opener-property-tagger 3.1.1 → 3.3.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: 3c33a6b3d4f58d431c9b6eca1c4a4a66239fa651
-  data.tar.gz: 40f3ca343fe303d8b5bbe69576bafec5e0f555cf
+SHA256:
+  metadata.gz: 200701c1c3f256b0fbfa6fb284e321a16d10f399ced1fa750f7f9d4b66ed5a6f
+  data.tar.gz: 6da856c2c5beb56d778f12e93db99c6e5109728c1c89a6a8f8cceec488a556d2
 SHA512:
-  metadata.gz: 810e45722da0010b75c49add4d9f83cab411c06b72c458402a6e37997c607c49a107bc874b9dd6f06f5ea61f93c7c2e155e03e1c234ba6aa567bc86f9ca2a8c4
-  data.tar.gz: 01dd805a7355cd5109cc61085ca2a564bb5a81a5a1cdf8f90f2d8b199a18f1c4d69e9f263b8fc9872e918bd10e74e6a2ba9a18103e59287ccd52441cf5885488
+  metadata.gz: 5d7056c547b3c7b845b4d3ad7d9f43019629c57e447b7ba5d10ada342809e708f4b62a6ac468daef61617a62e60b35391b541e9f393722d26590b531f6292559
+  data.tar.gz: 7a79912c1317fb629c95244c8f0bb9cf73690bf3b194984c8a7c818a6cc3123f8eaa5957d6bbe32db01f000457472a9db91ba7f1c3b8b556fdfcd440ff033c1a

data/lib/opener/property_tagger.rb

@@ -2,13 +2,18 @@ require 'open3'
 require 'slop'
 require 'oga'
 require 'monitor'
+require 'httpclient'
+require 'hashie'
+require 'json'
+require 'active_support/all'
 
 require 'rexml/document'
 require 'rexml/formatters/pretty'
 
 require_relative 'property_tagger/version'
 require_relative 'property_tagger/cli'
-require_relative 'property_tagger/aspects_cache'
+require_relative 'property_tagger/remote_aspects_cache'
+require_relative 'property_tagger/file_aspects_cache'
 require_relative 'property_tagger/processor'
 
 module Opener
@@ -43,14 +48,17 @@ module Opener
     # @return [String]
     #
     def path
-      path = options[:resource_path] || ENV['RESOURCE_PATH'] ||
+      return @path if @path
+
+      @path = options[:resource_path] || ENV['RESOURCE_PATH'] ||
         ENV['PROPERTY_TAGGER_LEXICONS_PATH']
+      return unless @path
 
-      unless path
-        raise ArgumentError, 'No lexicon path provided'
-      end
+      @path = File.expand_path @path
+    end
 
-      return File.expand_path(path)
+    def remote_url
+      @remote_url ||= ENV['PROPERTY_TAGGER_LEXICONS_URL']
     end
 
     ##
@@ -59,11 +67,18 @@ module Opener
     # @param [String] input
     # @return [String]
     #
-    def run(input)
+    def run input, params = {}
       timestamp = !options[:no_time]
 
-      return Processor.new(input, path, timestamp, options[:pretty]).process
+      Processor.new(input,
+        params:    params,
+        url:       remote_url,
+        path:      path,
+        timestamp: timestamp,
+        pretty:    options[:pretty],
+      ).process
     end
-  end # PolarityTagger
-end # Opener
+
+  end
+end
 

data/lib/opener/property_tagger/{aspects_cache.rb → file_aspects_cache.rb}

@@ -3,7 +3,8 @@ module Opener
     ##
     # Thread-safe cache for storing the contents of aspect files.
     #
-    class AspectsCache
+    class FileAspectsCache
+
       include MonitorMixin
 
       def initialize
@@ -22,8 +23,6 @@ module Opener
         synchronize do
           @cache[path] = load_aspects(path) unless @cache.key?(path)
         end
-
-        return @cache[path]
       end
 
       alias_method :get, :[]
@@ -44,6 +43,7 @@ module Opener
 
         return mapping
       end
-    end # AspectsCache
-  end # PropertyTagger
-end # Opener
+
+    end
+  end
+end

data/lib/opener/property_tagger/processor.rb

@@ -4,14 +4,18 @@ module Opener
     # Class that applies property tagging to a given input KAF file.
     #
     class Processor
-      attr_accessor :document, :aspects_path, :timestamp, :pretty
+
+      attr_accessor :document
+      attr_accessor :aspects, :aspects_path, :aspects_url
+      attr_accessor :timestamp, :pretty
 
       ##
       # Global cache used for storing loaded aspects.
       #
-      # @return [Opener::PropertyTagger::AspectsCache.new]
+      # @return [Opener::PropertyTagger::FileAspectsCache.new]
       #
-      ASPECTS_CACHE = AspectsCache.new
+      FILE_ASPECTS_CACHE   = FileAspectsCache.new
+      REMOTE_ASPECTS_CACHE = RemoteAspectsCache.new
 
       ##
       # @param [String|IO] file The KAF file/input to process.
@@ -20,13 +24,19 @@ module Opener
       # @param [TrueClass|FalseClass] pretty Enable pretty formatting, disabled
       #  by default due to the performance overhead.
       #
-      def initialize(file, aspects_path, timestamp = true, pretty = false)
-        @document     = Oga.parse_xml(file)
-        @aspects_path = aspects_path
+      def initialize file, params: {}, url: nil, path: nil, timestamp: true, pretty: false
+        @document     = Oga.parse_xml file
+        raise 'Error parsing input. Input is required to be KAF' unless is_kaf?
         @timestamp    = timestamp
         @pretty       = pretty
 
