From 1b3ba42f6cb2838f19efddb1244f3eeacaa438d9 Mon Sep 17 00:00:00 2001
From: Matthew Bernhardt <mjbernha@mit.edu>
Date: Tue, 16 Jul 2024 10:50:05 -0400
Subject: [PATCH 1/5] Basic rails model generation

bin/rails generate model Detector::SuggestedResource title:string url:string phrase:string:uniq fingerprint:string:uniq
---
 app/models/detector/suggested_resource.rb     |  2 ++
 ...850_create_detector_suggested_resources.rb | 14 +++++++++++++
 db/schema.rb                                  | 13 +++++++++++-
 .../fixtures/detector/suggested_resources.yml | 20 +++++++++++++++++++
 .../detector/suggested_resource_test.rb       |  7 +++++++
 5 files changed, 55 insertions(+), 1 deletion(-)
 create mode 100644 app/models/detector/suggested_resource.rb
 create mode 100644 db/migrate/20240716143850_create_detector_suggested_resources.rb
 create mode 100644 test/fixtures/detector/suggested_resources.yml
 create mode 100644 test/models/detector/suggested_resource_test.rb

diff --git a/app/models/detector/suggested_resource.rb b/app/models/detector/suggested_resource.rb
new file mode 100644
index 0000000..7675cbc
--- /dev/null
+++ b/app/models/detector/suggested_resource.rb
@@ -0,0 +1,2 @@
+class Detector::SuggestedResource < ApplicationRecord
+end
diff --git a/db/migrate/20240716143850_create_detector_suggested_resources.rb b/db/migrate/20240716143850_create_detector_suggested_resources.rb
new file mode 100644
index 0000000..b13e90b
--- /dev/null
+++ b/db/migrate/20240716143850_create_detector_suggested_resources.rb
@@ -0,0 +1,14 @@
+class CreateDetectorSuggestedResources < ActiveRecord::Migration[7.1]
+  def change
+    create_table :detector_suggested_resources do |t|
+      t.string :title
+      t.string :url
+      t.string :phrase
+      t.string :fingerprint
+
+      t.timestamps
+    end
+    add_index :detector_suggested_resources, :phrase, unique: true
+    add_index :detector_suggested_resources, :fingerprint, unique: true
+  end
+end
diff --git a/db/schema.rb b/db/schema.rb
index 3acaf32..52c9d2d 100644
--- a/db/schema.rb
+++ b/db/schema.rb
@@ -10,7 +10,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema[7.1].define(version: 2024_07_01_205444) do
+ActiveRecord::Schema[7.1].define(version: 2024_07_16_143850) do
   create_table "detector_journals", force: :cascade do |t|
     t.string "name"
     t.json "additional_info"
@@ -19,6 +19,17 @@
     t.index ["name"], name: "index_detector_journals_on_name"
   end
 
+  create_table "detector_suggested_resources", force: :cascade do |t|
+    t.string "title"
+    t.string "url"
+    t.string "phrase"
+    t.string "fingerprint"
+    t.datetime "created_at", null: false
+    t.datetime "updated_at", null: false
+    t.index ["fingerprint"], name: "index_detector_suggested_resources_on_fingerprint", unique: true
+    t.index ["phrase"], name: "index_detector_suggested_resources_on_phrase", unique: true
+  end
+
   create_table "metrics_algorithms", force: :cascade do |t|
     t.date "month"
     t.integer "doi"
diff --git a/test/fixtures/detector/suggested_resources.yml b/test/fixtures/detector/suggested_resources.yml
new file mode 100644
index 0000000..c5ace34
--- /dev/null
+++ b/test/fixtures/detector/suggested_resources.yml
@@ -0,0 +1,20 @@
+# Read about fixtures at https://api.rubyonrails.org/classes/ActiveRecord/FixtureSet.html
+
+
+one:
+  title: JSTOR
+  url: https://libguides.mit.edu/jstor
+  phrase: jstor
+  fingerprint: jstor
+
+two:
+  title: Web of Science
+  url: https://libguides.mit.edu/webofsci
+  phrase: web of science
+  fingerprint: of science web
+
+three:
+  title: Web of Knowledge
+  url: https://libguides.mit.edu/webofsci
+  phrase: web of knowledge
+  fingerprint: knowledge of web
diff --git a/test/models/detector/suggested_resource_test.rb b/test/models/detector/suggested_resource_test.rb
new file mode 100644
index 0000000..7a4c844
--- /dev/null
+++ b/test/models/detector/suggested_resource_test.rb
@@ -0,0 +1,7 @@
+require "test_helper"
+
+class Detector::SuggestedResourceTest < ActiveSupport::TestCase
+  # test "the truth" do
+  #   assert true
+  # end
+end

