html2rss 0.23.0 → 0.25.0

data/lib/html2rss/html/article_extractor/image_extractor.rb

@@ -32,27 +32,16 @@ module Html2rss
         # @see <https://developer.mozilla.org/en-US/docs/Learn/HTML/Multimedia_and_embedding/Responsive_images>
         # @see <https://developer.mozilla.org/en-US/docs/Web/HTML/Element/img#srcset>
         # @see <https://developer.mozilla.org/en-US/docs/Web/HTML/Element/picture>
-        def self.from_source(article_tag) # rubocop:disable Metrics/AbcSize
-          hash = article_tag.css('img[srcset], picture > source[srcset]').flat_map do |source|
-            source['srcset'].to_s.scan(/(\S+)\s+(\d+w|\d+h)[\s,]?/).map do |url, width|
-              next if url.nil? || url.start_with?('data:')
-
-              width_value = width.to_i.zero? ? 0 : width.scan(/\d+/).first.to_i
-
-              [width_value, url.strip]
-            end
-          end.compact.to_h
-
-          hash[hash.keys.max]
+        def self.from_source(article_tag)
+          srcsets = article_tag.css('img[srcset], picture > source[srcset]').map { |source| source['srcset'] }
+          ArticleRules::Image.largest_from_srcsets(srcsets)
         end
 
         # @param article_tag [Nokogiri::XML::Element] article container node
         # @return [String, nil] best style-based background image URL
         def self.from_style(article_tag)
-          article_tag.css('[style*="url"]')
-                     .filter_map { |tag| tag['style'][/url\(['"]?(.*?)['"]?\)/, 1] }
-                     .reject { |src| src.start_with?('data:') }
-                     .max_by(&:size)
+          styles = article_tag.css('[style*="url"]').map { |tag| tag['style'] }
+          ArticleRules::Image.best_from_styles(styles)
         end
       end
     end

data/lib/html2rss/html/article_rules/category.rb

@@ -0,0 +1,55 @@
+# frozen_string_literal: true
+
+module Html2rss
+  module Html
+    module ArticleRules
+      ##
+      # Shared category term vocabulary and text accumulation (DOM walk stays in adapters).
+      module Category
+        # Class/data attribute tokens that mark category metadata.
+        CATEGORY_TERMS = %w[
+          category categories tag tags topic topics section sections
+          label labels theme themes subject subjects
+        ].freeze
+
+        # Regex matching category-related attribute or class names.
+        CATEGORY_ATTR_PATTERN = /#{CATEGORY_TERMS.join('|')}/i
+
+        class << self
+          ##
+          # @param name [String]
+          # @return [Boolean]
+          def attr_name_match?(name)
+            name.to_s.match?(CATEGORY_ATTR_PATTERN)
+          end
+
+          ##
+          # @param class_attr [String, nil]
+          # @return [Boolean]
+          def class_match?(class_attr)
+            class_attr.to_s.match?(CATEGORY_ATTR_PATTERN)
+          end
+
+          ##
+          # @param categories [Set<String>]
+          # @param text [String, nil]
+          # @return [void]
+          def add_text!(categories, text)
+            value = text.to_s.strip
+            categories.add(value) unless value.empty?
+          end
+
+          ##
+          # @param categories [Set<String>]
+          # @param text [String, nil]
+          # @return [void]
+          def add_split_text!(categories, text)
+            return unless text
+
+            text.split(/\n+/).each { |line| add_text!(categories, line) }
+          end
+        end
+      end
+    end
+  end
+end

data/lib/html2rss/html/article_rules/date.rb

@@ -0,0 +1,25 @@
+# frozen_string_literal: true
+
+module Html2rss
+  module Html
+    module ArticleRules
+      ##
+      # DOM-agnostic datetime aggregation for article published_at.
+      module Date
+        class << self
+          ##
+          # @param datetime_strings [Array<String, nil>] raw datetime attribute values
+          # @return [DateTime, nil] earliest successfully parsed instant
+          def earliest(datetime_strings)
+            times = Array(datetime_strings).filter_map do |value|
+              DateTime.parse(value.to_s)
+            rescue ArgumentError, TypeError
+              nil
+            end
+            times.min
+          end
+        end
+      end
+    end
+  end
+end

