[Completes #76012166] searchers should surface pics with tag "apollo1…

…1" when searching on "apollo 11"
GSA · Aug 11, 2014 · 9a29de8 · 9a29de8
1 parent 0ab10cd
commit 9a29de8
Show file tree

Hide file tree

Showing 5 changed files with 49 additions and 8 deletions.
diff --git a/app/models/flickr_photo.rb b/app/models/flickr_photo.rb
@@ -13,7 +13,7 @@ class FlickrPhoto
   attribute :title, String, mapping: { type: 'string', analyzer: 'en_analyzer', copy_to: 'bigram' }
   attribute :description, String, mapping: { type: 'string', analyzer: 'en_analyzer', copy_to: 'bigram' }
   attribute :taken_at, Date
-  attribute :tags, String, mapping: ElasticSettings::KEYWORD
+  attribute :tags, String, mapping: ElasticSettings::TAG
   attribute :url, String, mapping: ElasticSettings::KEYWORD
   attribute :thumbnail_url, String, mapping: ElasticSettings::KEYWORD
   attribute :popularity, Integer, default: 0, mapping: { type: 'integer', index: :not_analyzed }

diff --git a/app/models/image_search.rb b/app/models/image_search.rb
@@ -5,7 +5,7 @@ class ImageSearch
   DEFAULT_PRE_TAG = '<strong>'
   DEFAULT_POST_TAG = '</strong>'
   NO_HITS = { 'hits' => { 'total' => 0, 'offset' => 0, 'hits' => [] } }
-  TEXT_FIELDS = %w(title description caption tags^2)
+  TEXT_FIELDS = %w(title description caption)
 
   def initialize(query, options)
     @query = (query || '').squish
@@ -140,15 +140,33 @@ def owner_terms(json, profiles)
 
   def filtered_query_query(json)
     json.query do
-      json.simple_query_string do
-        json.fields TEXT_FIELDS
+      json.bool do
+        json.set! :should do
+          json.child! { match_tags(json) }
+          json.child! { simple_query_string(json) }
+        end
+      end
+    end
+  end
+
+  def match_tags(json)
+    json.match do
+      json.tags do
         json.query @query
-        json.analyzer "en_analyzer"
-        json.default_operator "AND"
+        json.analyzer "tag_analyzer"
       end
     end
   end
 
+  def simple_query_string(json)
+    json.simple_query_string do
+      json.fields TEXT_FIELDS
+      json.query @query
+      json.analyzer "en_analyzer"
+      json.default_operator "AND"
+    end
+  end
+
   def pre_tags
     [DEFAULT_PRE_TAG]
   end

diff --git a/app/models/instagram_photo.rb b/app/models/instagram_photo.rb
@@ -9,7 +9,7 @@ class InstagramPhoto
   end
 
   attribute :username, String, mapping: ElasticSettings::KEYWORD
-  attribute :tags, String, mapping: ElasticSettings::KEYWORD
+  attribute :tags, String, mapping: ElasticSettings::TAG
   attribute :url, String, mapping: ElasticSettings::KEYWORD
   attribute :thumbnail_url, String, mapping: ElasticSettings::KEYWORD
   attribute :caption, String, mapping: { type: 'string', analyzer: 'en_analyzer', copy_to: 'bigram' }

diff --git a/lib/elastic_settings.rb b/lib/elastic_settings.rb
@@ -1,12 +1,14 @@
 module ElasticSettings
   KEYWORD = { type: 'string', analyzer: 'case_insensitive_keyword_analyzer' }
+  TAG = { type: 'string', analyzer: 'tag_analyzer' }
   ENGLISH_STOPWORDS = %w(a an and are as at be but by for if in into is no not of on or s such t that the their then there these they this to was with)
 
   COMMON = {
     index: {
       analysis: {
         char_filter: {
-          ignore_chars: { type: "mapping", mappings: ["'=>", "’=>", "`=>"] }
+          ignore_chars: { type: "mapping", mappings: ["'=>", "’=>", "`=>"] },
+          strip_whitespace: { type: "mapping", mappings: ["\\u0020=>"] }
         },
         filter: {
           bigram_filter: { type: 'shingle' },
@@ -26,6 +28,12 @@ module ElasticSettings
             char_filter: %w(ignore_chars),
             filter: %w(standard asciifolding lowercase bigram_filter)
           },
+          tag_analyzer: {
+            type: "custom",
+            tokenizer: "standard",
+            char_filter: %w(strip_whitespace),
+            filter: %w(standard asciifolding lowercase)
+          },
           case_insensitive_keyword_analyzer: {
             tokenizer: 'keyword',
             char_filter: %w(ignore_chars),

diff --git a/spec/models/image_search_spec.rb b/spec/models/image_search_spec.rb
@@ -17,6 +17,21 @@
     end
   end
 
+  context 'when smooshed user query matches tag in either Instagram or Flickr indexes' do
+    before do
+      FlickrPhoto.create(id: "photo1", owner: "owner1", profile_type: 'user', tags: %w(apollo11 space), title: "title1 earth", description: "desc 1", taken_at: Date.current, popularity: 100, url: "http://photo1", thumbnail_url: "http://photo_thumbnail1")
+      FlickrPhoto.refresh_index!
+      InstagramPhoto.create(id: "123456", username: 'user1', tags: %w(earth apollo11), caption: 'first photo of earth', taken_at: Date.current, popularity: 101, url: "http://photo2", thumbnail_url: "http://photo_thumbnail2")
+      InstagramPhoto.refresh_index!
+    end
+
+    it 'should return results from both indexes' do
+      image_search = ImageSearch.new("apollo 11", {})
+      image_search_results = image_search.search
+      expect(image_search_results.results.collect(&:type).uniq).to match_array(["InstagramPhoto", "FlickrPhoto"])
+    end
+  end
+
   context 'when search term yields no results but a similar spelling does have results' do
     before do
       FlickrPhoto.create(id: "photo1", owner: "owner1", profile_type: 'user', tags: [], title: "title1 earth", description: "desc 1", taken_at: Date.current, popularity: 100, url: "http://photo1", thumbnail_url: "http://photo_thumbnail1")