From 03597ba02b757a6e29c980d53884d473d9a29275 Mon Sep 17 00:00:00 2001
From: Matthew Bernhardt <mjbernha@mit.edu>
Date: Tue, 16 Jul 2024 14:30:19 -0400
Subject: [PATCH 2/5] Model annotation

Output of `make annotate`
---
 app/models/detector/suggested_resource.rb       | 12 ++++++++++++
 test/fixtures/detector/suggested_resources.yml  | 12 ++++++++++++
 test/models/detector/suggested_resource_test.rb | 12 ++++++++++++
 3 files changed, 36 insertions(+)

diff --git a/app/models/detector/suggested_resource.rb b/app/models/detector/suggested_resource.rb
index 7675cbc..12ed72b 100644
--- a/app/models/detector/suggested_resource.rb
+++ b/app/models/detector/suggested_resource.rb
@@ -1,2 +1,14 @@
+# == Schema Information
+#
+# Table name: detector_suggested_resources
+#
+#  id          :integer          not null, primary key
+#  title       :string
+#  url         :string
+#  phrase      :string
+#  fingerprint :string
+#  created_at  :datetime         not null
+#  updated_at  :datetime         not null
+#
 class Detector::SuggestedResource < ApplicationRecord
 end
diff --git a/test/fixtures/detector/suggested_resources.yml b/test/fixtures/detector/suggested_resources.yml
index c5ace34..a145aab 100644
--- a/test/fixtures/detector/suggested_resources.yml
+++ b/test/fixtures/detector/suggested_resources.yml
@@ -1,3 +1,15 @@
+# == Schema Information
+#
+# Table name: detector_suggested_resources
+#
+#  id          :integer          not null, primary key
+#  title       :string
+#  url         :string
+#  phrase      :string
+#  fingerprint :string
+#  created_at  :datetime         not null
+#  updated_at  :datetime         not null
+#
 # Read about fixtures at https://api.rubyonrails.org/classes/ActiveRecord/FixtureSet.html
 
 
diff --git a/test/models/detector/suggested_resource_test.rb b/test/models/detector/suggested_resource_test.rb
index 7a4c844..8f43900 100644
--- a/test/models/detector/suggested_resource_test.rb
+++ b/test/models/detector/suggested_resource_test.rb
@@ -1,3 +1,15 @@
+# == Schema Information
+#
+# Table name: detector_suggested_resources
+#
+#  id          :integer          not null, primary key
+#  title       :string
+#  url         :string
+#  phrase      :string
+#  fingerprint :string
+#  created_at  :datetime         not null
+#  updated_at  :datetime         not null
+#
 require "test_helper"
 
 class Detector::SuggestedResourceTest < ActiveSupport::TestCase

From d8d45e5d1ca91a77f9ab36dbef8b1d3bee703cb4 Mon Sep 17 00:00:00 2001
From: Matthew Bernhardt <mjbernha@mit.edu>
Date: Wed, 24 Jul 2024 16:50:24 -0400
Subject: [PATCH 3/5] Build out fingerprinting logic and tests

** Why are these changes being introduced:

* With the SuggestedResource model being auto-generated and annotated in
  previous commits, we now need to implement the custom fingerprinting
  logic from the Bento application, and write some tests about that
  logic.

** Relevant ticket(s):

* https://mitlibraries.atlassian.net/browse/tco-22

** How does this address that need:

* The bulk of this commit is the calculate_fingerprint method in the
  detector::suggestedresource model. This is nearly the same as the
  model from Bento, with two changes:

  1. We no longer preserve c++ or c#, and run the gsub command in this
     method.
  2. The position of the .to_ascii method moves up in the order, to
     match the current documentation from OpenRefine. This also resolves
     some flaws in ordering and term handling that happened when the
     .to_ascii happens last.

