picky 3.0.0.pre1 → 3.0.0.pre2
Sign up to get free protection for your applications and to get access to all the features.
- data/lib/picky/adapters/rack/search.rb +3 -3
- data/lib/picky/adapters/rack.rb +3 -3
- data/lib/picky/application.rb +5 -3
- data/lib/picky/backend/file/json.rb +7 -0
- data/lib/picky/backend/file/text.rb +2 -2
- data/lib/picky/character_substituters/west_european.rb +3 -0
- data/lib/picky/extensions/symbol.rb +4 -4
- data/lib/picky/indexes/index.rb +11 -85
- data/lib/picky/indexes/index_indexed.rb +13 -2
- data/lib/picky/indexes/index_indexing.rb +38 -11
- data/lib/picky/loader.rb +0 -4
- data/lib/picky/search.rb +71 -78
- data/lib/picky/tokenizers/base.rb +4 -2
- data/lib/picky.rb +6 -0
- data/spec/lib/application_spec.rb +31 -11
- data/spec/lib/backend/files_spec.rb +1 -1
- data/spec/lib/categories_indexed_spec.rb +2 -2
- data/spec/lib/category_indexed_spec.rb +7 -2
- data/spec/lib/category_indexing_spec.rb +1 -1
- data/spec/lib/category_spec.rb +1 -1
- data/spec/lib/frontend_adapters/rack_spec.rb +4 -4
- data/spec/lib/indexed/bundle/memory_spec.rb +2 -2
- data/spec/lib/indexed/bundle/redis_spec.rb +1 -1
- data/spec/lib/indexed/wrappers/exact_first_spec.rb +2 -2
- data/spec/lib/indexes/index_indexed_spec.rb +25 -3
- data/spec/lib/indexes/index_indexing_spec.rb +58 -10
- data/spec/lib/indexes/index_spec.rb +13 -7
- data/spec/lib/indexes/redis_spec.rb +1 -1
- data/spec/lib/indexing/bundle/memory_partial_generation_speed_spec.rb +1 -1
- data/spec/lib/indexing/bundle/memory_spec.rb +1 -1
- data/spec/lib/indexing/bundle/redis_spec.rb +1 -1
- data/spec/lib/search_spec.rb +18 -18
- data/spec/lib/tokenizers/base_spec.rb +3 -3
- metadata +3 -6
- data/lib/picky/loggers/search.rb +0 -26
- data/spec/lib/loggers/search_spec.rb +0 -19
@@ -7,7 +7,7 @@ module Picky
|
|
7
7
|
# adapter for it.
|
8
8
|
#
|
9
9
|
# For example, if you give it a query, it will extract the query param etc.
|
10
|
-
# and call
|
10
|
+
# and call search on it if it is called by Rack.
|
11
11
|
#
|
12
12
|
module Rack
|
13
13
|
|
@@ -35,9 +35,9 @@ module Picky
|
|
35
35
|
lambda do |env|
|
36
36
|
params = ::Rack::Request.new(env).params
|
37
37
|
|
38
|
-
results = query.
|
38
|
+
results = query.search *extracted(params)
|
39
39
|
|
40
|
-
|
40
|
+
Picky.logger && Picky.logger.info(results.to_log(params[query_key]))
