval.to_s
|
63
|
+
when "hour", "day", "week", "month", "quarter", "year"
|
64
|
+
@klass.group_by_period(interval, column, **gd_options)
|
65
|
+
when /\A\d+s\z/
|
66
|
+
@klass.group_by_second(column, n: interval.to_i, **gd_options)
|
67
|
+
when /\A\d+m\z/
|
68
|
+
@klass.group_by_minute(column, n: interval.to_i, **gd_options)
|
69
|
+
else
|
70
|
+
raise ArgumentError, "Invalid interval: #{interval}"
|
71
|
+
end
|
72
|
+
end
|
73
|
+
|
74
|
+
def set_dimension_names(dimension_names, relation)
|
75
|
+
groups = relation.group_values[0..-2]
|
76
|
+
|
77
|
+
if dimension_names
|
78
|
+
Utils.check_dimensions
|
79
|
+
if dimension_names.size != groups.size
|
80
|
+
raise ArgumentError, "Expected dimension_names to be size #{groups.size}, not #{dimension_names.size}"
|
81
|
+
end
|
82
|
+
dimension_names
|
83
|
+
else
|
84
|
+
Utils.check_dimensions if groups.any?
|
85
|
+
groups.map { |group| determine_dimension_name(group) }
|
86
|
+
end
|
87
|
+
end
|
88
|
+
|
89
|
+
def determine_dimension_name(group)
|
90
|
+
# split by ., ->>, and -> and remove whitespace
|
91
|
+
value = group.to_s.split(/\s*((\.)|(->>)|(->))\s*/).last
|
92
|
+
|
93
|
+
# removing starting and ending quotes
|
94
|
+
# for simplicity, they don't need to be the same
|
95
|
+
value = value[1..-2] if value.match(/\A["'`].+["'`]\z/)
|
96
|
+
|
97
|
+
unless value.match(/\A\w+\z/)
|
98
|
+
raise "Cannot determine dimension name: #{group}. Use the dimension_names option"
|
99
|
+
end
|
100
|
+
|
101
|
+
value
|
102
|
+
end
|
103
|
+
|
104
|
+
# calculation can mutate relation, but that's fine
|
105
|
+
def perform_calculation(relation, &block)
|
106
|
+
if block_given?
|
107
|
+
yield(relation)
|
108
|
+
else
|
109
|
+
relation.count
|
110
|
+
end
|
111
|
+
end
|
112
|
+
|
113
|
+
def prepare_result(result, name, dimension_names, interval)
|
114
|
+
raise "Expected calculation to return Hash, not #{result.class.name}" unless result.is_a?(Hash)
|
115
|
+
|
116
|
+
time_class = Utils.date_interval?(interval) ? Date : Time
|
117
|
+
dimensions_supported = Utils.dimensions_supported?
|
118
|
+
expected_key_size = dimension_names.size + 1
|
119
|
+
|
120
|
+
result.map do |key, value|
|
121
|
+
dimensions = {}
|
122
|
+
if dimensions_supported && dimension_names.any?
|
123
|
+
unless key.is_a?(Array) && key.size == expected_key_size
|
124
|
+
raise "Expected result key to be Array with size #{expected_key_size}"
|
125
|
+
end
|
126
|
+
time = key[-1]
|
127
|
+
# may be able to support dimensions in SQLite by sorting dimension names
|
128
|
+
dimension_names.each_with_index do |dn, i|
|
129
|
+
dimensions[dn] = key[i]
|
130
|
+
end
|
131
|
+
else
|
132
|
+
time = key
|
133
|
+
end
|
134
|
+
|
135
|
+
raise "Expected time to be #{time_class.name}, not #{time.class.name}" unless time.is_a?(time_class)
|
136
|
+
raise "Expected value to be Numeric or nil, not #{value.class.name}" unless value.is_a?(Numeric) || value.nil?