* Fixture names are made more meaningful

* We add the stringex gem to get the .to_ascii method in the first
  place, which may end up not being needed as we continue to develop
  the application. For now, though, I suspect that we'll benefit from
  having it.

* Add some tests around how the fingerprint logic operates, focusing on
  character manipulation and the before_save hook's impact.

** Document any side effects to this change:

* The method applies the punctuation and symbol cleaning twice, because
  the .to_ascii method can introduce new punctuation when converting
  some letters. However, just running the regex once still had problems
  with these tests, so twice is necessary?

* This also removes a comment in the fixtures that was introduced when
  I re-ran the annotation command.
---
 Gemfile                                       |  2 +
 Gemfile.lock                                  |  2 +
 app/models/detector/suggested_resource.rb     | 36 ++++++++++++++
 .../fixtures/detector/suggested_resources.yml |  7 ++-
 .../detector/suggested_resource_test.rb       | 49 +++++++++++++++++--
 5 files changed, 89 insertions(+), 7 deletions(-)

diff --git a/Gemfile b/Gemfile
index c07aed4..ffe03d9 100644
--- a/Gemfile
+++ b/Gemfile
@@ -46,6 +46,8 @@ gem 'rack-cors'
 # Use Redis adapter to run Action Cable in production
 # gem "redis", ">= 4.0.1"
 
+gem 'stringex'
+
 # Use Kredis to get higher-level data types in Redis [https://github.com/rails/kredis]
 # gem "kredis"
 
diff --git a/Gemfile.lock b/Gemfile.lock
index 3ffa3b7..0748244 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -306,6 +306,7 @@ GEM
     sqlite3 (1.7.3-x86_64-linux)
     stimulus-rails (1.3.3)
       railties (>= 6.0.0)
+    stringex (2.8.6)
     stringio (3.1.1)
     strscan (3.1.0)
     thor (1.3.1)
@@ -368,6 +369,7 @@ DEPENDENCIES
   sprockets-rails
   sqlite3
   stimulus-rails
+  stringex
   turbo-rails
   tzinfo-data
   vcr
diff --git a/app/models/detector/suggested_resource.rb b/app/models/detector/suggested_resource.rb
index 12ed72b..f527622 100644
--- a/app/models/detector/suggested_resource.rb
+++ b/app/models/detector/suggested_resource.rb
@@ -10,5 +10,41 @@
 #  created_at  :datetime         not null
 #  updated_at  :datetime         not null
 #
+
+require 'stringex/core_ext'
+
 class Detector::SuggestedResource < ApplicationRecord
+  before_save :update_fingerprint
+
+  def update_fingerprint
+    self.fingerprint = calculate_fingerprint
+  end
+
+  # This implements the OpenRefine fingerprinting algorithm. See
+  # https://openrefine.org/docs/technical-reference/clustering-in-depth#fingerprint
+  def calculate_fingerprint
+    temp = self.phrase
+    temp.strip!
+    temp.downcase!
+
+    # This removes all punctuation and symbol characters from the string.
+    temp.gsub!(/\p{P}|\p{S}/, '')
+
+    # Normalize to ASCII (e.g. gödel and godel are liable to be intended to
+    # find the same thing)
+    temp = temp.to_ascii
+
+    # Coercion to ASCII can introduce new symbols, so we remove those now.
+    temp.gsub!(/\p{P}|\p{S}/, '')
+
+    # Tokenize
+    array = temp.split
+
+    # Remove duplicates and sort
+    array.uniq!
+    array.sort!
+
+    # Rejoin tokens
+    new_fingerprint = array.join(' ')
+  end
 end
diff --git a/test/fixtures/detector/suggested_resources.yml b/test/fixtures/detector/suggested_resources.yml
index a145aab..82c43a7 100644
--- a/test/fixtures/detector/suggested_resources.yml
+++ b/test/fixtures/detector/suggested_resources.yml
@@ -10,22 +10,21 @@
 #  created_at  :datetime         not null
 #  updated_at  :datetime         not null
 #
-# Read about fixtures at https://api.rubyonrails.org/classes/ActiveRecord/FixtureSet.html
 
 
-one:
+jstor:
   title: JSTOR
   url: https://libguides.mit.edu/jstor
   phrase: jstor
   fingerprint: jstor
 