|
41
41
|
|
42
42
|
respond_with results.to_response, content_type
|
43
43
|
end
|
data/lib/picky/adapters/rack.rb
CHANGED
@@ -8,7 +8,7 @@ module Picky
|
|
8
8
|
# adapter for it.
|
9
9
|
#
|
10
10
|
# For example, if you give it a query, it will extract the query param etc.
|
11
|
-
# and call
|
11
|
+
# and call search on it if it is called by Rack.
|
12
12
|
#
|
13
13
|
# Usage:
|
14
14
|
# Adapters::Rack.app_for(thing, options)
|
@@ -18,8 +18,8 @@ module Picky
|
|
18
18
|
# Generates the appropriate app for Rack.
|
19
19
|
#
|
20
20
|
@@mapping = {
|
21
|
-
:
|
22
|
-
:parameters
|
21
|
+
:search => Search,
|
22
|
+
:parameters => LiveParameters
|
23
23
|
}
|
24
24
|
def self.app_for thing, options = {}
|
25
25
|
@@mapping.each_pair do |method, adapter|
|
data/lib/picky/application.rb
CHANGED
@@ -175,7 +175,10 @@ module Picky
|
|
175
175
|
|
176
176
|
# Routes.
|
177
177
|
#
|
178
|
-
|
178
|
+
def route options
|
179
|
+
raise "Warning: block passed into #route method, not into Search.new!" if block_given?
|
180
|
+
rack_adapter.route options
|
181
|
+
end
|
179
182
|
|
180
183
|
# A Picky application implements the Rack interface.
|
181
184
|
#
|
@@ -194,8 +197,7 @@ module Picky
|
|
194
197
|
# Reloads & finalizes the apps.
|
195
198
|
#
|
196
199
|
def reload
|
197
|
-
Loader.load_user 'app' # Sinatra app_file.
|
198
|
-
Loader.load_user 'app/logging' # Standard Picky logging.
|
200
|
+
Loader.load_user 'app' # Sinatra app_file. TODO Needed?
|
199
201
|
Loader.load_user 'app/application' # Standard Picky appfile.
|
200
202
|
finalize_apps
|
201
203
|
exclaim "Application #{apps.map(&:name).join(', ')} loaded."
|
@@ -17,6 +17,13 @@ module Picky
|
|
17
17
|
#
|
18
18
|
def load
|
19
19
|
Yajl::Parser.parse ::File.open(cache_path, 'r'), symbolize_keys: true
|
20
|
+
|
21
|
+
# Note: Circumvents the yajl symbolize utf-8 characters problem.
|
22
|
+
#
|
23
|
+
# Yajl::Parser.parse(::File.open(cache_path, 'r')).inject({}) do |hash, (k, v)|
|
24
|
+
# hash[k.to_sym] = v
|
25
|
+
# hash
|
26
|
+
# end
|
20
27
|
end
|
21
28
|
# Dumps the index hash in json format.
|
22
29
|
#
|
@@ -36,7 +36,7 @@ module Picky
|
|
36
36
|
def retrieve
|
37
37
|
id = nil
|
38
38
|
token = nil
|
39
|
-
::File.open(cache_path, 'r:
|
39
|
+
::File.open(cache_path, 'r:utf-8') do |file|
|
40
40
|
file.each_line do |line|
|
41
41
|
id, token = line.split ?,, 2
|
42
42
|
yield id, (token.chomp! || token).to_sym
|
@@ -47,7 +47,7 @@ module Picky
|
|
47
47
|
#
|
48
48
|
#
|
49
49
|
def open &block
|
50
|
-
::File.open cache_path, 'w:
