embulk-input-mixpanel 0.5.2 → 0.5.3.alpha.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/embulk-input-mixpanel.gemspec +1 -2
- data/lib/embulk/input/mixpanel.rb +62 -31
- data/lib/range_generator.rb +1 -1
- data/test/embulk/input/test_mixpanel.rb +122 -1
- metadata +5 -5
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 28a4798aa8352ee4fbf06a23a57ff44aeb4cb902
|
4
|
+
data.tar.gz: 433333137ed26c3a2d62b822b767621885095cf4
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 2d5fbba923cf5e52c169ed94001633001a2fcd8bb83cf0138541581d5a7768e4fd51bf7d690f7105c3fb31cf480a27a1af4fa9633636c8e4a237ba5786c99722
|
7
|
+
data.tar.gz: a147657608530a5f0ef513fdd150e93cf62bee74abeeadc3046ad30df42d19c1b94c4f42606260c827f1785dfc3c5c9fef6cf231efbc55ef4351d7be38cd4002
|
@@ -1,7 +1,6 @@
|
|
1
|
-
|
2
1
|
Gem::Specification.new do |spec|
|
3
2
|
spec.name = "embulk-input-mixpanel"
|
4
|
-
spec.version = "0.5.
|
3
|
+
spec.version = "0.5.3.alpha.1"
|
5
4
|
spec.authors = ["yoshihara", "uu59"]
|
6
5
|
spec.summary = "Mixpanel input plugin for Embulk"
|
7
6
|
spec.description = "Loads records from Mixpanel."
|
@@ -29,17 +29,32 @@ module Embulk
|
|
29
29
|
# between each 7 (SLICE_DAYS_COUNT) days.
|
30
30
|
SLICE_DAYS_COUNT = 7
|
31
31
|
|
32
|
+
DEFAULT_TIME_COLUMN = 'time'
|
33
|
+
|
32
34
|
def self.transaction(config, &control)
|
33
35
|
timezone = config.param(:timezone, :string)
|
34
36
|
TimezoneValidator.new(timezone).validate
|
35
37
|
|
36
38
|
from_date = config.param(:from_date, :string, default: (Date.today - 2).to_s)
|
37
39
|
fetch_days = config.param(:fetch_days, :integer, default: nil)
|
38
|
-
|
39
|
-
Embulk.logger.info "Try to fetch data from #{range.first} to #{range.last}"
|
40
|
+
|
40
41
|
|
41
42
|
fetch_unknown_columns = config.param(:fetch_unknown_columns, :bool, default: false)
|
42
43
|
|
44
|
+
incremental_column = config.param(:incremental_column, :string, default: nil)
|
45
|
+
incremental = config.param(:incremental, :bool, default: true)
|
46
|
+
latest_fetched_time = config.param(:latest_fetched_time, :integer, default: 0)
|
47
|
+
|
48
|
+
# Backfill from date if incremental and an incremental field is set and we are in incremental run
|
49
|
+
if incremental && !incremental_column.nil? && latest_fetched_time !=0
|
50
|
+
back_fill_days = config.param(:back_fill_days, :integer, default: 5)
|
51
|
+
puts "Backfill days #{back_fill_days}"
|
52
|
+
from_date = (Date.parse(from_date) - back_fill_days).to_s
|
53
|
+
fetch_days = fetch_days.nil? ? nil : fetch_days + back_fill_days
|
54
|
+
end
|
55
|
+
|
56
|
+
range = RangeGenerator.new(from_date, fetch_days).generate_range
|
57
|
+
Embulk.logger.info "Try to fetch data from #{range.first} to #{range.last}"
|
43
58
|
task = {
|
44
59
|
params: export_params(config),
|
45
60
|
dates: range,
|
@@ -50,8 +65,10 @@ module Embulk
|
|
50
65
|
fetch_unknown_columns: fetch_unknown_columns,
|
51
66
|
fetch_custom_properties: config.param(:fetch_custom_properties, :bool, default: true),
|
52
67
|
retry_initial_wait_sec: config.param(:retry_initial_wait_sec, :integer, default: 1),
|
68
|
+
incremental_column: incremental_column,
|
53
69
|
retry_limit: config.param(:retry_limit, :integer, default: 5),
|
54
|
-
latest_fetched_time:
|
70
|
+
latest_fetched_time: latest_fetched_time,
|
71
|
+
incremental: incremental
|
55
72
|
}
|
56
73
|
|
57
74
|
if task[:fetch_unknown_columns] && task[:fetch_custom_properties]
