searchkick 4.4.0 → 5.3.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +160 -3
- data/LICENSE.txt +1 -1
- data/README.md +567 -421
- data/lib/searchkick/bulk_reindex_job.rb +12 -8
- data/lib/searchkick/controller_runtime.rb +40 -0
- data/lib/searchkick/index.rb +167 -74
- data/lib/searchkick/index_cache.rb +30 -0
- data/lib/searchkick/index_options.rb +465 -404
- data/lib/searchkick/indexer.rb +15 -8
- data/lib/searchkick/log_subscriber.rb +57 -0
- data/lib/searchkick/middleware.rb +9 -2
- data/lib/searchkick/model.rb +50 -51
- data/lib/searchkick/process_batch_job.rb +9 -25
- data/lib/searchkick/process_queue_job.rb +4 -3
- data/lib/searchkick/query.rb +106 -77
- data/lib/searchkick/record_data.rb +1 -1
- data/lib/searchkick/record_indexer.rb +136 -51
- data/lib/searchkick/reindex_queue.rb +51 -9
- data/lib/searchkick/reindex_v2_job.rb +10 -34
- data/lib/searchkick/relation.rb +247 -0
- data/lib/searchkick/relation_indexer.rb +155 -0
- data/lib/searchkick/results.rb +131 -96
- data/lib/searchkick/version.rb +1 -1
- data/lib/searchkick/where.rb +11 -0
- data/lib/searchkick.rb +202 -96
- data/lib/tasks/searchkick.rake +14 -10
- metadata +18 -85
- data/CONTRIBUTING.md +0 -53
- data/lib/searchkick/bulk_indexer.rb +0 -173
- data/lib/searchkick/logging.rb +0 -246
@@ -1,173 +0,0 @@
|
|
1
|
-
module Searchkick
|
2
|
-
class BulkIndexer
|
3
|
-
attr_reader :index
|
4
|
-
|
5
|
-
def initialize(index)
|
6
|
-
@index = index
|
7
|
-
end
|
8
|
-
|
9
|
-
def import_scope(relation, resume: false, method_name: nil, async: false, batch: false, batch_id: nil, full: false, scope: nil)
|
10
|
-
if scope
|
11
|
-
relation = relation.send(scope)
|
12
|
-
elsif relation.respond_to?(:search_import)
|
13
|
-
relation = relation.search_import
|
14
|
-
end
|
15
|
-
|
16
|
-
if batch
|
17
|
-
import_or_update relation.to_a, method_name, async
|
18
|
-
Searchkick.with_redis { |r| r.srem(batches_key, batch_id) } if batch_id
|
19
|
-
elsif full && async
|
20
|
-
full_reindex_async(relation)
|
21
|
-
elsif relation.respond_to?(:find_in_batches)
|
22
|
-
if resume
|
23
|
-
# use total docs instead of max id since there's not a great way
|
24
|
-
# to get the max _id without scripting since it's a string
|
25
|
-
|
26
|
-
# TODO use primary key and prefix with table name
|
27
|
-
relation = relation.where("id > ?", index.total_docs)
|
28
|
-
end
|
29
|
-
|
30
|
-
relation = relation.select("id").except(:includes, :preload) if async
|
31
|
-
|
32
|
-
relation.find_in_batches batch_size: batch_size do |items|
|
33
|
-
import_or_update items, method_name, async
|
34
|
-
end
|
35
|
-
else
|
36
|
-
each_batch(relation) do |items|
|
37
|
-
import_or_update items, method_name, async
|
38
|
-
end
|
39
|
-
end
|
40
|
-
end
|
41
|
-
|
42
|
-
def bulk_index(records)
|
43
|
-
Searchkick.indexer.queue(records.map { |r| RecordData.new(index, r).index_data })
|
44
|
-
end
|
45
|
-
|
46
|
-
def bulk_delete(records)
|
47
|
-
Searchkick.indexer.queue(records.reject { |r| r.id.blank? }.map { |r| RecordData.new(index, r).delete_data })
|
48
|
-
end
|
49
|
-
|
50
|
-
def bulk_update(records, method_name)
|
51
|
-
Searchkick.indexer.queue(records.map { |r| RecordData.new(index, r).update_data(method_name) })
|
52
|
-
end
|
53
|
-
|
54
|
-
def batches_left
|
55
|
-
Searchkick.with_redis { |r| r.scard(batches_key) }
|
56
|
-
end
|
57
|
-
|
58
|
-
private
|
59
|
-
|
60
|
-
def import_or_update(records, method_name, async)
|
61
|
-
if records.any?