|
50
|
+
::File.open cache_path, 'w:utf-8', &block
|
51
51
|
end
|
52
52
|
|
53
53
|
|
@@ -1,20 +1,20 @@
|
|
1
1
|
# Extending the Symbol class.
|
2
2
|
#
|
3
3
|
class Symbol # :nodoc:all
|
4
|
-
|
4
|
+
|
5
5
|
# :keys.each_subtoken # => yields each of [:keys, :key, :ke, :k]
|
6
6
|
# :keys.each_subtoken(2) # => yields each of [:keys, :key, :ke]
|
7
7
|
#
|
8
8
|
def each_subtoken from_length = 1
|
9
9
|
sub = self.id2name
|
10
|
-
|
10
|
+
|
11
11
|
size = sub.size
|
12
12
|
from_length = size + from_length + 1 if from_length < 0
|
13
13
|
from_length = size if size < from_length
|
14
14
|
from_length = 1 if from_length < 1
|
15
|
-
|
15
|
+
|
16
16
|
yield self
|
17
17
|
size.downto(from_length + 1) { yield sub.chop!.intern }
|
18
18
|
end
|
19
|
-
|
19
|
+
|
20
20
|
end
|
data/lib/picky/indexes/index.rb
CHANGED
@@ -99,41 +99,25 @@ module Picky
|
|
99
99
|
# === Parameters
|
100
100
|
# * name: A name that will be used for the index directory and in the Picky front end.
|
101
101
|
#
|
102
|
-
# === Options
|
102
|
+
# === Options (all are used in the block, see examples)
|
103
103
|
# * source: Where the data comes from, e.g. Sources::CSV.new(...). Optional, can be defined in the block using #source.
|
104
104
|
# * result_identifier: Use if you'd like a different identifier/name in the results than the name of the index.
|
105
105
|
# * after_indexing: As of this writing only used in the db source. Executes the given after_indexing as SQL after the indexing process.
|
106
|
-
# * tokenizer:
|
107
|
-
# * key_format:
|
108
|
-
#
|
109
|
-
# Examples:
|
110
|
-
# my_index = Indexes::Memory.new(:my_index, source: some_source) do
|
111
|
-
# category :bla
|
112
|
-
# end
|
106
|
+
# * tokenizer: Call and pass either a tokenizer (responds to #tokenize) or the options for a tokenizer..
|
107
|
+
# * key_format: Call and pass in a format method for the ids (default is #to_i).
|
113
108
|
#
|
109
|
+
# Example:
|
114
110
|
# my_index = Indexes::Memory.new(:my_index) do
|
115
|
-
# source
|
116
|
-
#
|
111
|
+
# source Sources::CSV.new(file: 'data/index.csv')
|
112
|
+
# key_format :to_sym
|
113
|
+
# category :bla
|
114
|
+
# result_identifier :my_special_results
|
117
115
|
# end
|
118
116
|
#
|
119
|
-
#
|
120
117
|
def initialize name, options = {}
|
121
|
-
|
122
|
-
@name = name.to_sym
|
123
|
-
|
124
|
-
check_options options
|
125
|
-
|
126
|
-
@source = options[:source]
|
127
|
-
|
128
|
-
@after_indexing = options[:after_indexing]
|
129
|
-
@tokenizer = options[:tokenizer]
|
130
|
-
@key_format = options[:key_format]
|
131
|
-
|
132
|
-
# Indexed.
|
133
|
-
#
|
134
|
-
@result_identifier = options[:result_identifier] || name
|
118
|
+
@name = name.to_sym
|
135
119
|
|
136
|
-
# TODO Move ignore_unassigned_tokens to query, somehow.
|
120
|
+
# TODO Move ignore_unassigned_tokens to query, somehow. Then, remove options.
|
137
121
|
#
|
138
122
|
@categories = Categories.new ignore_unassigned_tokens: (options[:ignore_unassigned_tokens] || false)
|
139
123
|
|
@@ -141,13 +125,7 @@ module Picky
|
|
141
125
|
#
|
142
126
|
Indexes.register self
|
143
127
|
|
144
|
-
#
|
145
|
-
#
|
146
128
|
instance_eval(&Proc.new) if block_given?