|
@@ -82,14 +99,16 @@ module Embulk
|
|
82
99
|
|
83
100
|
# NOTE: If this plugin supports to run by multi threads, this
|
84
101
|
# implementation is terrible.
|
85
|
-
|
86
|
-
|
87
|
-
|
88
|
-
|
89
|
-
|
90
|
-
|
91
|
-
|
92
|
-
|
102
|
+
if task[:incremental]
|
103
|
+
task_report = task_reports.first
|
104
|
+
next_to_date = Date.parse(task_report[:to_date])
|
105
|
+
next_config_diff = {
|
106
|
+
from_date: next_to_date.to_s,
|
107
|
+
latest_fetched_time: task_report[:latest_fetched_time],
|
108
|
+
}
|
109
|
+
return next_config_diff
|
110
|
+
end
|
111
|
+
return {}
|
93
112
|
end
|
94
113
|
|
95
114
|
def self.guess(config)
|
@@ -109,7 +128,6 @@ module Embulk
|
|
109
128
|
"from_date" => range.first,
|
110
129
|
"to_date" => range.last,
|
111
130
|
)
|
112
|
-
|
113
131
|
columns = guess_from_records(client.export_for_small_dataset(params))
|
114
132
|
return {"columns" => columns}
|
115
133
|
end
|
@@ -133,6 +151,8 @@ module Embulk
|
|
133
151
|
@schema = task[:schema]
|
134
152
|
@dates = task[:dates]
|
135
153
|
@fetch_unknown_columns = task[:fetch_unknown_columns]
|
154
|
+
@incremental_column = task[:incremental_column]
|
155
|
+
@incremental = task[:incremental]
|
136
156
|
end
|
137
157
|
|
138
158
|
def run
|
@@ -146,19 +166,25 @@ module Embulk
|
|
146
166
|
unless preview?
|
147
167
|
Embulk.logger.info "Fetching data from #{dates.first} to #{dates.last} ..."
|
148
168
|
end
|
149
|
-
|
150
|
-
fetch(dates).each do |record|
|
151
|
-
|
152
|
-
|
153
|
-
|
154
|
-
|
155
|
-
|
156
|
-
|
157
|
-
|
158
|
-
|
159
|
-
|
160
|
-
|
161
|
-
|
169
|
+
record_time_column=@incremental_column || DEFAULT_TIME_COLUMN
|
170
|
+
fetch(dates,prev_latest_fetched_time).each do |record|
|
171
|
+
if @incremental
|
172
|
+
if !record["properties"].include?(record_time_column)
|
173
|
+
raise Embulk::ConfigError.new("Incremental column not exists in fetched data #{record_time_column}")
|
174
|
+
end
|
175
|
+
record_time = record["properties"][record_time_column]
|
176
|
+
if @incremental_column.nil?
|
177
|
+
if record_time <= prev_latest_fetched_time
|
178
|
+
ignored_record_count += 1
|
179
|
+
next
|
180
|
+
end
|
181
|
+
end
|
182
|
+
|
183
|
+
current_latest_fetched_time= [
|
184
|
+
current_latest_fetched_time,
|
185
|
+
record_time,
|
186
|
+
].max
|
187
|
+
end
|
162
188
|
values = extract_values(record)
|
163
189
|
if @fetch_unknown_columns
|
164
190
|
unknown_values = extract_unknown_values(record)
|
@@ -175,14 +201,12 @@ module Embulk
|
|
175
201
|
end
|
176
202
|
break if preview?
