reindex.rb
12.1 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
module Searchkick
module Reindex
# https://gist.github.com/jarosan/3124884
# http://www.elasticsearch.org/blog/changing-mapping-with-zero-downtime/
def reindex(options = {})
skip_import = options[:import] == false
alias_name = searchkick_index.name
new_name = "#{alias_name}_#{Time.now.strftime('%Y%m%d%H%M%S%L')}"
index = Searchkick::Index.new(new_name)
clean_indices
index.create searchkick_index_options
# check if alias exists
if Searchkick.client.indices.exists_alias(name: alias_name)
# import before swap
searchkick_import(index: index) unless skip_import
# get existing indices to remove
old_indices = Searchkick.client.indices.get_alias(name: alias_name).keys
actions = old_indices.map{|name| {remove: {index: name, alias: alias_name}} } + [{add: {index: new_name, alias: alias_name}}]
Searchkick.client.indices.update_aliases body: {actions: actions}
clean_indices
else
searchkick_index.delete if searchkick_index.exists?
Searchkick.client.indices.update_aliases body: {actions: [{add: {index: new_name, alias: alias_name}}]}
# import after swap
searchkick_import(index: index) unless skip_import
end
index.refresh
true
end
# remove old indices that start w/ index_name
def clean_indices
all_indices = Searchkick.client.indices.get_aliases
indices = all_indices.select{|k, v| v["aliases"].empty? && k =~ /\A#{Regexp.escape(searchkick_index.name)}_\d{14,17}\z/ }.keys
indices.each do |index|
Searchkick::Index.new(index).delete
end
indices
end
def self.extended(klass)
@descendents ||= []
@descendents << klass unless @descendents.include?(klass)
end
def searchkick_import(options = {})
index = options[:index] || searchkick_index
batch_size = searchkick_options[:batch_size] || 1000
# use scope for import
scope = searchkick_klass
scope = scope.search_import if scope.respond_to?(:search_import)
if scope.respond_to?(:find_in_batches)
scope.find_in_batches batch_size: batch_size do |batch|
index.import batch.select{|item| item.should_index? }
end
else
# https://github.com/karmi/tire/blob/master/lib/tire/model/import.rb
# use cursor for Mongoid
items = []
scope.all.each do |item|
items << item if item.should_index?
if items.length == batch_size
index.import items
items = []
end
end
index.import items
end
end
def searchkick_index_options
options = searchkick_options
if options[:mappings] and !options[:merge_mappings]
settings = options[:settings] || {}
mappings = options[:mappings]
else
settings = {
analysis: {
analyzer: {
searchkick_keyword: {
type: "custom",
tokenizer: "keyword",
filter: ["lowercase", "searchkick_stemmer"]
},
default_index: {
type: "custom",
tokenizer: "standard",
# synonym should come last, after stemming and shingle
# shingle must come before searchkick_stemmer
filter: ["standard", "lowercase", "asciifolding", "searchkick_index_shingle", "searchkick_stemmer"]
},
searchkick_search: {
type: "custom",
tokenizer: "standard",
filter: ["standard", "lowercase", "asciifolding", "searchkick_search_shingle", "searchkick_stemmer"]
},
searchkick_search2: {
type: "custom",
tokenizer: "standard",
filter: ["standard", "lowercase", "asciifolding", "searchkick_stemmer"]
},
# https://github.com/leschenko/elasticsearch_autocomplete/blob/master/lib/elasticsearch_autocomplete/analyzers.rb
searchkick_autocomplete_index: {
type: "custom",
tokenizer: "searchkick_autocomplete_ngram",
filter: ["lowercase", "asciifolding"]
},
searchkick_autocomplete_search: {
type: "custom",
tokenizer: "keyword",
filter: ["lowercase", "asciifolding"]
},
searchkick_word_search: {
type: "custom",
tokenizer: "standard",
filter: ["lowercase", "asciifolding"]
},
searchkick_suggest_index: {
type: "custom",
tokenizer: "standard",
filter: ["lowercase", "asciifolding", "searchkick_suggest_shingle"]
},
searchkick_text_start_index: {
type: "custom",
tokenizer: "keyword",
filter: ["lowercase", "asciifolding", "searchkick_edge_ngram"]
},
searchkick_text_middle_index: {
type: "custom",
tokenizer: "keyword",
filter: ["lowercase", "asciifolding", "searchkick_ngram"]
},
searchkick_text_end_index: {
type: "custom",
tokenizer: "keyword",
filter: ["lowercase", "asciifolding", "reverse", "searchkick_edge_ngram", "reverse"]
},
searchkick_word_start_index: {
type: "custom",
tokenizer: "standard",
filter: ["lowercase", "asciifolding", "searchkick_edge_ngram"]
},
searchkick_word_middle_index: {
type: "custom",
tokenizer: "standard",
filter: ["lowercase", "asciifolding", "searchkick_ngram"]
},
searchkick_word_end_index: {
type: "custom",
tokenizer: "standard",
filter: ["lowercase", "asciifolding", "reverse", "searchkick_edge_ngram", "reverse"]
}
},
filter: {
searchkick_index_shingle: {
type: "shingle",
token_separator: ""
},
# lucky find http://web.archiveorange.com/archive/v/AAfXfQ17f57FcRINsof7
searchkick_search_shingle: {
type: "shingle",
token_separator: "",
output_unigrams: false,
output_unigrams_if_no_shingles: true
},
searchkick_suggest_shingle: {
type: "shingle",
max_shingle_size: 5
},
searchkick_edge_ngram: {
type: "edgeNGram",
min_gram: 1,
max_gram: 50
},
searchkick_ngram: {
type: "nGram",
min_gram: 1,
max_gram: 50
},
searchkick_stemmer: {
type: "snowball",
language: options[:language] || "English"
}
},
tokenizer: {
searchkick_autocomplete_ngram: {
type: "edgeNGram",
min_gram: 1,
max_gram: 50
}
}
}
}
if Searchkick.env == "test"
settings.merge!(number_of_shards: 1, number_of_replicas: 0)
end
settings.deep_merge!(options[:settings] || {})
# synonyms
synonyms = options[:synonyms] || []
if synonyms.any?