|
147
|
-
|
148
|
-
# Check if any source has been given in the block or the options.
|
149
|
-
#
|
150
|
-
check_source @source
|
151
129
|
end
|
152
130
|
|
153
131
|
# Default bundles.
|
@@ -319,58 +297,6 @@ module Picky
|
|
319
297
|
end
|
320
298
|
alias define_geo_categories geo_categories
|
321
299
|
|
322
|
-
#
|
323
|
-
# Since this is an API, we fail hard quickly.
|
324
|
-
#
|
325
|
-
def check_name name # :nodoc:
|
326
|
-
raise ArgumentError.new(<<-NAME
|
327
|
-
|
328
|
-
|
329
|
-
The index identifier (you gave "#{name}") for Indexes::Memory/Indexes::Redis should be a Symbol/String,
|
330
|
-
Examples:
|
331
|
-
Indexes::Memory.new(:my_cool_index) # Recommended
|
332
|
-
Indexes::Redis.new("a-redis-index")
|
333
|
-
NAME
|
334
|
-
|
335
|
-
|
336
|
-
) unless name.respond_to?(:to_sym)
|
337
|
-
end
|
338
|
-
def check_options options # :nodoc:
|
339
|
-
raise ArgumentError.new(<<-OPTIONS
|
340
|
-
|
341
|
-
|
342
|
-
Sources are not passed in as second parameter for #{self.class.name} anymore, but either
|
343
|
-
* as :source option:
|
344
|
-
#{self.class.name}.new(#{name.inspect}, source: #{options})
|
345
|
-
or
|
346
|
-
* given to the #source method inside the config block:
|
347
|
-
#{self.class.name}.new(#{name.inspect}) do
|
348
|
-
source #{options}
|
349
|
-
end
|
350
|
-
|
351
|
-
Sorry about that breaking change (in 2.2.0), didn't want to go to 3.0.0 yet!
|
352
|
-
|
353
|
-
All the best
|
354
|
-
-- Picky
|
355
|
-
|
356
|
-
|
357
|
-
OPTIONS
|
358
|
-
) unless options.respond_to?(:[])
|
359
|
-
end
|
360
|
-
def check_source source # :nodoc:
|
361
|
-
raise ArgumentError.new(<<-SOURCE
|
362
|
-
|
363
|
-
|
364
|
-
The index "#{name}" should use a data source that responds to either the method #each, or the method #harvest, which yields(id, text), OR it can be a lambda/block, returning such a source.
|
365
|
-
Or it could use one of the built-in sources:
|
366
|
-
Sources::#{(Sources.constants - [:Base, :Wrappers, :NoCSVFileGiven, :NoCouchDBGiven]).join(',
|
367
|
-
Sources::')}
|
368
|
-
|
369
|
-
|
370
|
-
SOURCE
|
371
|
-
) unless source.respond_to?(:each) || source.respond_to?(:harvest) || source.respond_to?(:call)
|
372
|
-
end
|
373
|
-
|
374
300
|
def to_stats # :nodoc:
|
375
301
|
stats = <<-INDEX
|
376
302
|
#{name} (#{self.class}):
|
@@ -390,7 +316,7 @@ module Picky
|
|
390
316
|
#
|
391
317
|
#
|
392
318
|
def to_s
|
393
|
-
"#{self.class}(#{name}, result_id: #{result_identifier}, source: #{source}, categories: #{categories})"
|
319
|
+
"#{self.class}(#{name}, result_id: #{result_identifier}, source: #{@source}, categories: #{categories})"
|
394
320
|
end
|
395
321
|
|
396
322
|
end
|
@@ -6,8 +6,7 @@ module Picky
|
|
6
6
|
#
|
7
7
|
class Index
|
8
8
|
|
9
|
-
attr_reader :