-two:
+web_of_science:
   title: Web of Science
   url: https://libguides.mit.edu/webofsci
   phrase: web of science
   fingerprint: of science web
 
-three:
+web_of_knowledge:
   title: Web of Knowledge
   url: https://libguides.mit.edu/webofsci
   phrase: web of knowledge
diff --git a/test/models/detector/suggested_resource_test.rb b/test/models/detector/suggested_resource_test.rb
index 8f43900..6754473 100644
--- a/test/models/detector/suggested_resource_test.rb
+++ b/test/models/detector/suggested_resource_test.rb
@@ -13,7 +13,50 @@
 require "test_helper"
 
 class Detector::SuggestedResourceTest < ActiveSupport::TestCase
-  # test "the truth" do
-  #   assert true
-  # end
+  test 'fingerprints are generated automatically' do
+    resource = {
+      title: 'Our latest resource',
+      url: 'https://example.org',
+      phrase: 'Our latest resource'
+    }
+
+    new_resource = Detector::SuggestedResource.create(resource)
+
+    assert new_resource.fingerprint == 'latest our resource'
+  end
+
+  test 'fingerprints are recalculated on save' do
+    resource = detector_suggested_resources('jstor')
+    refute resource.fingerprint == 'A brand new phrase'
+
+    resource.phrase = 'This is a brand new phrase'
+    resource.save
+    resource.reload
+
+    assert resource.fingerprint == 'a brand is new phrase this'
+  end
+
+  test 'fingerprints standardize characters used' do
+    resource = {
+      title: 'A wide range of characters',
+      url: 'https://example.org',
+      phrase: 'This phrase uses: WeIrD caPital letters, * (punctuation), and symbols™ like ¥€$'
+    }
+
+    new_resource = Detector::SuggestedResource.create(resource)
+
+    assert new_resource.fingerprint == 'and capital letters like phrase punctuation symbols this uses weird'
+  end
+
+  test 'fingerprints coerce characters to ASCII' do
+    resource = {
+      title: 'A wide range of characters',
+      url: 'https://example.org',
+      phrase: 'а а̀ а̂ а̄ ӓ б в г ґ д ђ ѓ е ѐ е̄ е̂ ё є ж з з́ ѕ и і ї ꙇ ѝ и̂ ӣ й ј к л љ м н њ о о̀ о̂ ō ӧ п р с с́ т ћ ќ у у̀ у̂ ӯ ў ӱ ф х ц ч џ ш щ ꙏ ъ ъ̀ ы ь ѣ э ю ю̀ я'
+    }
+
+    new_resource = Detector::SuggestedResource.create(resource)
+
+    assert new_resource.fingerprint == 'a b ch d dj dz dzh e f g gh gj i ia ie io iu j k kh kj l lj m n nj o p r s sh shch t ts tsh u v y yi z zh'
+  end
 end

From 76a62d69ff1c5336f77d92fff13604ede4dac9ad Mon Sep 17 00:00:00 2001
From: Matthew Bernhardt <mjbernha@mit.edu>
Date: Thu, 25 Jul 2024 16:21:37 -0400
Subject: [PATCH 4/5] Rubocop-inspired fixes after informal code review

---
 app/models/detector/suggested_resource.rb     | 57 ++++++++------
 .../detector/suggested_resource_test.rb       | 78 ++++++++++---------
 2 files changed, 74 insertions(+), 61 deletions(-)

diff --git a/app/models/detector/suggested_resource.rb b/app/models/detector/suggested_resource.rb
index f527622..f7c1180 100644
--- a/app/models/detector/suggested_resource.rb
+++ b/app/models/detector/suggested_resource.rb
@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 # == Schema Information
 #
 # Table name: detector_suggested_resources
@@ -13,38 +15,43 @@
 
 require 'stringex/core_ext'
 
-class Detector::SuggestedResource < ApplicationRecord
-  before_save :update_fingerprint
+module Detector
+  # Detector::SuggestedResource stores custom hints that we want to send to the
+  # user in response to specific strings. For example, a search for "web of
+  # science" should be met with our custom login link to Web of Science via MIT.
+  class SuggestedResource < ApplicationRecord
+    before_save :update_fingerprint
 