|
137
|
+
|
138
|
+
record = {
|
139
|
+
name: name,
|
140
|
+
interval: interval,
|
141
|
+
time: time,
|
142
|
+
value: value
|
143
|
+
}
|
144
|
+
record[:dimensions] = dimensions if dimensions_supported
|
145
|
+
record
|
146
|
+
end
|
147
|
+
end
|
148
|
+
|
149
|
+
def maybe_clear(clear, name, interval)
|
150
|
+
if clear
|
151
|
+
Rollup.transaction do
|
152
|
+
Rollup.where(name: name, interval: interval).delete_all
|
153
|
+
yield
|
154
|
+
end
|
155
|
+
else
|
156
|
+
yield
|
157
|
+
end
|
158
|
+
end
|
159
|
+
|
160
|
+
def save_records(records)
|
161
|
+
# order must match unique index
|
162
|
+
# consider using index name instead
|
163
|
+
conflict_target = [:name, :interval, :time]
|
164
|
+
conflict_target << :dimensions if Utils.dimensions_supported?
|
165
|
+
|
166
|
+
if ActiveRecord::VERSION::MAJOR >= 6
|
167
|
+
options = Utils.mysql? ? {} : {unique_by: conflict_target}
|
168
|
+
Rollup.upsert_all(records, **options)
|
169
|
+
else
|
170
|
+
update = Utils.mysql? ? [:value] : {columns: [:value], conflict_target: conflict_target}
|
171
|
+
Rollup.import(records,
|
172
|
+
on_duplicate_key_update: update,
|
173
|
+
validate: false
|
174
|
+
)
|
175
|
+
end
|
176
|
+
end
|
177
|
+
end
|
178
|
+
end
|
data/lib/rollup/model.rb
ADDED
data/lib/rollup/utils.rb
ADDED
@@ -0,0 +1,75 @@
|
|
1
|
+
class Rollup
|
2
|
+
module Utils
|
3
|
+
DATE_INTERVALS = %w(day week month quarter year)
|
4
|
+
|
5
|
+
class << self
|
6
|
+
def time_sql(interval)
|
7
|
+
if date_interval?(interval)
|
8
|
+
if postgresql?
|
9
|
+
"rollups.time::date"
|
10
|
+
elsif sqlite?
|
11
|
+
"date(rollups.time)"
|
12
|
+
else
|
13
|
+
"CAST(rollups.time AS date)"
|
14
|
+
end
|
15
|
+
else
|
16
|
+
:time
|
17
|
+
end
|
18
|
+
end
|
19
|
+
|
20
|
+
def date_interval?(interval)
|
21
|
+
DATE_INTERVALS.include?(interval.to_s)
|
22
|
+
end
|
23
|
+
|
24
|
+
def dimensions_supported?
|
25
|
+
unless defined?(@dimensions_supported)
|
26
|
+
@dimensions_supported = postgresql? && Rollup.column_names.include?("dimensions")
|
27
|
+
end
|
28
|
+
@dimensions_supported
|
29
|
+
end
|
30
|
+
|
31
|
+
def check_dimensions
|
32
|
+
raise "Dimensions not supported" unless dimensions_supported?
|
33
|
+
end
|
34
|
+
|
35
|
+
def adapter_name
|
36
|
+
Rollup.connection.adapter_name
|
37
|
+
end
|
38
|
+
|
39
|
+
def postgresql?
|
40
|
+
adapter_name =~ /postg/i
|
41
|
+
end
|
42
|
+
|
43
|
+
def mysql?
|
44
|
+
adapter_name =~ /mysql/i
|
45
|
+
end
|
46
|
+
|
47
|
+
def sqlite?