|
10
|
-
:combinator
|
9
|
+
attr_reader :combinator
|
11
10
|
|
12
11
|
delegate :load_from_cache,
|
13
12
|
:analyze,
|
@@ -17,6 +16,18 @@ module Picky
|
|
17
16
|
|
18
17
|
alias reload load_from_cache
|
19
18
|
|
19
|
+
# Define how the results of this index are identified.
|
20
|
+
# (Shown in the client, for example)
|
21
|
+
#
|
22
|
+
# Default is the name of the index.
|
23
|
+
#
|
24
|
+
def result_identifier result_identifier = nil
|
25
|
+
result_identifier ? define_result_identifier(result_identifier) : (@result_identifier || @name)
|
26
|
+
end
|
27
|
+
def define_result_identifier result_identifier
|
28
|
+
@result_identifier = result_identifier
|
29
|
+
end
|
30
|
+
|
20
31
|
end
|
21
32
|
|
22
33
|
end
|
@@ -6,8 +6,7 @@ module Picky
|
|
6
6
|
#
|
7
7
|
class Index
|
8
8
|
|
9
|
-
attr_reader :
|
10
|
-
:bundle_class
|
9
|
+
attr_reader :bundle_class
|
11
10
|
|
12
11
|
# Delegators for indexing.
|
13
12
|
#
|
@@ -35,6 +34,19 @@ module Picky
|
|
35
34
|
end
|
36
35
|
end
|
37
36
|
|
37
|
+
# Define an index tokenizer on the index.
|
38
|
+
#
|
39
|
+
# Parameters are the exact same as for indexing.
|
40
|
+
#
|
41
|
+
def indexing options = {}
|
42
|
+
@tokenizer = if options.respond_to?(:tokenize)
|
43
|
+
options
|
44
|
+
else
|
45
|
+
options && Tokenizers::Index.new(options)
|
46
|
+
end
|
47
|
+
end
|
48
|
+
alias define_indexing indexing
|
49
|
+
|
38
50
|
# Check if the given enumerable source is empty.
|
39
51
|
#
|
40
52
|
# Note: Checking as early as possible to tell the
|
@@ -68,15 +80,6 @@ module Picky
|
|
68
80
|
categories.each &:cache
|
69
81
|
end
|
70
82
|
|
71
|
-
# Define an index tokenizer on the index.
|
72
|
-
#
|
73
|
-
# Parameters are the exact same as for indexing.
|
74
|
-
#
|
75
|
-
def indexing options = {}
|
76
|
-
@tokenizer = Tokenizers::Index.new options
|
77
|
-
end
|
78
|
-
alias define_indexing indexing
|
79
|
-
|
80
83
|
# Returns the installed tokenizer or the default.
|
81
84
|
#
|
82
85
|
def tokenizer
|
@@ -102,6 +105,7 @@ module Picky
|
|
102
105
|
@source = @source.respond_to?(:call) ? @source.call : @source
|
103
106
|
end
|
104
107
|
def define_source source
|
108
|
+
check_source source
|
105
109
|
@source = source
|
106
110
|
end
|
107
111
|
def raise_no_source
|
@@ -119,6 +123,19 @@ module Picky
|
|
119
123
|
NO_SOURCE
|
120
124
|
)
|
121
125
|
end
|
126
|
+
def check_source source # :nodoc:
|
127
|
+
raise ArgumentError.new(<<-SOURCE
|
128
|
+
|
129
|
+
|
130
|
+
The index "#{name}" should use a data source that responds to either the method #each, or the method #harvest, which yields(id, text), OR it can be a lambda/block, returning such a source.
|
131
|
+
Or it could use one of the built-in sources:
|
132
|
+
Sources::#{(Sources.constants - [:Base, :Wrappers, :NoCSVFileGiven, :NoCouchDBGiven]).join(',
|
133
|
+
Sources::')}
|
134
|
+
|
135
|
+
|
136
|
+
SOURCE
|
137
|
+
) unless source.respond_to?(:each) || source.respond_to?(:harvest) || source.respond_to?(:call)