|
177
203
|
end
|
178
|
-
|
179
204
|
page_builder.finish
|
180
|
-
|
181
205
|
task_report = {
|
182
206
|
latest_fetched_time: current_latest_fetched_time,
|
183
207
|
to_date: @dates.last || Date.today - 1,
|
184
208
|
}
|
185
|
-
|
209
|
+
task_report
|
186
210
|
end
|
187
211
|
|
188
212
|
private
|
@@ -236,13 +260,19 @@ module Embulk
|
|
236
260
|
end
|
237
261
|
end
|
238
262
|
|
239
|
-
def fetch(dates, &block)
|
263
|
+
def fetch(dates,last_fetch_time, &block)
|
240
264
|
from_date = dates.first
|
241
265
|
to_date = dates.last
|
242
266
|
params = @params.merge(
|
243
267
|
"from_date" => from_date,
|
244
|
-
"to_date" => to_date
|
268
|
+
"to_date" => to_date
|
245
269
|
)
|
270
|
+
if !@incremental_column.nil? && !last_fetch_time.nil? && last_fetch_time!=0 # can't do filter on time column, time column need to be filter manually.
|
271
|
+
params = params.merge(
|
272
|
+
"where" => "#{params['where'].nil? ? '' : "(#{params['where']}) and " }properties[\"#{@incremental_column}\"] > #{last_fetch_time}"
|
273
|
+
)
|
274
|
+
end
|
275
|
+
puts "Where params is #{params["where"]}"
|
246
276
|
client = MixpanelApi::Client.new(@api_key, @api_secret, self.class.perfect_retry(task))
|
247
277
|
|
248
278
|
if preview?
|
@@ -299,7 +329,7 @@ module Embulk
|
|
299
329
|
end
|
300
330
|
|
301
331
|
def self.guess_from_records(records)
|
302
|
-
sample_props = records.first(GUESS_RECORDS_COUNT).map{|r| r["properties"]}
|
332
|
+
sample_props = records.first(GUESS_RECORDS_COUNT).map {|r| r["properties"]}
|
303
333
|
schema = Guess::SchemaGuess.from_hash_records(sample_props)
|
304
334
|
columns = schema.map do |col|
|
305
335
|
next if col.name == "time"
|
@@ -311,6 +341,7 @@ module Embulk
|
|
311
341
|
result
|
312
342
|
end.compact
|
313
343
|
columns.unshift(name: NOT_PROPERTY_COLUMN, type: :string)
|
344
|
+
# Shift incremental column to top
|
314
345
|
columns.unshift(name: "time", type: :long)
|
315
346
|
end
|
316
347
|
end
|
data/lib/range_generator.rb
CHANGED
@@ -116,7 +116,7 @@ module Embulk
|
|
116
116
|
|
117
117
|
def test_json_type
|
118
118
|
sample_records = records.map do |r|
|
119
|
-
r.merge("properties" => {"array" => [1,2], "hash" => {foo: "FOO"}})
|
119
|
+
r.merge("properties" => {"time" => 1, "array" => [1, 2], "hash" => {foo: "FOO"}})
|
120
120
|
end
|
121
121
|
actual = Mixpanel.guess_from_records(sample_records)
|
122
122
|
assert actual.include?(name: "array", type: :json)
|
@@ -193,6 +193,14 @@ module Embulk
|
|
193
193
|
end
|
194
194
|
end
|
195
195
|
|
196
|
+
def test_default_configuration
|
197
|
+
stub(Mixpanel).resume {|task|
|
198
|
+
assert_nil(task[:incremental_column])
|
199
|
+
assert_true(task[:incremental])
|
200
|
+
}
|
201
|
+
Mixpanel.transaction(transaction_config(Date.today))
|
202
|
+
end
|
203
|
+
|
196
204
|
private
|
197
205
|
|
198
206
|
def transaction_config(from_date)
|
@@ -282,6 +290,8 @@ module Embulk
|
|
282
290
|
dates: DATES.map {|date| date.to_s},
|
283
291
|
api_key: API_KEY,
|
284