data/lib/html2rss/html/article_rules/enclosure.rb

@@ -0,0 +1,72 @@
+# frozen_string_literal: true
+
+module Html2rss
+  module Html
+    module ArticleRules
+      ##
+      # DOM-agnostic enclosure URL/type normalization.
+      module Enclosure
+        # Href path suffixes treated as download/archive enclosure links.
+        ARCHIVE_HREF_SUFFIXES = %w[.pdf .zip .tar.gz .tgz].freeze
+
+        class << self
+          ##
+          # @param href [String]
+          # @return [Boolean]
+          def archive_href?(href)
+            ARCHIVE_HREF_SUFFIXES.any? { |suffix| href.to_s.end_with?(suffix) }
+          end
+
+          ##
+          # @param src [String]
+          # @param base_url [String, Html2rss::Url]
+          # @return [Hash{Symbol => Object}, nil]
+          def from_image(src, base_url)
+            return if src.to_s.empty?
+
+            abs = Url.from_relative(src, base_url)
+            { url: abs, type: Article::Enclosure.guess_content_type_from_url(abs, default: 'image/jpeg') }
+          end
+
+          ##
+          # @param src [String]
+          # @param type [String, nil]
+          # @param base_url [String, Html2rss::Url]
+          # @return [Hash{Symbol => Object}, nil]
+          def from_media(src, type, base_url)
+            return if src.to_s.empty?
+
+            { url: Url.from_relative(src, base_url), type: }
+          end
+
+          ##
+          # @param src [String]
+          # @param base_url [String, Html2rss::Url]
+          # @return [Hash{Symbol => Object}, nil]
+          def from_iframe(src, base_url)
+            return if src.to_s.empty?
+
+            abs = Url.from_relative(src, base_url)
+            { url: abs, type: Article::Enclosure.guess_content_type_from_url(abs, default: 'text/html') }
+          end
+
+          ##
+          # @param href [String]
+          # @param base_url [String, Html2rss::Url]
+          # @return [Hash{Symbol => Object}, nil]
+          def from_anchor(href, base_url)
+            return if href.to_s.empty?
+
+            abs = Url.from_relative(href, base_url)
+            type = if href.end_with?('.pdf')
+                     Article::Enclosure.guess_content_type_from_url(abs)
+                   else
+                     'application/zip'
+                   end
+            { url: abs, type: }
+          end
+        end
+      end
+    end
+  end
+end

data/lib/html2rss/html/article_rules/image.rb

@@ -0,0 +1,109 @@
+# frozen_string_literal: true
+
+module Html2rss
+  module Html
+    module ArticleRules
+      ##
+      # DOM-agnostic image URL selection (srcset sizing, CSS background urls).
+      #
+      # Uses string scans instead of Regexp: CodeQL rb/polynomial-redos flags the
+      # prior patterns on uncontrolled HTML attributes, and length guards alone
+      # do not clear the query.
+      module Image
+        # Skip absurdly large attributes before scanning.
+        MAX_ATTR_CHARS = 4_096
+        # Optional quoting around a CSS `url(...)` value.
+        QUOTES = ['"', "'"].freeze
+        # Srcset width/height descriptor suffixes.
+        UNITS = %w[w h].freeze
+        # HTML/CSS whitespace treated as srcset token separators.
+        WS = [' ', "\t", "\n", "\r", "\f"].freeze
+
+        class << self
+          ##
+          # @param srcset_strings [Array<String>] raw srcset attribute values
+          # @return [String, nil] URL of the largest width/height candidate
+          def largest_from_srcsets(srcset_strings)
+            by_size = {}
+            Array(srcset_strings).each { |srcset| merge_srcset!(by_size, srcset) }
+            by_size[by_size.keys.max]
+          end
+
+          ##
+          # @param style_strings [Array<String>] CSS style attribute values
+          # @return [String, nil] longest non-data background url
+          def best_from_styles(style_strings)
+            Array(style_strings)
+              .filter_map { |style| style_url(style) }
+              .reject { |src| src.start_with?('data:') }
+              .max_by(&:size)
+          end
+
+          private
+
+          def merge_srcset!(by_size, srcset)
+            s = srcset.to_s
+            return if s.length > MAX_ATTR_CHARS
+
+            each_srcset_pair(s) do |url, size|
+              next if url.start_with?('data:')
+
+              by_size[size] = url
+            end
+          end
+
+          # URL token + `Nw`/`Nh` pairs. Commas may appear inside URLs; a candidate
+          # boundary is whitespace before a width/height descriptor.
+          def each_srcset_pair(string) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
+            i = 0
+            len = string.length
+            while i < len
+              i += 1 while i < len && sep?(string[i])
+              break if i >= len
+
+              start = i
+              i += 1 while i < len && !ws?(string[i])
+              url = string[start, i - start]
+
+              i += 1 while i < len && ws?(string[i])
+              size_start = i
+              i += 1 while i < len && string[i].between?('0', '9')
+              next unless i > size_start && i < len && UNITS.include?(string[i])
+
+              yield url, string[size_start, i - size_start].to_i
+              i += 1
+            end
+          end
+
+          def style_url(style)
+            s = style.to_s
+            return if s.length > MAX_ATTR_CHARS
+
+            pos = s.index('url(')
+            return unless pos
+
+            extract_url_body(s, pos + 4)
+          end
+
+          def extract_url_body(string, index)
+            quote = string[index]
+            if QUOTES.include?(quote)
+              slice_until(string, index + 1, quote)
+            else
+              slice_until(string, index, ')')
+            end
+          end
+
+          def slice_until(string, start, closer)
+            close = string.index(closer, start)
+            string[start, close - start] if close
+          end
+
+          def sep?(char) = ws?(char) || char == ','
+
+          def ws?(char) = WS.include?(char)
+        end
+      end
+    end
+  end
+end

data/lib/html2rss/html/article_rules.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+module Html2rss
+  module Html
+    ##
+    # DOM-agnostic article field rules shared by Nokogiri and SST adapters.
+    module ArticleRules
+    end
+  end
+end

data/lib/html2rss/html/rendering/audio_renderer.rb

@@ -1,12 +1,12 @@
 # frozen_string_literal: true
 
-require 'cgi'
-
 module Html2rss
   module Html
     module Rendering
       # Renders an HTML <audio> tag from a URL and type.
       class AudioRenderer
+        include EscapedAttributes
+
         # @param url [String, Html2rss::Url] media URL for the audio source
         # @param type [String] MIME type for the audio source
         def initialize(url:, type:)
@@ -22,16 +22,6 @@ module Html2rss
             '</audio>'
           ].join
         end
-
-        private
-
-        def escaped_url
-          CGI.escapeHTML(@url.to_s)
-        end
-
-        def escaped_type
-          CGI.escapeHTML(@type.to_s)
-        end
       end
     end
   end

data/lib/html2rss/html/rendering/escaped_attributes.rb

@@ -0,0 +1,27 @@
+# frozen_string_literal: true
+
+require 'cgi'
+
+module Html2rss
+  module Html
+    module Rendering
+      ##
+      # Shared HTML attribute escaping for media renderers.
+      module EscapedAttributes
+        private
+
+        def escaped_url
+          CGI.escapeHTML(@url.to_s)
+        end
+
+        def escaped_type
+          CGI.escapeHTML(@type.to_s)
+        end
+
+        def escaped_title
+          CGI.escapeHTML(@title.to_s)
+        end
+      end
+    end
+  end
+end

data/lib/html2rss/html/rendering/image_renderer.rb

@@ -1,12 +1,12 @@
 # frozen_string_literal: true
 
-require 'cgi'
-
 module Html2rss
   module Html
     module Rendering
       # Renders an HTML <img> tag from a URL and title.
       class ImageRenderer
+        include EscapedAttributes
+
         # @param url [String, Html2rss::Url] image URL for the src attribute
         # @param title [String, nil] title/alt text for the image
         def initialize(url:, title:)
@@ -27,16 +27,6 @@ module Html2rss
           ]
           "<img #{attributes.join(' ')}>"
         end
-
-        private
-
-        def escaped_url
-          CGI.escapeHTML(@url.to_s)
-        end
-
-        def escaped_title
-          CGI.escapeHTML(@title.to_s)
-        end
       end
     end
   end

data/lib/html2rss/html/rendering/pdf_renderer.rb

@@ -1,12 +1,12 @@
 # frozen_string_literal: true
 
-require 'cgi'
-
 module Html2rss
   module Html
     module Rendering
       # Renders an HTML <iframe> for PDF documents.
       class PdfRenderer
+        include EscapedAttributes
+
         # @param url [String, Html2rss::Url] PDF URL rendered in the iframe
         def initialize(url:)
           @url = url
@@ -24,12 +24,6 @@ module Html2rss
           ]
           "<iframe #{attributes.join(' ')}></iframe>"
         end
-
-        private
-
-        def escaped_url
-          CGI.escapeHTML(@url.to_s)
-        end
       end
     end
   end

data/lib/html2rss/html/rendering/video_renderer.rb

@@ -1,12 +1,12 @@
 # frozen_string_literal: true
 
-require 'cgi'
-
 module Html2rss
   module Html
     module Rendering
       # Renders an HTML <video> tag from a URL and type.
       class VideoRenderer
+        include EscapedAttributes
+
         # @param url [String, Html2rss::Url] media URL for the video source
         # @param type [String] MIME type for the video source
         def initialize(url:, type:)
@@ -22,16 +22,6 @@ module Html2rss
             '</video>'
           ].join
         end
-
-        private
-
-        def escaped_url
-          CGI.escapeHTML(@url.to_s)
-        end
-
-        def escaped_type
-          CGI.escapeHTML(@type.to_s)
-        end
       end
     end
   end