|
138
|
+
end
|
122
139
|
|
123
140
|
# Define a key_format on the index.
|
124
141
|
#
|
@@ -131,6 +148,16 @@ module Picky
|
|
131
148
|
@key_format = key_format
|
132
149
|
end
|
133
150
|
|
151
|
+
# Define what to do after indexing.
|
152
|
+
# (Only used in the Sources::DB)
|
153
|
+
#
|
154
|
+
def after_indexing after_indexing = nil
|
155
|
+
after_indexing ? define_after_indexing(after_indexing) : @after_indexing
|
156
|
+
end
|
157
|
+
def define_after_indexing after_indexing
|
158
|
+
@after_indexing = after_indexing
|
159
|
+
end
|
160
|
+
|
134
161
|
end
|
135
162
|
|
136
163
|
end
|
data/lib/picky/loader.rb
CHANGED
data/lib/picky/search.rb
CHANGED
@@ -1,14 +1,16 @@
|
|
1
|
+
# encoding: utf-8
|
2
|
+
#
|
1
3
|
module Picky
|
2
4
|
|
3
|
-
# = Picky
|
5
|
+
# = Picky Searches
|
4
6
|
#
|
5
7
|
# A Picky Search is an object which:
|
6
|
-
#
|
7
|
-
#
|
8
|
+
# * holds one or more indexes
|
9
|
+
# * offers an interface to query these indexes.
|
8
10
|
#
|
9
11
|
# You connect URL paths to indexes via a Query.
|
10
12
|
#
|
11
|
-
# We recommend not
|
13
|
+
# We recommend to not use this directly, but connect it to an URL and query through one of these
|
12
14
|
# (Protip: Use "curl 'localhost:8080/query/path?query=exampletext')" in a Terminal.
|
13
15
|
#
|
14
16
|
class Search
|
@@ -16,39 +18,40 @@ module Picky
|
|
16
18
|
include Helpers::Measuring
|
17
19
|
|
18
20
|
attr_reader :indexes
|
19
|
-
|
20
|
-
attr_accessor :weights
|
21
|
+
attr_accessor :tokenizer, :weights
|
21
22
|
|
22
23
|
# Takes:
|
23
24
|
# * A number of indexes
|
24
|
-
# * Options hash (optional) with:
|
25
|
-
# * tokenizer: Tokenizers::Query.default by default.
|
26
|
-
# * weights: A hash of weights, or a Query::Weights object.
|
27
25
|
#
|
28
26
|
# TODO Add identifiers_to_remove (rename) and reduce_allocations_to_amount (rename).
|
29
27
|
# TODO categories_to_remove ?
|
30
28
|
#
|
31
29
|
# It is also possible to define the tokenizer and weights like so.
|
32
30
|
# Example:
|
33
|
-
# Search.new(index1, index2, index3) do
|
31
|
+
# search = Search.new(index1, index2, index3) do
|
34
32
|
# searching removes_characters: /[^a-z]/, etc.
|
35
33
|
# weights [:author, :title] => +3, [:title, :isbn] => +1
|
36
34
|
# end
|
37
35
|
#
|
38
36
|
def initialize *index_definitions
|
39
|
-
options = Hash === index_definitions.last ? index_definitions.pop : {}
|
40
|
-
|
41
37
|
@indexes = Query::Indexes.new *index_definitions, combinations_type_for(index_definitions)
|
42
|
-
searching options[:tokenizer]
|
43
|
-
boost options[:weights]
|
44
38
|
|
45
39
|
instance_eval(&Proc.new) if block_given?