292
|
api_secret: API_SECRET,
|
293
|
+
incremental: true,
|
294
|
+
incremental_column: nil,
|
285
295
|
timezone: timezone,
|
286
296
|
schema: schema
|
287
297
|
)
|
@@ -304,6 +314,25 @@ module Embulk
|
|
304
314
|
Mixpanel.transaction(transaction_config(days), &control)
|
305
315
|
end
|
306
316
|
|
317
|
+
def test_valid_days_with_backfill
|
318
|
+
days = 5
|
319
|
+
|
320
|
+
stub(Mixpanel).resume() do |task|
|
321
|
+
assert_equal(["2015-02-17", "2015-02-18", "2015-02-19", "2015-02-20", "2015-02-21", "2015-02-22", "2015-02-23", "2015-02-24", "2015-02-25", "2015-02-26"], task[:dates])
|
322
|
+
end
|
323
|
+
config=transaction_config(days).merge("back_fill_days" => 5, "incremental_column" => "test_column", "latest_fetched_time" => 1501599491000)
|
324
|
+
Mixpanel.transaction(config, &control)
|
325
|
+
end
|
326
|
+
|
327
|
+
def test_valid_days_with_backfill_first_run
|
328
|
+
days = 5
|
329
|
+
stub(Mixpanel).resume() do |task|
|
330
|
+
assert_equal(transaction_task(days)[:dates], task[:dates])
|
331
|
+
end
|
332
|
+
config=transaction_config(days).merge("back_fill_days" => 5, "incremental_column" => "test_column")
|
333
|
+
Mixpanel.transaction(config, &control)
|
334
|
+
end
|
335
|
+
|
307
336
|
def test_invalid_days
|
308
337
|
days = 0
|
309
338
|
|
@@ -549,6 +578,96 @@ module Embulk
|
|
549
578
|
@plugin.run
|
550
579
|
end
|
551
580
|
|
581
|
+
class NonIncrementalRunTest < self
|
582
|
+
|
583
|
+
def test_non_incremental_run
|
584
|
+
|
585
|
+
mock(@page_builder).add(anything).times(records.length * 2)
|
586
|
+
mock(@page_builder).finish
|
587
|
+
task_report = @plugin.run
|
588
|
+
assert_equal(0, task_report[:latest_fetched_time])
|
589
|
+
end
|
590
|
+
|
591
|
+
def task
|
592
|
+
super.merge(incremental: false)
|
593
|
+
end
|
594
|
+
|
595
|
+
end
|
596
|
+
|
597
|
+
class IncrementalRunTest < self
|
598
|
+
|
599
|
+
def test_incremental_run
|
600
|
+
dont_allow(mock(@page_builder)).add(anything)
|
601
|
+
mock(@page_builder).finish
|
602
|
+
task_report = @plugin.run
|
603
|
+
assert_equal(record_epoch+1, task_report[:latest_fetched_time])
|
604
|
+
end
|
605
|
+
|
606
|
+
def task
|
607
|
+
super.merge(incremental: true, latest_fetched_time: record_epoch+1)
|
608
|
+
end
|
609
|
+
|
610
|
+
end
|
611
|
+
|
612
|
+
class IncrementalColumnTest < self
|
613
|
+
|
614
|
+
def setup
|
615
|
+
end
|
616
|
+
|
617
|
+
def test_incremental_column_with_where
|
618
|
+
page_builder = Object.new
|
619
|
+
plugin = Mixpanel.new(task.merge(params: task[:params].merge("where" => "abc==def")), nil, nil, page_builder)
|
620
|
+
stub(plugin).preview? {false}
|
621
|
+
adjusted = record_epoch - timezone_offset_seconds
|
622
|
+
mock(page_builder).add(["FOO", adjusted, "event"]).times(records.length * 2)
|
623
|
+
mock(page_builder).finish
|
624
|
+
any_instance_of(MixpanelApi::Client) do |klass|
|
625
|
+
stub(klass).export() do |params, block|
|
626
|
+
assert_equal('(abc==def) and properties["mp_processing_time_ms"] > 0',params["where"])
