relax criteria for uniqueness in CollectUnique

LibraryOfCongress · Nov 27, 2017 · 1c05c32 · 1c05c32
1 parent 63fa7e5
commit 1c05c32
Show file tree

Hide file tree

Showing 6 changed files with 55 additions and 1 deletion.
diff --git a/app/assets/javascripts/components/browse/browse-panel.cjsx b/app/assets/javascripts/components/browse/browse-panel.cjsx
@@ -140,6 +140,26 @@ module.exports = React.createClass
                                 {subj.data.category}
                               </div>
                           }
+
+                          {
+                            if subj.data.values && subj.data.values[0].caption
+                              <div className="caption">
+                                {subj.data.values[0].caption}
+                              </div>
+                          }
+                          {
+                            if subj.data.values && subj.data.values[0].creator
+                              <div className="creator">
+                                By {subj.data.values[0].creator}
+                              </div>
+                          }
+                          {
+                            if subj.data.values && subj.data.values[0].category
+                              <div className="category">
+                                {subj.data.values[0].category}
+                              </div>
+                          }
+
                           {
                             if subj.meta_data.subject_description
                               <div className="citation">

diff --git a/app/assets/javascripts/components/view/view-panel.cjsx b/app/assets/javascripts/components/view/view-panel.cjsx
@@ -55,6 +55,27 @@ module.exports = React.createClass
                         {subj.data.category}
                       </div>
                   }
+
+                  {
+                    if subj.data.values && subj.data.values[0].caption
+                      <div className="caption">
+                        {subj.data.values[0].caption}
+                      </div>
+                  }
+                  {
+                    if subj.data.values && subj.data.values[0].creator
+                      <div className="creator">
+                        By {subj.data.values[0].creator}
+                      </div>
+                  }
+                  {
+                    if subj.data.values && subj.data.values[0].category
+                      <div className="category">
+                        {subj.data.values[0].category}
+                      </div>
+                  }
+
+
                   {
                     if subj.meta_data.subject_description
                       <div className="citation">

diff --git a/app/models/subject_generation_methods/collect_unique.rb b/app/models/subject_generation_methods/collect_unique.rb
@@ -16,7 +16,11 @@ def process_classification(classification)
       if classification.child_subject.persisted?
 
         values = classification.child_subject.data['values'].nil? ? [] : classification.child_subject.data['values']
-        values.push ann unless values.include? ann
+        if classification.workflow.fuzzy_match
+          values.push ann unless values.map {|x| scrub(x)}.include? scrub(ann)
+        else
+          values.push ann unless values.include? ann
+        end
 
         atts[:data] = {'values' => values}
 
@@ -74,6 +78,11 @@ def process_classification(classification)
       classification.child_subject
     end
 
+    def scrub(annotation)
+      # remove all extra whitespaces and punctuations, convert to lower case
+      Hash[*annotation.map{|k,str| [k,str.gsub(/[[:punct:]]/, ' ').gsub(/\s+/, ' ').downcase.strip]}.flatten]
+    end
+
   end
 
 end
diff --git a/app/models/workflow.rb b/app/models/workflow.rb
@@ -23,6 +23,7 @@ class Workflow
   field    :order,                                           type: Integer,   default: 0
   field    :show_completion_assessment_task,                 type: Boolean,   default: true
   field    :show_bad_subject_button,                         type: Boolean,   default: true
+  field    :fuzzy_match,                                     type: Boolean,   default: false
 
   has_many     :subjects
   has_many     :classifications

diff --git a/docker-compose-dev.yml b/docker-compose-dev.yml
@@ -3,6 +3,8 @@ services:
   mongodb:
     image: mongo:2.6
     command: mongod --setParameter textSearchEnabled=true
+    ports:
+      - 27017:27017
     volumes:
       - /Users/twan/data/db:/data/db
 

diff --git a/project/loc/workflows/transcribe.json b/project/loc/workflows/transcribe.json
@@ -6,6 +6,7 @@
   "generates_subjects_after": 3,
   "generates_subjects_method": "collect-unique",
   "subject_fetch_limit": 50,
+  "fuzzy_match": true,
 
   "tasks": {