|
40
|
+
|
41
|
+
@tokenizer ||= Tokenizers::Query.default
|
42
|
+
@weights ||= Query::Weights.new
|
43
|
+
|
44
|
+
self
|
46
45
|
end
|
47
46
|
|
48
|
-
#
|
49
|
-
# Search.new(index1, index2, index3) do
|
50
|
-
# searching removes_characters: /[^a-z]/,
|
51
|
-
#
|
47
|
+
# Examples:
|
48
|
+
# search = Search.new(index1, index2, index3) do
|
49
|
+
# searching removes_characters: /[^a-z]/,
|
50
|
+
# # etc.
|
51
|
+
# end
|
52
|
+
#
|
53
|
+
# search = Search.new(index1, index2, index3) do
|
54
|
+
# searching MyTokenizerThatRespondsToTheMethodTokenize.new
|
52
55
|
# end
|
53
56
|
#
|
54
57
|
def searching options
|
@@ -59,63 +62,17 @@ module Picky
|
|
59
62
|
end
|
60
63
|
end
|
61
64
|
|
62
|
-
# Returns the tokenizer if set or if not, the query tokenizer.
|
63
|
-
#
|
64
|
-
def tokenizer
|
65
|
-
@tokenizer || Tokenizers::Query.default
|
66
|
-
end
|
67
|
-
|
68
65
|
# Example:
|
69
|
-
# Search.new(
|
70
|
-
#
|
71
|
-
#
|
66
|
+
# search = Search.new(books_index, dvd_index, mp3_index) do
|
67
|
+
# boost [:author, :title] => +3,
|
68
|
+
# [:title, :isbn] => +1
|
72
69
|
# end
|
73
70
|
#
|
74
|
-
def boost
|
75
|
-
weights
|
71
|
+
def boost weights
|
72
|
+
weights ||= Query::Weights.new
|
76
73
|
@weights = Hash === weights ? Query::Weights.new(weights) : weights
|
77
74
|
end
|
78
75
|
|
79
|
-
# Returns the right combinations strategy for
|
80
|
-
# a number of query indexes.
|
81
|
-
#
|
82
|
-
# Currently it isn't possible using Memory and Redis etc.
|
83
|
-
# indexes in the same query index group.
|
84
|
-
#
|
85
|
-
# Picky will raise a Query::Indexes::DifferentTypesError.
|
86
|
-
#
|
87
|
-
@@mapping = {
|
88
|
-
Indexes::Memory => Query::Combinations::Memory,
|
89
|
-
Indexes::Redis => Query::Combinations::Redis
|
90
|
-
}
|
91
|
-
def combinations_type_for index_definitions_ary
|
92
|
-
index_types = extract_index_types index_definitions_ary
|
93
|
-
!index_types.empty? && @@mapping[*index_types] || Query::Combinations::Memory
|
94
|
-
end
|
95
|
-
def extract_index_types index_definitions_ary
|
96
|
-
index_types = index_definitions_ary.map(&:class)
|
97
|
-
index_types.uniq!
|
98
|
-
check_index_types index_types
|
99
|
-
index_types
|
100
|
-
end
|
101
|
-
def check_index_types index_types
|
102
|
-
raise_different index_types if index_types.size > 1
|
103
|
-
end
|
104
|
-
# Currently it isn't possible using Memory and Redis etc.
|
105
|
-
# indexes in the same query index group.
|
106
|
-
#
|
107
|
-
class DifferentTypesError < StandardError
|
108
|
-
def initialize types
|
109
|
-
@types = types
|
110
|
-
end
|
111
|
-
def to_s
|
112
|
-
"Currently it isn't possible to mix #{@types.join(" and ")} Indexes in the same Search instance."
|
113
|
-
end
|
114
|
-
end
|
115
|
-
def raise_different index_types
|
116
|
-
raise DifferentTypesError.new(index_types)