|
48
|
+
adapter_name =~ /sqlite/i
|
49
|
+
end
|
50
|
+
|
51
|
+
def make_series(result, interval)
|
52
|
+
series = {}
|
53
|
+
if Utils.date_interval?(interval)
|
54
|
+
result.each do |row|
|
55
|
+
series[row[0].to_date] = row[1]
|
56
|
+
end
|
57
|
+
else
|
58
|
+
time_zone = Rollup.time_zone
|
59
|
+
if result.any? && result[0][0].is_a?(Time)
|
60
|
+
result.each do |row|
|
61
|
+
series[row[0].in_time_zone(time_zone)] = row[1]
|
62
|
+
end
|
63
|
+
else
|
64
|
+
utc = ActiveSupport::TimeZone["Etc/UTC"]
|
65
|
+
result.each do |row|
|
66
|
+
# row can be time or string
|
67
|
+
series[utc.parse(row[0]).in_time_zone(time_zone)] = row[1]
|
68
|
+
end
|
69
|
+
end
|
70
|
+
end
|
71
|
+
series
|
72
|
+
end
|
73
|
+
end
|
74
|
+
end
|
75
|
+
end
|
data/lib/rollups.rb
ADDED
@@ -0,0 +1,16 @@
|
|
1
|
+
# dependencies
|
2
|
+
require "active_support"
|
3
|
+
require "groupdate"
|
4
|
+
|
5
|
+
ActiveSupport.on_load(:active_record) do
|
6
|
+
# must come first
|
7
|
+
require "rollup"
|
8
|
+
|
9
|
+
require "rollup/model"
|
10
|
+
extend Rollup::Model
|
11
|
+
|
12
|
+
# modules
|
13
|
+
require "rollup/aggregator"
|
14
|
+
require "rollup/utils"
|
15
|
+
require "rollup/version"
|
16
|
+
end
|
metadata
ADDED
@@ -0,0 +1,180 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: rollups
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 0.1.0
|
5
|
+
platform: ruby
|
6
|
+
authors:
|
7
|
+
- Andrew Kane
|
8
|
+
autorequire:
|
9
|
+
bindir: bin
|
10
|
+
cert_chain: []
|
11
|
+
date: 2020-09-07 00:00:00.000000000 Z
|
12
|
+
dependencies:
|
13
|
+
- !ruby/object:Gem::Dependency
|
14
|
+
name: activesupport
|
15
|
+
requirement: !ruby/object:Gem::Requirement
|
16
|
+
requirements:
|
17
|
+
- - ">="
|
18
|
+
- !ruby/object:Gem::Version
|
19
|
+
version: '5.1'
|
20
|
+
type: :runtime
|
21
|
+
prerelease: false
|
22
|
+
version_requirements: !ruby/object:Gem::Requirement
|
23
|
+
requirements:
|
24
|
+
- - ">="
|
25
|
+
- !ruby/object:Gem::Version
|
26
|
+
version: '5.1'
|
27
|
+
- !ruby/object:Gem::Dependency
|
28
|
+
name: groupdate
|
29
|
+
requirement: !ruby/object:Gem::Requirement
|
30
|
+
requirements:
|
31
|
+
- - ">="
|
32
|
+
- !ruby/object:Gem::Version
|
33
|
+
version: '5.2'
|
34
|
+
type: :runtime
|
35
|
+
prerelease: false
|
36
|
+
version_requirements: !ruby/object:Gem::Requirement
|
37
|
+
requirements:
|
38
|
+
- - ">="
|
39
|
+
- !ruby/object:Gem::Version
|
40
|
+
version: '5.2'
|
41
|
+
- !ruby/object:Gem::Dependency
|
42
|
+
name: bundler
|
43
|
+
requirement: !ruby/object:Gem::Requirement
|
44
|
+
requirements:
|
45
|
+
- - ">="
|
46
|
+
- !ruby/object:Gem::Version
|
47
|
+
version: '0'
|
48
|
+
type: :development
|
49
|
+
prerelease: false
|
50
|
+
version_requirements: !ruby/object:Gem::Requirement
|
51
|
+
requirements:
|
52
|
+
- - ">="
|
53
|
+
- !ruby/object:Gem::Version
|
54
|
+
version: '0'
|
55
|
+
- !ruby/object:Gem::Dependency
|
56
|
+
name: rake
|
57
|
+
requirement: !ruby/object:Gem::Requirement
|
58
|
+
requirements:
|
59
|
+
- - ">="
|
60
|
+
- !ruby/object:Gem::Version
|
61
|
+
version: '0'
|
62
|
+
type: :development
|
63
|
+
prerelease: false
|
64
|
+
version_requirements: !ruby/object:Gem::Requirement
|
65
|
+
requirements:
|
66
|
+
- - ">="
|
67
|
+
- !ruby/object:Gem::Version
|
68
|
+
version: '0'
|
69
|
+
- !ruby/object:Gem::Dependency
|
70
|
+
name: minitest
|
71
|
+
requirement: !ruby/object:Gem::Requirement