-  def update_fingerprint
-    self.fingerprint = calculate_fingerprint
-  end
+    def update_fingerprint
+      self.fingerprint = calculate_fingerprint
+    end
 
-  # This implements the OpenRefine fingerprinting algorithm. See
-  # https://openrefine.org/docs/technical-reference/clustering-in-depth#fingerprint
-  def calculate_fingerprint
-    temp = self.phrase
-    temp.strip!
-    temp.downcase!
+    # This implements the OpenRefine fingerprinting algorithm. See
+    # https://openrefine.org/docs/technical-reference/clustering-in-depth#fingerprint
+    def calculate_fingerprint
+      temp = phrase
+      temp.strip!
+      temp.downcase!
 
-    # This removes all punctuation and symbol characters from the string.
-    temp.gsub!(/\p{P}|\p{S}/, '')
+      # This removes all punctuation and symbol characters from the string.
+      temp.gsub!(/\p{P}|\p{S}/, '')
 
-    # Normalize to ASCII (e.g. gödel and godel are liable to be intended to
-    # find the same thing)
-    temp = temp.to_ascii
+      # Normalize to ASCII (e.g. gödel and godel are liable to be intended to
+      # find the same thing)
+      temp = temp.to_ascii
 
-    # Coercion to ASCII can introduce new symbols, so we remove those now.
-    temp.gsub!(/\p{P}|\p{S}/, '')
+      # Coercion to ASCII can introduce new symbols, so we remove those now.
+      temp.gsub!(/\p{P}|\p{S}/, '')
 
-    # Tokenize
-    array = temp.split
+      # Tokenize
+      array = temp.split
 
-    # Remove duplicates and sort
-    array.uniq!
-    array.sort!
+      # Remove duplicates and sort
+      array.uniq!
+      array.sort!
 
-    # Rejoin tokens
-    new_fingerprint = array.join(' ')
+      # Rejoin tokens
+      array.join(' ')
+    end
   end
 end
diff --git a/test/models/detector/suggested_resource_test.rb b/test/models/detector/suggested_resource_test.rb
index 6754473..b981c40 100644
--- a/test/models/detector/suggested_resource_test.rb
+++ b/test/models/detector/suggested_resource_test.rb
@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 # == Schema Information
 #
 # Table name: detector_suggested_resources
@@ -10,53 +12,57 @@
 #  created_at  :datetime         not null
 #  updated_at  :datetime         not null
 #
-require "test_helper"
+require 'test_helper'
 
-class Detector::SuggestedResourceTest < ActiveSupport::TestCase
-  test 'fingerprints are generated automatically' do
-    resource = {
-      title: 'Our latest resource',
-      url: 'https://example.org',
-      phrase: 'Our latest resource'
-    }
+module Detector
+  class SuggestedResourceTest < ActiveSupport::TestCase
+    test 'fingerprints are generated automatically' do
+      resource = {
+        title: 'Our latest resource',
+        url: 'https://example.org',
+        phrase: 'Our latest resource'
+      }
 
-    new_resource = Detector::SuggestedResource.create(resource)
+      new_resource = Detector::SuggestedResource.create(resource)
 
-    assert new_resource.fingerprint == 'latest our resource'
-  end
+      assert new_resource.fingerprint == 'latest our resource'
+    end
 
-  test 'fingerprints are recalculated on save' do
-    resource = detector_suggested_resources('jstor')
-    refute resource.fingerprint == 'A brand new phrase'
+    test 'fingerprints are recalculated on save' do
+      resource = detector_suggested_resources('jstor')
+      refute resource.fingerprint == 'A brand new phrase'
 
-    resource.phrase = 'This is a brand new phrase'
-    resource.save
-    resource.reload
+      resource.phrase = 'This is a brand new phrase'
+      resource.save
+      resource.reload
 
-    assert resource.fingerprint == 'a brand is new phrase this'
-  end
+      assert resource.fingerprint == 'a brand is new phrase this'
+    end
 
-  test 'fingerprints standardize characters used' do
-    resource = {
-      title: 'A wide range of characters',
-      url: 'https://example.org',
-      phrase: 'This phrase uses: WeIrD caPital letters, * (punctuation), and symbols™ like ¥€$'
-    }
+    test 'fingerprints standardize characters used' do
+      resource = {
+        title: 'A wide range of characters',
+        url: 'https://example.org',
+        phrase: 'This phrase uses: WeIrD caPital letters, * (punctuation), and symbols™ like ¥€$'
+      }
 
-    new_resource = Detector::SuggestedResource.create(resource)
+      new_resource = Detector::SuggestedResource.create(resource)
 
-    assert new_resource.fingerprint == 'and capital letters like phrase punctuation symbols this uses weird'
-  end
+      assert new_resource.fingerprint == 'and capital letters like phrase punctuation symbols this uses weird'
+    end
 
-  test 'fingerprints coerce characters to ASCII' do
-    resource = {
-      title: 'A wide range of characters',
-      url: 'https://example.org',
-      phrase: 'а а̀ а̂ а̄ ӓ б в г ґ д ђ ѓ е ѐ е̄ е̂ ё є ж з з́ ѕ и і ї ꙇ ѝ и̂ ӣ й ј к л љ м н њ о о̀ о̂ ō ӧ п р с с́ т ћ ќ у у̀ у̂ ӯ ў ӱ ф х ц ч џ ш щ ꙏ ъ ъ̀ ы ь ѣ э ю ю̀ я'
-    }
+    test 'fingerprints coerce characters to ASCII' do
+      resource = {
+        title: 'A wide range of characters',
+        url: 'https://example.org',
+        phrase: 'а а̀ а̂ а̄ ӓ б в г ґ д ђ ѓ е ѐ е̄ е̂ ё є ж з з́ ѕ и і ї ꙇ ѝ и̂ ӣ й ј к л љ м н њ о о̀ о̂ ō ӧ п р с с́'\
+        ' т ћ ќ у у̀ у̂ ӯ ў ӱ ф х ц ч џ ш щ ꙏ ъ ъ̀ ы ь ѣ э ю ю̀ я'
+      }
 
-    new_resource = Detector::SuggestedResource.create(resource)
+      new_resource = Detector::SuggestedResource.create(resource)
 
-    assert new_resource.fingerprint == 'a b ch d dj dz dzh e f g gh gj i ia ie io iu j k kh kj l lj m n nj o p r s sh shch t ts tsh u v y yi z zh'
+      assert new_resource.fingerprint == 'a b ch d dj dz dzh e f g gh gj i ia ie io iu j k kh kj l lj m n nj o p r s '\
+      'sh shch t ts tsh u v y yi z zh'
+    end
   end
 end

From b5786115a7ed0124fb5c7f0dc13806319571564c Mon Sep 17 00:00:00 2001
From: Matthew Bernhardt <mjbernha@mit.edu>
Date: Fri, 26 Jul 2024 12:08:05 -0400
Subject: [PATCH 5/5] Updates after informal code review

* There are more tests, each explicitly testing one aspect of the
  fingerprinting. Most notably is one to confirm that the original
  phrase is not affected while calculating the fingerprint.

* The calculate_fingerprint method now has an explicit argument to
  receive the phrase

* Variable names internal to this method have been clarified

* Bang methods are no longer being used
---
 app/models/detector/suggested_resource.rb     | 24 +++----
 .../detector/suggested_resource_test.rb       | 71 ++++++++++++++++---
 2 files changed, 75 insertions(+), 20 deletions(-)

diff --git a/app/models/detector/suggested_resource.rb b/app/models/detector/suggested_resource.rb
index f7c1180..526074e 100644
--- a/app/models/detector/suggested_resource.rb
+++ b/app/models/detector/suggested_resource.rb
@@ -23,35 +23,35 @@ class SuggestedResource < ApplicationRecord
     before_save :update_fingerprint
 
     def update_fingerprint
-      self.fingerprint = calculate_fingerprint
+      self.fingerprint = calculate_fingerprint(phrase)
     end
 
     # This implements the OpenRefine fingerprinting algorithm. See
     # https://openrefine.org/docs/technical-reference/clustering-in-depth#fingerprint
-    def calculate_fingerprint
-      temp = phrase
-      temp.strip!
-      temp.downcase!
+    def calculate_fingerprint(old_phrase)
+      modified_phrase = old_phrase
+      modified_phrase = modified_phrase.strip
+      modified_phrase = modified_phrase.downcase
 