|
117
|
-
end
|
118
|
-
|
119
76
|
# This is the main entry point for a query.
|
120
77
|
# Use this in specs and also for running queries.
|
121
78
|
#
|
@@ -126,19 +83,15 @@ module Picky
|
|
126
83
|
#
|
127
84
|
# Note: The Rack adapter calls this method after unravelling the HTTP request.
|
128
85
|
#
|
129
|
-
|
130
|
-
|
131
|
-
def search_with_text text, ids = 20, offset = 0
|
132
|
-
search tokenized(text), ids.to_i, offset.to_i
|
86
|
+
def search text, ids = 20, offset = 0
|
87
|
+
search_with tokenized(text), ids.to_i, offset.to_i
|
133
88
|
end
|
134
89
|
|
135
90
|
# Runs the actual search using Query::Tokens.
|
136
91
|
#
|
137
|
-
# Note: Internal method, use #
|
92
|
+
# Note: Internal method, use #search
|
138
93
|
#
|
139
|
-
|
140
|
-
#
|
141
|
-
def search tokens, ids = 20, offset = 0
|
94
|
+
def search_with tokens, ids = 20, offset = 0
|
142
95
|
results = nil
|
143
96
|
|
144
97
|
duration = timed do
|
@@ -151,7 +104,7 @@ module Picky
|
|
151
104
|
|
152
105
|
# Execute a search using Query::Tokens.
|
153
106
|
#
|
154
|
-
# Note: Internal method, use #
|
107
|
+
# Note: Internal method, use #search.
|
155
108
|
#
|
156
109
|
def execute tokens, ids, offset
|
157
110
|
Results.from ids, offset, sorted_allocations(tokens)
|
@@ -172,6 +125,46 @@ module Picky
|
|
172
125
|
indexes.prepared_allocations_for tokens, weights
|
173
126
|
end
|
174
127
|
|
128
|
+
# Returns the right combinations strategy for
|
129
|
+
# a number of query indexes.
|
130
|
+
#
|
131
|
+
# Currently it isn't possible using Memory and Redis etc.
|
132
|
+
# indexes in the same query index group.
|
133
|
+
#
|
134
|
+
# Picky will raise a Query::Indexes::DifferentTypesError.
|
135
|
+
#
|
136
|
+
@@mapping = {
|
137
|
+
Indexes::Memory => Query::Combinations::Memory,
|
138
|
+
Indexes::Redis => Query::Combinations::Redis
|
139
|
+
}
|
140
|
+
def combinations_type_for index_definitions_ary
|
141
|
+
index_types = extract_index_types index_definitions_ary
|
142
|
+
!index_types.empty? && @@mapping[*index_types] || Query::Combinations::Memory
|
143
|
+
end
|
144
|
+
def extract_index_types index_definitions_ary
|
145
|
+
index_types = index_definitions_ary.map(&:class)
|
146
|
+
index_types.uniq!
|
147
|
+
check_index_types index_types
|
148
|
+
index_types
|
149
|
+
end
|
150
|
+
def check_index_types index_types
|
151
|
+
raise_different index_types if index_types.size > 1
|
152
|
+
end
|
153
|
+
# Currently it isn't possible using Memory and Redis etc.
|
154
|
+
# indexes in the same query index group.
|
155
|
+
#
|
156
|
+
class DifferentTypesError < StandardError
|
157
|
+
def initialize types
|
158
|
+
@types = types
|
159
|
+
end
|
160
|
+
def to_s
|
161
|
+
"Currently it isn't possible to mix #{@types.join(" and ")} Indexes in the same Search instance."
|
162
|
+
end
|
163
|
+
end
|
164
|
+
def raise_different index_types
|
165
|
+
raise DifferentTypesError.new(index_types)
|
166
|
+
end
|
167
|
+
|
175
168
|
# Display some nice information for the user.
|
176
169
|
#
|
177
170
|
def to_s
|
@@ -1,3 +1,5 @@
|
|
1
|
+
# encoding: utf-8
|
2
|
+
#
|
1
3
|
module Picky
|
2
4
|
|
3
5
|
module Tokenizers # :nodoc:all
|
@@ -120,7 +122,7 @@ Case sensitive? #{@case_sensitive ? "Yes." : "-"}
|
|
120
122
|
#
|
121
123
|
# Note: Currently only for indexing.
|
122
124
|
#
|
123
|
-
def
|
125
|
+
def rejects_token_if &condition
|
124
126
|
@reject_condition = condition
|
125
127
|
end
|
126
128
|
def reject tokens
|
@@ -170,7 +172,7 @@ Case sensitive? #{@case_sensitive ? "Yes." : "-"}
|
|
170
172
|
# Defaults.
|
171
173
|
#
|
172
174
|
splits_text_on options[:splits_text_on] || /\s/
|
173
|
-
|
175
|
+
rejects_token_if &(options[:rejects_token_if] || :blank?)
|
174
176
|
end
|
175
177
|
|
176
178
|
# Default preprocessing hook.
|