|
72
|
+
requirements:
|
73
|
+
- - ">="
|
74
|
+
- !ruby/object:Gem::Version
|
75
|
+
version: '5'
|
76
|
+
type: :development
|
77
|
+
prerelease: false
|
78
|
+
version_requirements: !ruby/object:Gem::Requirement
|
79
|
+
requirements:
|
80
|
+
- - ">="
|
81
|
+
- !ruby/object:Gem::Version
|
82
|
+
version: '5'
|
83
|
+
- !ruby/object:Gem::Dependency
|
84
|
+
name: activerecord
|
85
|
+
requirement: !ruby/object:Gem::Requirement
|
86
|
+
requirements:
|
87
|
+
- - ">="
|
88
|
+
- !ruby/object:Gem::Version
|
89
|
+
version: '0'
|
90
|
+
type: :development
|
91
|
+
prerelease: false
|
92
|
+
version_requirements: !ruby/object:Gem::Requirement
|
93
|
+
requirements:
|
94
|
+
- - ">="
|
95
|
+
- !ruby/object:Gem::Version
|
96
|
+
version: '0'
|
97
|
+
- !ruby/object:Gem::Dependency
|
98
|
+
name: pg
|
99
|
+
requirement: !ruby/object:Gem::Requirement
|
100
|
+
requirements:
|
101
|
+
- - ">="
|
102
|
+
- !ruby/object:Gem::Version
|
103
|
+
version: '0'
|
104
|
+
type: :development
|
105
|
+
prerelease: false
|
106
|
+
version_requirements: !ruby/object:Gem::Requirement
|
107
|
+
requirements:
|
108
|
+
- - ">="
|
109
|
+
- !ruby/object:Gem::Version
|
110
|
+
version: '0'
|
111
|
+
- !ruby/object:Gem::Dependency
|
112
|
+
name: mysql2
|
113
|
+
requirement: !ruby/object:Gem::Requirement
|
114
|
+
requirements:
|
115
|
+
- - ">="
|
116
|
+
- !ruby/object:Gem::Version
|
117
|
+
version: '0'
|
118
|
+
type: :development
|
119
|
+
prerelease: false
|
120
|
+
version_requirements: !ruby/object:Gem::Requirement
|
121
|
+
requirements:
|
122
|
+
- - ">="
|
123
|
+
- !ruby/object:Gem::Version
|
124
|
+
version: '0'
|
125
|
+
- !ruby/object:Gem::Dependency
|
126
|
+
name: sqlite3
|
127
|
+
requirement: !ruby/object:Gem::Requirement
|
128
|
+
requirements:
|
129
|
+
- - ">="
|
130
|
+
- !ruby/object:Gem::Version
|
131
|
+
version: '0'
|
132
|
+
type: :development
|
133
|
+
prerelease: false
|
134
|
+
version_requirements: !ruby/object:Gem::Requirement
|
135
|
+
requirements:
|
136
|
+
- - ">="
|
137
|
+
- !ruby/object:Gem::Version
|
138
|
+
version: '0'
|
139
|
+
description:
|
140
|
+
email: andrew@chartkick.com
|
141
|
+
executables: []
|
142
|
+
extensions: []
|
143
|
+
extra_rdoc_files: []
|
144
|
+
files:
|
145
|
+
- CHANGELOG.md
|
146
|
+
- LICENSE.txt
|
147
|
+
- README.md
|
148
|
+
- lib/generators/rollups_generator.rb
|
149
|
+
- lib/generators/templates/dimensions.rb.tt
|
150
|
+
- lib/generators/templates/standard.rb.tt
|
151
|
+
- lib/rollup.rb
|
152
|
+
- lib/rollup/aggregator.rb
|
153
|
+
- lib/rollup/model.rb
|
154
|
+
- lib/rollup/utils.rb
|
155
|
+
- lib/rollup/version.rb
|
156
|
+
- lib/rollups.rb
|
157
|
+
homepage: https://github.com/ankane/rollup
|
158
|
+
licenses:
|
159
|
+
- MIT
|
160
|
+
metadata: {}
|
161
|
+
post_install_message:
|
162
|
+
rdoc_options: []
|
163
|
+
require_paths:
|
164
|
+
- lib
|
165
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
166
|
+
requirements:
|
167
|
+
- - ">="
|
168
|
+
- !ruby/object:Gem::Version
|
169
|
+
version: '2.6'
|
170
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
171
|
+
requirements:
|
172
|
+
- - ">="
|
173
|
+
- !ruby/object:Gem::Version
|
174
|
+
version: '0'
|
175
|
+
requirements: []
|
176
|
+
rubygems_version: 3.1.2
|
177
|
+
signing_key:
|
178
|
+
specification_version: 4
|
179
|
+
summary: Rollup time-series data in Rails
|
180
|
+
test_files: []
|