settings[:analysis][:filter][:searchkick_synonym] = {
type: "synonym",
synonyms: synonyms.select{|s| s.size > 1 }.map{|s| s.join(",") }
}
# choosing a place for the synonym filter when stemming is not easy
# https://groups.google.com/forum/#!topic/elasticsearch/p7qcQlgHdB8
# TODO use a snowball stemmer on synonyms when creating the token filter
# http://elasticsearch-users.115913.n3.nabble.com/synonym-multi-words-search-td4030811.html
# I find the following approach effective if you are doing multi-word synonyms (synonym phrases):
# - Only apply the synonym expansion at index time
# - Don't have the synonym filter applied search
# - Use directional synonyms where appropriate. You want to make sure that you're not injecting terms that are too general.
settings[:analysis][:analyzer][:default_index][:filter].insert(4, "searchkick_synonym")
settings[:analysis][:analyzer][:default_index][:filter] << "searchkick_synonym"
end
if options[:wordnet]
settings[:analysis][:filter][:searchkick_wordnet] = {
type: "synonym",
format: "wordnet",
synonyms_path: Searchkick.wordnet_path
}
settings[:analysis][:analyzer][:default_index][:filter].insert(4, "searchkick_wordnet")
settings[:analysis][:analyzer][:default_index][:filter] << "searchkick_wordnet"
end
if options[:special_characters] == false
settings[:analysis][:analyzer].each do |analyzer, analyzer_settings|
analyzer_settings[:filter].reject!{|f| f == "asciifolding" }
end
end
mapping = {}
# conversions
if options[:conversions]
mapping[:conversions] = {
type: "nested",
properties: {
query: {type: "string", analyzer: "searchkick_keyword"},
count: {type: "integer"}
}
}
end
mapping_options = Hash[
[:autocomplete, :suggest, :text_start, :text_middle, :text_end, :word_start, :word_middle, :word_end, :highlight]
.map{|type| [type, (options[type] || []).map(&:to_s)] }
]
mapping_options.values.flatten.uniq.each do |field|
field_mapping = {
type: "multi_field",
fields: {
field => {type: "string", index: "not_analyzed"},
"analyzed" => {type: "string", index: "analyzed"}
# term_vector: "with_positions_offsets" for fast / correct highlighting
# http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-request-highlighting.html#_fast_vector_highlighter
}
}
mapping_options.except(:highlight).each do |type, fields|
if fields.include?(field)
field_mapping[:fields][type] = {type: "string", index: "analyzed", analyzer: "searchkick_#{type}_index"}
end
end
if mapping_options[:highlight].include?(field)
field_mapping[:fields]["analyzed"][:term_vector] = "with_positions_offsets"
end
mapping[field] = field_mapping
end
(options[:locations] || []).each do |field|
mapping[field] = {
type: "geo_point"
}
end
mappings = {
_default_: {
properties: mapping,
# https://gist.github.com/kimchy/2898285
dynamic_templates: [
{
string_template: {
match: "*",
match_mapping_type: "string",
mapping: {
# http://www.elasticsearch.org/guide/reference/mapping/multi-field-type/
type: "multi_field",
fields: {
# analyzed field must be the default field for include_in_all
# http://www.elasticsearch.org/guide/reference/mapping/multi-field-type/
# however, we can include the not_analyzed field in _all
# and the _all index analyzer will take care of it
"{name}" => {type: "string", index: "not_analyzed"},
"analyzed" => {type: "string", index: "analyzed"}
}
}
}
}
]
}
}.deep_merge(options[:mappings] || {})
end
{
settings: settings,
mappings: mappings
}
end
end
end