|
62
|
-
if async
|
63
|
-
Searchkick::BulkReindexJob.perform_later(
|
64
|
-
class_name: records.first.class.searchkick_options[:class_name],
|
65
|
-
record_ids: records.map(&:id),
|
66
|
-
index_name: index.name,
|
67
|
-
method_name: method_name ? method_name.to_s : nil
|
68
|
-
)
|
69
|
-
else
|
70
|
-
records = records.select(&:should_index?)
|
71
|
-
if records.any?
|
72
|
-
with_retries do
|
73
|
-
# call out to index for ActiveSupport notifications
|
74
|
-
if method_name
|
75
|
-
index.bulk_update(records, method_name)
|
76
|
-
else
|
77
|
-
index.bulk_index(records)
|
78
|
-
end
|
79
|
-
end
|
80
|
-
end
|
81
|
-
end
|
82
|
-
end
|
83
|
-
end
|
84
|
-
|
85
|
-
def full_reindex_async(scope)
|
86
|
-
if scope.respond_to?(:primary_key)
|
87
|
-
# TODO expire Redis key
|
88
|
-
primary_key = scope.primary_key
|
89
|
-
|
90
|
-
scope = scope.select(primary_key).except(:includes, :preload)
|
91
|
-
|
92
|
-
starting_id =
|
93
|
-
begin
|
94
|
-
scope.minimum(primary_key)
|
95
|
-
rescue ActiveRecord::StatementInvalid
|
96
|
-
false
|
97
|
-
end
|
98
|
-
|
99
|
-
if starting_id.nil?
|
100
|
-
# no records, do nothing
|
101
|
-
elsif starting_id.is_a?(Numeric)
|
102
|
-
max_id = scope.maximum(primary_key)
|
103
|
-
batches_count = ((max_id - starting_id + 1) / batch_size.to_f).ceil
|
104
|
-
|
105
|
-
batches_count.times do |i|
|
106
|
-
batch_id = i + 1
|
107
|
-
min_id = starting_id + (i * batch_size)
|
108
|
-
bulk_reindex_job scope, batch_id, min_id: min_id, max_id: min_id + batch_size - 1
|
109
|
-
end
|
110
|
-
else
|
111
|
-
scope.find_in_batches(batch_size: batch_size).each_with_index do |batch, i|
|
112
|
-
batch_id = i + 1
|
113
|
-
|
114
|
-
bulk_reindex_job scope, batch_id, record_ids: batch.map { |record| record.id.to_s }
|
115
|
-
end
|
116
|
-
end
|
117
|
-
else
|
118
|
-
batch_id = 1
|
119
|
-
# TODO remove any eager loading
|
120
|
-
scope = scope.only(:_id) if scope.respond_to?(:only)
|
121
|
-
each_batch(scope) do |items|
|
122
|
-
bulk_reindex_job scope, batch_id, record_ids: items.map { |i| i.id.to_s }
|
123
|
-
batch_id += 1
|
124
|
-
end
|
125
|
-
end
|
126
|
-
end
|
127
|
-
|
128
|
-
def each_batch(scope)
|
129
|
-
# https://github.com/karmi/tire/blob/master/lib/tire/model/import.rb
|
130
|
-
# use cursor for Mongoid
|
131
|
-
items = []
|
132
|
-
scope.all.each do |item|
|
133
|
-
items << item
|
134
|
-
if items.length == batch_size
|
135
|
-
yield items
|
136
|
-
items = []
|
137
|
-
end
|
138
|
-
end
|
139
|
-
yield items if items.any?
