From cc114650f57c73ea316defcc3c42e011304c5dba Mon Sep 17 00:00:00 2001
From: Marwin Hochfelsner <50826859+hlfan@users.noreply.github.com>
Date: Mon, 27 Oct 2025 21:01:54 +0100
Subject: [PATCH] Move tag2link transforms into module

---
 config/initializers/tag2link.rb | 18 ++----------
 lib/tag2link.rb                 | 20 ++++++++++++-
 test/lib/tag2link_test.rb       | 52 +++++++++++++++++++++++++++++++++
 3 files changed, 74 insertions(+), 16 deletions(-)

diff --git a/config/initializers/tag2link.rb b/config/initializers/tag2link.rb
index 771411e3e..e5fd8aea1 100644
--- a/config/initializers/tag2link.rb
+++ b/config/initializers/tag2link.rb
@@ -1,17 +1,5 @@
 # frozen_string_literal: true
 
-# A map of each OSM key to its formatter URL. For example:
-# { "ref:vatin" => "https://example.com/$1" }
-# The JSON data is an array with duplicate entries, which is not efficient for lookups.
-# So, convert it to a hash and only keep the item with the best rank.
-TAG2LINK = JSON.parse(Rails.root.join("node_modules/tag2link/index.json").read)
-               # exclude deprecated and third-party URLs
-               .reject { |item| item["rank"] == "deprecated" || item["source"] == "wikidata:P3303" }
-               .group_by { |item| item["key"] }
-               .transform_keys { |key| key.sub(/^Key:/, "") }
-               # move preferred to the start of the array
-               .transform_values { |items| items.sort_by { |item| item["rank"] == "preferred" ? 0 : 1 }.uniq { |item| item["url"] } }
-               # exclude any that are ambiguous, i.e. the best and second-best have the same rank
-               .reject { |_key, value| value[1] && value[0]["rank"] == value[1]["rank"] }
-               # keep only the best match
-               .transform_values { |items| items[0]["url"] }
+require "tag2link"
+
+Tag2link.load(Rails.root.join("node_modules/tag2link/index.json"))
diff --git a/lib/tag2link.rb b/lib/tag2link.rb
index 8eb9484f5..b4ab9c3b7 100644
--- a/lib/tag2link.rb
+++ b/lib/tag2link.rb
@@ -1,13 +1,31 @@
 # frozen_string_literal: true
 
 module Tag2link
+  def self.load(path)
+    @dict = build_dict(JSON.parse(path.read)).freeze
+  end
+
   def self.link(key, value)
     # skip if it's a full URL
     return nil if %r{\Ahttps?://}.match?(value)
 
-    url_template = TAG2LINK[key]
+    url_template = @dict[key]
     return nil unless url_template
 
     url_template.gsub("$1", value)
   end
+
+  def self.build_dict(data)
+    data
+      # exclude deprecated and third-party URLs
+      .reject { |item| item["rank"] == "deprecated" || item["source"] == "wikidata:P3303" }
+      .group_by { |item| item["key"] }
+      .transform_keys { |key| key.sub(/^Key:/, "") }
+      # move preferred to the start of the array
+      .transform_values { |items| items.sort_by { |item| item["rank"] == "preferred" ? 0 : 1 }.uniq { |item| item["url"] } }
+      # exclude any that are ambiguous, i.e. the best and second-best have the same rank
+      .reject { |_key, value| value[1] && value[0]["rank"] == value[1]["rank"] }
+      # keep only the best match
+      .transform_values { |items| items[0]["url"] }
+  end
 end
diff --git a/test/lib/tag2link_test.rb b/test/lib/tag2link_test.rb
index 6d4fec58d..f7dbd6abd 100644
--- a/test/lib/tag2link_test.rb
+++ b/test/lib/tag2link_test.rb
@@ -15,4 +15,56 @@ class Tag2linkTest < ActiveSupport::TestCase
     url = Tag2link.link("wikidata", "Q936")
     assert_equal "https://www.wikidata.org/entity/Q936", url
   end
+
+  def test_build_dict_rejects_deprecated_and_third_party
+    data = [
+      { "key" => "Key:example", "url" => "http://example.com/$1", "rank" => "deprecated", "source" => "osmwiki:P8" },
+      { "key" => "Key:example2", "url" => "http://example2.com/$1", "rank" => "preferred", "source" => "wikidata:P3303" },
+      { "key" => "Key:example3", "url" => "http://example3.com/$1", "rank" => "preferred", "source" => "osmwiki:P8" }
+    ]
+    dict = Tag2link.build_dict(data)
+    assert_not_includes dict, "example"
+    assert_not_includes dict, "example2"
+    assert_includes dict, "example3"
+  end
+
+  def test_build_dict_chooses_single_preferred_item
+    data = [
+      { "key" => "Key:example", "url" => "http://example2.com/$1", "rank" => "normal", "source" => "osmwiki:P8" },
+      { "key" => "Key:example", "url" => "http://example.com/$1", "rank" => "preferred", "source" => "osmwiki:P8" }
+    ]
+    dict = Tag2link.build_dict(data)
+    assert_equal "http://example.com/$1", dict["example"]
+
+    data = [
+      { "key" => "Key:example", "url" => "http://example2.com/$1", "rank" => "preferred", "source" => "osmwiki:P8" }
+    ]
+    dict = Tag2link.build_dict(data)
+    assert_equal "http://example2.com/$1", dict["example"]
+  end
+
+  def test_build_dict_deduplicates_urls
+    data = [
+      { "key" => "Key:example", "url" => "http://example.com/$1", "rank" => "preferred", "source" => "osmwiki:P8" },
+      { "key" => "Key:example", "url" => "http://example.com/$1", "rank" => "normal", "source" => "wikidata:P1630" }
+    ]
+    dict = Tag2link.build_dict(data)
+    assert_equal "http://example.com/$1", dict["example"]
+  end
+
+  def test_build_dict_rejects_multiple_equally_preferred_items
+    data = [
+      { "key" => "Key:example", "url" => "http://example1.com/$1", "rank" => "preferred", "source" => "osmwiki:P8" },
+      { "key" => "Key:example", "url" => "http://example2.com/$1", "rank" => "preferred", "source" => "osmwiki:P8" }
+    ]
+    dict = Tag2link.build_dict(data)
+    assert_not_includes dict, "example"
+
+    data = [
+      { "key" => "Key:example", "url" => "http://example1.com/$1", "rank" => "normal", "source" => "osmwiki:P8" },
+      { "key" => "Key:example", "url" => "http://example2.com/$1", "rank" => "normal", "source" => "osmwiki:P8" }
+    ]
+    dict = Tag2link.build_dict(data)
+    assert_not_includes dict, "example"
+  end
 end
-- 
2.39.5