Allow passing multiple pattern databases (#70)

- Support `patterns_path` argument but deprecate `patterns_path` attribute accessor
- Add new `patterns_paths` array argument and accessor to enable loading multiple patterns files

Fixes #11

Co-authored-by: Mikhail Doronin <mikhail.doronin@capitainetrain.com>

README.md

@@ -100,6 +100,11 @@ or when instantiating a `UserAgentParser::Parser`:
 UserAgentParser::Parser.new(patterns_path: '/some/path/to/regexes.yaml').parse(ua_string)
 ```
 
+Extending the standard database is possible by providing multiple files in `patterns_paths` (plural) array argument:
+```ruby
+UserAgentParser::Parser.new(patterns_paths: [UserAgentParser::DefaultPatternsPath, '/some/path/to/regexes.yaml'])
+```
+
 ## Command line tool
 
 The gem incldes a `user_agent_parser` bin command which will read from

lib/user_agent_parser.rb

@@ -10,7 +10,7 @@ module UserAgentParser
   DefaultPatternsPath = File.join(File.dirname(__FILE__), '../vendor/uap-core/regexes.yaml')
 
   # Parse the given +user_agent_string+, returning a +UserAgent+
-  def self.parse(user_agent_string, options = {})
-    Parser.new(options).parse(user_agent_string)
+  def self.parse(user_agent_string, **args)
+    Parser.new(**args).parse(user_agent_string)
   end
 end

lib/user_agent_parser/parser.rb

@@ -4,6 +4,8 @@
 
 module UserAgentParser
   class Parser
+    extend Gem::Deprecate
+
     FAMILY_REPLACEMENT_KEYS = %w[
       family_replacement
       v1_replacement
@@ -22,11 +24,13 @@ class Parser
 
     private_constant :FAMILY_REPLACEMENT_KEYS, :OS_REPLACEMENT_KEYS
 
-    attr_reader :patterns_path
+    attr_reader :patterns_paths
+
+    def initialize(patterns_path: nil, patterns_paths: [])
+      @patterns_paths = [patterns_path, *patterns_paths].compact
+      @patterns_paths = [UserAgentParser::DefaultPatternsPath] if @patterns_paths.empty?
 
-    def initialize(options = {})
-      @patterns_path = options[:patterns_path] || UserAgentParser::DefaultPatternsPath
-      @ua_patterns, @os_patterns, @device_patterns = load_patterns(patterns_path)
+      @ua_patterns, @os_patterns, @device_patterns = load_patterns(@patterns_paths)
     end
 
     def parse(user_agent)
@@ -35,9 +39,23 @@ def parse(user_agent)
       parse_ua(user_agent, os, device)
     end
 
+    def patterns_path
+      patterns_paths.first
+    end
+    deprecate :patterns_path, :patterns_paths, 2022, 12
+
     private
 
-    def load_patterns(path)
+    def load_patterns(patterns_paths)
+      patterns_paths.each_with_object([[], [], []]) do |path, patterns|
+        ua_patterns, os_patterns, device_patterns = load_patterns_file(path)
+        patterns[0] += ua_patterns
+        patterns[1] += os_patterns
+        patterns[2] += device_patterns
+      end
+    end
+
+    def load_patterns_file(path)
       yml = begin
         YAML.load_file(path, freeze: true)
       rescue ArgumentError
@@ -88,24 +106,11 @@ def parse_device(user_agent)
       end
     end
 
-    if Gem::Version.new(RUBY_VERSION) >= Gem::Version.new('2.4.0')
-      def first_pattern_match(patterns, value)
-        patterns.each do |pattern|
-          if pattern[:regex].match?(value)
-            return [pattern, pattern[:regex].match(value)]
-          end
-        end
-        nil
-      end
-    else
-      def first_pattern_match(patterns, value)
-        patterns.each do |pattern|
-          if (match = pattern[:regex].match(value))
-            return [pattern, match]
-          end
-        end
-        nil
+    def first_pattern_match(patterns, value)
+      patterns.each do |pattern|
+        return [pattern, pattern[:regex].match(value)] if pattern[:regex].match?(value)
       end
+      nil
     end
 
     def user_agent_from_pattern_match(pattern, match, os = nil, device = nil)

spec/custom_regexes.yaml

@@ -1,17 +1,17 @@
 user_agent_parsers:
-  - regex: '.*'
+  - regex: 'Any.*'
     family_replacement: 'Custom browser'
     v1_replacement: '1'
     v2_replacement: '2'
     v3_replacement: '3'
     v4_replacement: '4'
 
 os_parsers:
-  - regex: '.*'
+  - regex: 'Any.*'
     os_replacement: 'Custom OS'
     os_v1_replacement: '1'
     os_v2_replacement: '2'
 
 device_parsers:
-  - regex: '.*'
+  - regex: 'Any.*'
     device_replacement: 'Custom device'

spec/other_regexes.yaml

@@ -0,0 +1,17 @@
+user_agent_parsers:
+  - regex: 'Other.*'
+    family_replacement: 'Other browser'
+    v1_replacement: '1'
+    v2_replacement: '2'
+    v3_replacement: '3'
+    v4_replacement: '4'
+
+os_parsers:
+  - regex: 'Other.*'
+    os_replacement: 'Other OS'
+    os_v1_replacement: '1'
+    os_v2_replacement: '2'
+
+device_parsers:
+  - regex: 'Other.*'
+    device_replacement: 'Other device'

spec/parser_spec.rb

@@ -57,6 +57,10 @@ def custom_patterns_path
     File.join(File.dirname(__FILE__), 'custom_regexes.yaml')
   end
 
+  def other_patterns_path
+    File.join(File.dirname(__FILE__), 'other_regexes.yaml')
+  end
+
   describe '::parse' do
     it 'parses a UA' do
       ua = UserAgentParser.parse('Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-us) AppleWebKit/418.8 (KHTML, like Gecko) Safari/419.3')
@@ -69,10 +73,13 @@ def custom_patterns_path
   end
 
   describe '#initialize with a custom patterns path' do
-    it 'uses the custom patterns' do
+    it 'accepts a single patterns_path string' do
       parser = UserAgentParser::Parser.new(patterns_path: custom_patterns_path)
       ua = parser.parse('Any user agent string')
 
+      _(parser.patterns_paths).must_equal([custom_patterns_path])
+      _(parser.patterns_path).must_equal(custom_patterns_path)
+
       _(ua.family).must_equal('Custom browser')
       _(ua.version.major).must_equal('1')
       _(ua.version.minor).must_equal('2')
@@ -85,6 +92,20 @@ def custom_patterns_path
 
       _(ua.device.family).must_equal('Custom device')
     end
+
+    it 'accepts patterns_paths array' do
+      patterns_paths = [custom_patterns_path, other_patterns_path]
+      parser = UserAgentParser::Parser.new(patterns_paths: patterns_paths)
+
+      _(parser.patterns_paths).must_equal(patterns_paths)
+      _(parser.patterns_path).must_equal(custom_patterns_path)
+
+      ua = parser.parse('Any user agent string')
+      oua = parser.parse('Other user agent string')
+
+      _(ua.family).must_equal('Custom browser')
+      _(oua.family).must_equal('Other browser')
+    end
   end
 
   describe '#parse' do