-        raise 'Error parsing input. Input is required to be KAF' unless is_kaf?
+        @params       = params
+        @cache_keys   = params[:cache_keys] || {lang: language}
+        @remote       = !url.nil?
+        @aspects_path = path
+        @aspects_url  = url
+
+        @aspects = if @remote then REMOTE_ASPECTS_CACHE[**@cache_keys].aspects else FILE_ASPECTS_CACHE[aspects_file] end
       end
 
       ##
@@ -50,13 +60,6 @@ module Opener
         return pretty ? pretty_print(document) : document.to_xml
       end
 
-      ##
-      # @return [Hash]
-      #
-      def aspects
-        return ASPECTS_CACHE[aspects_file]
-      end
-
       ##
       # Get the language of the input file.
       #
@@ -228,9 +231,9 @@ module Opener
       # @return [String]
       #
       def aspects_file
-        return @aspects_file ||=
-          File.expand_path("#{aspects_path}/#{language}.txt", __FILE__)
+        @aspects_file ||= File.expand_path "#{aspects_path}/#{language}.txt", __FILE__
       end
-    end # Processor
-  end # PropertyTagger
-end # Opener
+
+    end
+  end
+end

data/lib/opener/property_tagger/remote_aspects_cache.rb

@@ -0,0 +1,60 @@
+module Opener
+  class PropertyTagger
+    ##
+    # Thread-safe cache for storing the contents of remote aspects.
+    #
+    class RemoteAspectsCache
+
+      include MonitorMixin
+
+      UPDATE_INTERVAL = (ENV['CACHE_EXPIRE_MINS']&.to_i || 5).minutes
+
+      def initialize
+        super
+
+        @url   = ENV['PROPERTY_TAGGER_LEXICONS_URL']
+        @cache = {}
+      end
+
+      def [] **params
+        synchronize do
+          existing = @cache[params]
+          break existing if existing and existing.from > UPDATE_INTERVAL.ago
+          @cache[params] = cache_update existing, **params
+        end
+      end
+      alias_method :get, :[]
+
+      def cache_update existing = nil, **params
+        from     = Time.now
+        lexicons = load_aspects cache: existing, **params
+
+        if existing and lexicons.blank?
+          existing.from = from
+          return existing
+        end
+
+        Hashie::Mash.new(
+          aspects: lexicons,
+          from:    from,
+        )
+      end
+
+      def load_aspects lang:, cache:, **params
+        url  = "#{@url}&language_code=#{lang}&#{params.to_query}"
+        url += "&if_updated_since=#{cache.from.utc.iso8601}" if cache
+        puts "#{lang}: loading aspects from #{url}"
+
+        lexicons = JSON.parse HTTPClient.new.get(url).body
+        lexicons = lexicons['data'].map{ |l| Hashie::Mash.new l }
+        mapping  = Hash.new{ |hash, key| hash[key] = [] }
+        lexicons.each do |l|
+          mapping[l.lemma.to_sym] << l.aspect
+        end
+
+        mapping
+      end
+
+    end
+  end
+end

data/lib/opener/property_tagger/version.rb

@@ -1,5 +1,7 @@
 module Opener
   class PropertyTagger
-    VERSION = '3.1.1'
-  end # PropertyTagger
-end # Opener
+
+    VERSION = '3.3.3'
+
+  end
+end

data/opener-property-tagger.gemspec

@@ -29,6 +29,9 @@ Gem::Specification.new do |gem|
   gem.add_dependency 'opener-core', '~> 2.2'
 
   gem.add_dependency 'oga', ['~> 1.0', '>= 1.3.1']
+  gem.add_dependency 'httpclient'
+  gem.add_dependency 'hashie'
+  gem.add_dependency 'activesupport'
 
   gem.add_development_dependency 'rspec', '~> 3.0'
   gem.add_development_dependency 'cucumber'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: opener-property-tagger
 version: !ruby/object:Gem::Version
-  version: 3.1.1
+  version: 3.3.3
 platform: ruby
 authors:
 - development@olery.com
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2015-09-07 00:00:00.000000000 Z
+date: 2020-11-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: opener-daemons
@@ -72,6 +72,48 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 1.3.1
+- !ruby/object:Gem::Dependency
+  name: httpclient
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: hashie
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: activesupport
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement
@@ -129,11 +171,11 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '2.0'
 description: Property tagger for hotels in Dutch and English.
-email: 
+email:
 executables:
+- property-tagger-server
 - property-tagger
 - property-tagger-daemon
-- property-tagger-server
 extensions: []
 extra_rdoc_files: []
 files:
@@ -145,10 +187,11 @@ files:
 - config.ru
 - exec/property-tagger.rb
 - lib/opener/property_tagger.rb
-- lib/opener/property_tagger/aspects_cache.rb
 - lib/opener/property_tagger/cli.rb
+- lib/opener/property_tagger/file_aspects_cache.rb
 - lib/opener/property_tagger/processor.rb
 - lib/opener/property_tagger/public/markdown.css
+- lib/opener/property_tagger/remote_aspects_cache.rb
 - lib/opener/property_tagger/server.rb
 - lib/opener/property_tagger/version.rb
 - lib/opener/property_tagger/views/index.erb
@@ -160,7 +203,7 @@ homepage: http://opener-project.github.com/
 licenses:
 - Apache 2.0
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -175,10 +218,9 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.4.8
-signing_key: 
+rubyforge_project:
+rubygems_version: 2.7.8
+signing_key:
 specification_version: 4
 summary: Property tagger for hotels in Dutch and English.
 test_files: []
-has_rdoc: