Added wordnet - #250

Andrew Kane
1 parent bb24e095
Showing 6 changed files with 45 additions and 1 deletions Show diff stats
CHANGELOG.md
README.md
lib/searchkick.rb
lib/searchkick/reindex.rb
test/synonyms_test.rb
test/test_helper.rb
 ## 0.8.2 [unreleased]
  
 - Added `async` to `callbacks` option
+- Added `wordnet` option
  
 ## 0.8.1
  
@@ -243,6 +243,27 @@ end
  
 Call `Product.reindex` after changing synonyms.
  
+### WordNet [master]
+
+Prepopulate English synonyms with the [WordNet database](http://en.wikipedia.org/wiki/WordNet).
+
+Download [WordNet 3.0](http://wordnetcode.princeton.edu/3.0/WNprolog-3.0.tar.gz) to each Elasticsearch server and move `wn_s.pl` to the `/var/lib` directory.
+
+```sh
+cd /tmp
+curl -o wordnet.tar.gz http://wordnetcode.princeton.edu/3.0/WNprolog-3.0.tar.gz
+tar -zxvf wordnet.tar.gz
+mv prolog/wn_s.pl /var/lib
+```
+
+Tell each model to use it:
+
+```ruby
+class Product < ActiveRecord::Base
+  searchkick wordnet: true
+end
+```
+
 ### Misspellings
  
 By default, Searchkick handles misspelled queries by returning results with an [edit distance](http://en.wikipedia.org/wiki/Levenshtein_distance) of one. To turn off this feature, use:
@@ -19,8 +19,10 @@ module Searchkick
  
   class << self
     attr_accessor :search_method_name
+    attr_accessor :wordnet_path
   end
   self.search_method_name = :search
+  self.wordnet_path = "/var/lib/wn_s.pl"
  
   def self.client
     @client ||= Elasticsearch::Client.new(url: ENV["ELASTICSEARCH_URL"])
@@ -230,6 +230,17 @@ module Searchkick
           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" }
@@ -42,4 +42,9 @@ class TestSynonyms &lt; Minitest::Test
     assert_search "scallions", ["Green Onions"]
   end
  
+  def test_wordnet
+    store_names ["Creature", "Beast", "Dragon"], Animal
+    assert_search "animal", ["Creature", "Beast"], {}, Animal
+  end
+
 end
@@ -171,7 +171,11 @@ class Store
 end
  
 class Animal
-  searchkick autocomplete: [:name], suggest: [:name], index_name: -> { "#{self.name.tableize}-#{Date.today.year}" }
+  searchkick \
+    autocomplete: [:name],
+    suggest: [:name],
+    index_name: -> { "#{self.name.tableize}-#{Date.today.year}" },
+    wordnet: true
 end
  
 Product.searchkick_index.delete if Product.searchkick_index.exists?
1	1	## 0.8.2 [unreleased]
2	2
3	3	- Added `async` to `callbacks` option
	4	+- Added `wordnet` option
4	5
5	6	## 0.8.1
6	7
...	...
...	...	@@ -243,6 +243,27 @@ end
243	243
244	244	Call `Product.reindex` after changing synonyms.
245	245
	246	+### WordNet [master]
	247	+
	248	+Prepopulate English synonyms with the [WordNet database](http://en.wikipedia.org/wiki/WordNet).
	249	+
	250	+Download [WordNet 3.0](http://wordnetcode.princeton.edu/3.0/WNprolog-3.0.tar.gz) to each Elasticsearch server and move `wn_s.pl` to the `/var/lib` directory.
	251	+
	252	+```sh
	253	+cd /tmp
	254	+curl -o wordnet.tar.gz http://wordnetcode.princeton.edu/3.0/WNprolog-3.0.tar.gz
	255	+tar -zxvf wordnet.tar.gz
	256	+mv prolog/wn_s.pl /var/lib
	257	+```
	258	+
	259	+Tell each model to use it:
	260	+
	261	+```ruby
	262	+class Product < ActiveRecord::Base
	263	+ searchkick wordnet: true
	264	+end
	265	+```
	266	+
246	267	### Misspellings
247	268
248	269	By default, Searchkick handles misspelled queries by returning results with an [edit distance](http://en.wikipedia.org/wiki/Levenshtein_distance) of one. To turn off this feature, use:
...	...
...	...	@@ -19,8 +19,10 @@ module Searchkick
19	19
20	20	class << self
21	21	attr_accessor :search_method_name
	22	+ attr_accessor :wordnet_path
22	23	end
23	24	self.search_method_name = :search
	25	+ self.wordnet_path = "/var/lib/wn_s.pl"
24	26
25	27	def self.client
26	28	@client \|\|= Elasticsearch::Client.new(url: ENV["ELASTICSEARCH_URL"])
...	...
...	...	@@ -230,6 +230,17 @@ module Searchkick
230	230	settings[:analysis][:analyzer][:default_index][:filter] << "searchkick_synonym"
231	231	end
232	232
	233	+ if options[:wordnet]
	234	+ settings[:analysis][:filter][:searchkick_wordnet] = {
	235	+ type: "synonym",
	236	+ format: "wordnet",
	237	+ synonyms_path: Searchkick.wordnet_path
	238	+ }
	239	+
	240	+ settings[:analysis][:analyzer][:default_index][:filter].insert(4, "searchkick_wordnet")
	241	+ settings[:analysis][:analyzer][:default_index][:filter] << "searchkick_wordnet"
	242	+ end
	243	+
233	244	if options[:special_characters] == false
234	245	settings[:analysis][:analyzer].each do \|analyzer, analyzer_settings\|
235	246	analyzer_settings[:filter].reject!{\|f\| f == "asciifolding" }
...	...
...	...	@@ -42,4 +42,9 @@ class TestSynonyms < Minitest::Test
42	42	assert_search "scallions", ["Green Onions"]
43	43	end
44	44
	45	+ def test_wordnet
	46	+ store_names ["Creature", "Beast", "Dragon"], Animal
	47	+ assert_search "animal", ["Creature", "Beast"], {}, Animal
	48	+ end
	49	+
45	50	end
...	...
...	...	@@ -171,7 +171,11 @@ class Store
171	171	end
172	172
173	173	class Animal
174		- searchkick autocomplete: [:name], suggest: [:name], index_name: -> { "#{self.name.tableize}-#{Date.today.year}" }
	174	+ searchkick \
	175	+ autocomplete: [:name],
	176	+ suggest: [:name],
	177	+ index_name: -> { "#{self.name.tableize}-#{Date.today.year}" },
	178	+ wordnet: true
175	179	end
176	180
177	181	Product.searchkick_index.delete if Product.searchkick_index.exists?
...	...