|
140
|
-
end
|
141
|
-
|
142
|
-
def bulk_reindex_job(scope, batch_id, options)
|
143
|
-
Searchkick.with_redis { |r| r.sadd(batches_key, batch_id) }
|
144
|
-
Searchkick::BulkReindexJob.perform_later(**{
|
145
|
-
class_name: scope.searchkick_options[:class_name],
|
146
|
-
index_name: index.name,
|
147
|
-
batch_id: batch_id
|
148
|
-
}.merge(options))
|
149
|
-
end
|
150
|
-
|
151
|
-
def with_retries
|
152
|
-
retries = 0
|
153
|
-
|
154
|
-
begin
|
155
|
-
yield
|
156
|
-
rescue Faraday::ClientError => e
|
157
|
-
if retries < 1
|
158
|
-
retries += 1
|
159
|
-
retry
|
160
|
-
end
|
161
|
-
raise e
|
162
|
-
end
|
163
|
-
end
|
164
|
-
|
165
|
-
def batches_key
|
166
|
-
"searchkick:reindex:#{index.name}:batches"
|
167
|
-
end
|
168
|
-
|
169
|
-
def batch_size
|
170
|
-
@batch_size ||= index.options[:batch_size] || 1000
|
171
|
-
end
|
172
|
-
end
|
173
|
-
end
|
data/lib/searchkick/logging.rb
DELETED
@@ -1,246 +0,0 @@
|
|
1
|
-
# based on https://gist.github.com/mnutt/566725
|
2
|
-
require "active_support/core_ext/module/attr_internal"
|
3
|
-
|
4
|
-
module Searchkick
|
5
|
-
module QueryWithInstrumentation
|
6
|
-
def execute_search
|
7
|
-
name = searchkick_klass ? "#{searchkick_klass.name} Search" : "Search"
|
8
|
-
event = {
|
9
|
-
name: name,
|
10
|
-
query: params
|
11
|
-
}
|
12
|
-
ActiveSupport::Notifications.instrument("search.searchkick", event) do
|
13
|
-
super
|
14
|
-
end
|
15
|
-
end
|
16
|
-
end
|
17
|
-
|
18
|
-
module IndexWithInstrumentation
|
19
|
-
def store(record)
|
20
|
-
event = {
|
21
|
-
name: "#{record.searchkick_klass.name} Store",
|
22
|
-
id: search_id(record)
|
23
|
-
}
|
24
|
-
if Searchkick.callbacks_value == :bulk
|
25
|
-
super
|
26
|
-
else
|
27
|
-
ActiveSupport::Notifications.instrument("request.searchkick", event) do
|
28
|
-
super
|
29
|
-
end
|
30
|
-
end
|
31
|
-
end
|
32
|
-
|
33
|
-
def remove(record)
|
34
|
-
name = record && record.searchkick_klass ? "#{record.searchkick_klass.name} Remove" : "Remove"
|
35
|
-
event = {
|
36
|
-
name: name,
|
37
|
-
id: search_id(record)
|
38
|
-
}
|
39
|
-
if Searchkick.callbacks_value == :bulk
|
40
|
-
super
|
41
|
-
else
|
42
|
-
ActiveSupport::Notifications.instrument("request.searchkick", event) do
|
43
|
-
super
|
44
|
-
end
|
45
|
-
end
|
46
|
-
end
|
47
|
-
|
48
|
-
def update_record(record, method_name)
|
49
|
-
event = {
|
50
|
-
name: "#{record.searchkick_klass.name} Update",
|
51
|
-
id: search_id(record)
|
52
|
-
}
|
53
|
-
if Searchkick.callbacks_value == :bulk
|
54
|
-
super
|
55
|
-
else
|
56
|
-
ActiveSupport::Notifications.instrument("request.searchkick", event) do
|
57
|
-
super
|
58
|
-
end
|
59
|
-
end
|
60
|
-
end
|
61
|
-
|
62
|
-
def bulk_index(records)
|
63
|
-
if records.any?