|
627
|
+
records.each{|record| block.call(record) }
|
628
|
+
end
|
629
|
+
end
|
630
|
+
task_report = plugin.run
|
631
|
+
assert_equal(1234567919, task_report[:latest_fetched_time])
|
632
|
+
end
|
633
|
+
|
634
|
+
def test_incremental_column
|
635
|
+
page_builder = Object.new
|
636
|
+
plugin = Mixpanel.new(task, nil, nil, page_builder)
|
637
|
+
stub(plugin).preview? {false}
|
638
|
+
adjusted = record_epoch - timezone_offset_seconds
|
639
|
+
mock(page_builder).add(["FOO", adjusted, "event"]).times(records.length * 2)
|
640
|
+
mock(page_builder).finish
|
641
|
+
any_instance_of(MixpanelApi::Client) do |klass|
|
642
|
+
stub(klass).export() do |params, block|
|
643
|
+
assert_equal('properties["mp_processing_time_ms"] > 0',params["where"])
|
644
|
+
records.each{|record| block.call(record) }
|
645
|
+
end
|
646
|
+
end
|
647
|
+
task_report = plugin.run
|
648
|
+
assert_equal(1234567919, task_report[:latest_fetched_time])
|
649
|
+
end
|
650
|
+
|
651
|
+
def records
|
652
|
+
super.each_with_index.map {|record, i|
|
653
|
+
record['properties']['mp_processing_time_ms'] = record_epoch+i
|
654
|
+
record
|
655
|
+
}
|
656
|
+
end
|
657
|
+
|
658
|
+
def schema
|
659
|
+
[
|
660
|
+
{"name" => "foo", "type" => "string"},
|
661
|
+
{"name" => "time", "type" => "integer"},
|
662
|
+
{"name" => "event", "type" => "string"},
|
663
|
+
]
|
664
|
+
end
|
665
|
+
|
666
|
+
def task
|
667
|
+
super.merge(incremental_column: 'mp_processing_time_ms')
|
668
|
+
end
|
669
|
+
end
|
670
|
+
|
552
671
|
class CustomPropertiesTest < self
|
553
672
|
def setup
|
554
673
|
super
|
@@ -671,6 +790,8 @@ module Embulk
|
|
671
790
|
api_key: API_KEY,
|
672
791
|
api_secret: API_SECRET,
|
673
792
|
timezone: TIMEZONE,
|
793
|
+
incremental: true,
|
794
|
+
incremental_column: nil,
|
674
795
|
schema: schema,
|
675
796
|
dates: DATES.to_a.map(&:to_s),
|
676
797
|
params: Mixpanel.export_params(embulk_config),
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: embulk-input-mixpanel
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.5.
|
4
|
+
version: 0.5.3.alpha.1
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- yoshihara
|
@@ -9,7 +9,7 @@ authors:
|
|
9
9
|
autorequire:
|
10
10
|
bindir: bin
|
11
11
|
cert_chain: []
|
12
|
-
date: 2017-
|
12
|
+
date: 2017-08-02 00:00:00.000000000 Z
|
13
13
|
dependencies:
|
14
14
|
- !ruby/object:Gem::Dependency
|
15
15
|
requirement: !ruby/object:Gem::Requirement
|
@@ -219,12 +219,12 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
219
219
|
version: '0'
|
220
220
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
221
221
|
requirements:
|
222
|
-
- - "
|
222
|
+
- - ">"
|
223
223
|
- !ruby/object:Gem::Version
|
224
|
-
version:
|
224
|
+
version: 1.3.1
|
225
225
|
requirements: []
|
226
226
|
rubyforge_project:
|
227
|
-
rubygems_version: 2.
|
227
|
+
rubygems_version: 2.6.12
|
228
228
|
signing_key:
|
229
229
|
specification_version: 4
|
230
230
|
summary: Mixpanel input plugin for Embulk
|