       # This removes all punctuation and symbol characters from the string.
-      temp.gsub!(/\p{P}|\p{S}/, '')
+      modified_phrase = modified_phrase.gsub(/\p{P}|\p{S}/, '')
 
       # Normalize to ASCII (e.g. gödel and godel are liable to be intended to
       # find the same thing)
-      temp = temp.to_ascii
+      modified_phrase = modified_phrase.to_ascii
 
       # Coercion to ASCII can introduce new symbols, so we remove those now.
-      temp.gsub!(/\p{P}|\p{S}/, '')
+      modified_phrase = modified_phrase.gsub(/\p{P}|\p{S}/, '')
 
       # Tokenize
-      array = temp.split
+      tokens = modified_phrase.split
 
       # Remove duplicates and sort
-      array.uniq!
-      array.sort!
+      tokens = tokens.uniq
+      tokens = tokens.sort
 
       # Rejoin tokens
-      array.join(' ')
+      tokens.join(' ')
     end
   end
 end
diff --git a/test/models/detector/suggested_resource_test.rb b/test/models/detector/suggested_resource_test.rb
index b981c40..620b35b 100644
--- a/test/models/detector/suggested_resource_test.rb
+++ b/test/models/detector/suggested_resource_test.rb
@@ -39,16 +39,49 @@ class SuggestedResourceTest < ActiveSupport::TestCase
       assert resource.fingerprint == 'a brand is new phrase this'
     end
 
-    test 'fingerprints standardize characters used' do
-      resource = {
-        title: 'A wide range of characters',
-        url: 'https://example.org',
-        phrase: 'This phrase uses: WeIrD caPital letters, * (punctuation), and symbols™ like ¥€$'
-      }
+    test 'generating fingerprints does not alter the phrase' do
+      resource = detector_suggested_resources('jstor')
+      benchmark = 'This is an updated phrase! '
 
-      new_resource = Detector::SuggestedResource.create(resource)
+      refute resource.phrase == benchmark
+      resource.phrase = benchmark
+      resource.save
+      resource.reload
+
+      assert resource.phrase == benchmark
+    end
+
+    test 'fingerprints strip extra spaces' do
+      resource = detector_suggested_resources('jstor')
+      refute resource.fingerprint == 'i need space'
+
+      resource.phrase = '  i  need  space  '
+      resource.save
+      resource.reload
+
+      assert resource.fingerprint == 'i need space'
+    end
+
+    test 'fingerprints are coerced to lowercase' do
+      resource = detector_suggested_resources('jstor')
+      refute resource.fingerprint == 'ftw intercapping'
+
+      resource.phrase = 'InterCapping FTW'
+      resource.save
+      resource.reload
 
-      assert new_resource.fingerprint == 'and capital letters like phrase punctuation symbols this uses weird'
+      assert resource.fingerprint == 'ftw intercapping'
+    end
+
+    test 'fingerprints remove punctuation and symbols' do
+      resource = detector_suggested_resources('jstor')
+      refute resource.fingerprint == 'bullets perfect phrase punctuation quoted symbols'
+
+      resource.phrase = 'symbols™ + punctuation: * bullets! - "quoted phrase" (perfect) ¥€$'
+      resource.save
+      resource.reload
+
+      assert resource.fingerprint == 'bullets perfect phrase punctuation quoted symbols'
     end
 
     test 'fingerprints coerce characters to ASCII' do
@@ -64,5 +97,27 @@ class SuggestedResourceTest < ActiveSupport::TestCase
       assert new_resource.fingerprint == 'a b ch d dj dz dzh e f g gh gj i ia ie io iu j k kh kj l lj m n nj o p r s '\
       'sh shch t ts tsh u v y yi z zh'
     end
+
+    test 'fingerprints remove repeated words' do
+      resource = detector_suggested_resources('jstor')
+      refute resource.fingerprint == 'double'
+
+      resource.phrase = 'double double'
+      resource.save
+      resource.reload
+
+      assert resource.fingerprint == 'double'
+    end
+
+    test 'fingerprints sort words alphabetically' do
+      resource = detector_suggested_resources('jstor')
+      refute resource.fingerprint == 'delta gamma'
+
+      resource.phrase = 'gamma delta'
+      resource.save
+      resource.reload
+
+      assert resource.fingerprint == 'delta gamma'
+    end
   end
 end