|
64
|
-
event = {
|
65
|
-
name: "#{records.first.searchkick_klass.name} Import",
|
66
|
-
count: records.size
|
67
|
-
}
|
68
|
-
event[:id] = search_id(records.first) if records.size == 1
|
69
|
-
if Searchkick.callbacks_value == :bulk
|
70
|
-
super
|
71
|
-
else
|
72
|
-
ActiveSupport::Notifications.instrument("request.searchkick", event) do
|
73
|
-
super
|
74
|
-
end
|
75
|
-
end
|
76
|
-
end
|
77
|
-
end
|
78
|
-
alias_method :import, :bulk_index
|
79
|
-
|
80
|
-
def bulk_update(records, *args)
|
81
|
-
if records.any?
|
82
|
-
event = {
|
83
|
-
name: "#{records.first.searchkick_klass.name} Update",
|
84
|
-
count: records.size
|
85
|
-
}
|
86
|
-
event[:id] = search_id(records.first) if records.size == 1
|
87
|
-
if Searchkick.callbacks_value == :bulk
|
88
|
-
super
|
89
|
-
else
|
90
|
-
ActiveSupport::Notifications.instrument("request.searchkick", event) do
|
91
|
-
super
|
92
|
-
end
|
93
|
-
end
|
94
|
-
end
|
95
|
-
end
|
96
|
-
|
97
|
-
def bulk_delete(records)
|
98
|
-
if records.any?
|
99
|
-
event = {
|
100
|
-
name: "#{records.first.searchkick_klass.name} Delete",
|
101
|
-
count: records.size
|
102
|
-
}
|
103
|
-
event[:id] = search_id(records.first) if records.size == 1
|
104
|
-
if Searchkick.callbacks_value == :bulk
|
105
|
-
super
|
106
|
-
else
|
107
|
-
ActiveSupport::Notifications.instrument("request.searchkick", event) do
|
108
|
-
super
|
109
|
-
end
|
110
|
-
end
|
111
|
-
end
|
112
|
-
end
|
113
|
-
end
|
114
|
-
|
115
|
-
module IndexerWithInstrumentation
|
116
|
-
def perform
|
117
|
-
if Searchkick.callbacks_value == :bulk
|
118
|
-
event = {
|
119
|
-
name: "Bulk",
|
120
|
-
count: queued_items.size
|
121
|
-
}
|
122
|
-
ActiveSupport::Notifications.instrument("request.searchkick", event) do
|
123
|
-
super
|
124
|
-
end
|
125
|
-
else
|
126
|
-
super
|
127
|
-
end
|
128
|
-
end
|
129
|
-
end
|
130
|
-
|
131
|
-
module SearchkickWithInstrumentation
|
132
|
-
def multi_search(searches)
|
133
|
-
event = {
|
134
|
-
name: "Multi Search",
|
135
|
-
body: searches.flat_map { |q| [q.params.except(:body).to_json, q.body.to_json] }.map { |v| "#{v}\n" }.join,
|
136
|
-
}
|
137
|
-
ActiveSupport::Notifications.instrument("multi_search.searchkick", event) do
|
138
|
-
super
|
139
|
-
end
|
140
|
-
end
|
141
|
-
end
|
142
|
-
|
143
|
-
# https://github.com/rails/rails/blob/master/activerecord/lib/active_record/log_subscriber.rb
|
144
|
-
class LogSubscriber < ActiveSupport::LogSubscriber
|
145
|
-
def self.runtime=(value)
|
146
|
-
Thread.current[:searchkick_runtime] = value
|
147
|
-
end
|
148
|
-
|
149
|
-
def self.runtime
|
150
|
-
Thread.current[:searchkick_runtime] ||= 0
|
151
|
-
end
|
152
|
-
|
153
|
-
def self.reset_runtime
|
154
|
-
rt = runtime
|
155
|
-
self.runtime = 0
|
156
|
-
rt
|
157
|
-
end
|
158
|
-
|
159
|
-
def search(event)
|
160
|
-
self.class.runtime += event.duration
|
161
|
-
return unless logger.debug?
|
162
|
-
|
163
|
-
payload = event.payload
|
164
|
-
name = "#{payload[:name]} (#{event.duration.round(1)}ms)"
|
165
|
-
|
166
|
-
index = payload[:query][:index].is_a?(Array) ? payload[:query][:index].join(",") : payload[:query][:index]
|
167
|
-
type = payload[:query][:type]
|
168
|
-
request_params = payload[:query].except(:index, :type, :body)
|
169
|
-
|
170
|
-
params = []
|
171
|
-
request_params.each do |k, v|
|
172
|
-
params << "#{CGI.escape(k.to_s)}=#{CGI.escape(v.to_s)}"
|
173
|
-
end
|
174
|
-
|
175
|
-
debug " #{color(name, YELLOW, true)} #{index}#{type ? "/#{type.join(',')}" : ''}/_search#{params.any? ? '?' + params.join('&') : nil} #{payload[:query][:body].to_json}"
|
176
|
-
end
|
177
|
-
|
178
|
-
def request(event)
|
179
|
-
self.class.runtime += event.duration
|
180
|
-
return unless logger.debug?
|
181
|
-
|
182
|
-
payload = event.payload
|
183
|
-
name = "#{payload[:name]} (#{event.duration.round(1)}ms)"
|
184
|
-
|
185
|
-
debug " #{color(name, YELLOW, true)} #{payload.except(:name).to_json}"
|
186
|
-
end
|
187
|
-
|
188
|
-
def multi_search(event)
|
189
|
-
self.class.runtime += event.duration
|
190
|
-
return unless logger.debug?
|
191
|
-
|
192
|
-
payload = event.payload
|
193
|
-
name = "#{payload[:name]} (#{event.duration.round(1)}ms)"
|
194
|
-
|
195
|
-
debug " #{color(name, YELLOW, true)} _msearch #{payload[:body]}"
|
196
|
-
end
|
197
|
-
end
|
198
|
-
|
199
|
-
# https://github.com/rails/rails/blob/master/activerecord/lib/active_record/railties/controller_runtime.rb
|
200
|
-
module ControllerRuntime
|
201
|
-
extend ActiveSupport::Concern
|
202
|
-
|
203
|
-
protected
|
204
|
-
|
205
|
-
attr_internal :searchkick_runtime
|
206
|
-
|
207
|
-
def process_action(action, *args)
|
208
|
-
# We also need to reset the runtime before each action
|
209
|
-
# because of queries in middleware or in cases we are streaming
|
210
|
-
# and it won't be cleaned up by the method below.
|
211
|
-
Searchkick::LogSubscriber.reset_runtime
|
212
|
-
super
|
213
|
-
end
|
214
|
-
|
215
|
-
def cleanup_view_runtime
|
216
|
-
searchkick_rt_before_render = Searchkick::LogSubscriber.reset_runtime
|
217
|
-
runtime = super
|
218
|
-
searchkick_rt_after_render = Searchkick::LogSubscriber.reset_runtime
|
219
|
-
self.searchkick_runtime = searchkick_rt_before_render + searchkick_rt_after_render
|
220
|
-
runtime - searchkick_rt_after_render
|
221
|
-
end
|
222
|
-
|
223
|
-
def append_info_to_payload(payload)
|
224
|
-
super
|
225
|
-
payload[:searchkick_runtime] = (searchkick_runtime || 0) + Searchkick::LogSubscriber.reset_runtime
|
226
|
-
end
|
227
|
-
|
228
|
-
module ClassMethods
|
229
|
-
def log_process_action(payload)
|
230
|
-
messages = super
|
231
|
-
runtime = payload[:searchkick_runtime]
|
232
|
-
messages << ("Searchkick: %.1fms" % runtime.to_f) if runtime.to_f > 0
|
233
|
-
messages
|
234
|
-
end
|
235
|
-
end
|
236
|
-
end
|
237
|
-
end
|
238
|
-
|
239
|
-
Searchkick::Query.prepend(Searchkick::QueryWithInstrumentation)
|
240
|
-
Searchkick::Index.prepend(Searchkick::IndexWithInstrumentation)
|
241
|
-
Searchkick::Indexer.prepend(Searchkick::IndexerWithInstrumentation)
|
242
|
-
Searchkick.singleton_class.prepend(Searchkick::SearchkickWithInstrumentation)
|
243
|
-
Searchkick::LogSubscriber.attach_to :searchkick
|
244
|
-
ActiveSupport.on_load(:action_controller) do
|
245
|
-
include Searchkick::ControllerRuntime
|
246
|
-
end
|