draftjs_html 0.24.0 → 0.25.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7c2703fb050f796abd4f8af851d16b654a2ab84f03e5f8865c92b373b2acfefa
-  data.tar.gz: eced6f85cfefd34f4c50c7fc383526675f26dd7c212a4f7de145721d254b1a28
+  metadata.gz: 20fc136c6cdbd1488b5ba808c3f55caca80f8f748241bc42b53028a544878542
+  data.tar.gz: a39a398da3a5c1513d62d008fb9f80f105a9d4b91137da4b22d072a09727fea1
 SHA512:
-  metadata.gz: 432d7b8900153f89f8ba506ded86208fe71a4b6b832911693059e0f57cb31b93412ef8fa5a40a8f7d857b370b65e8a523f29d0029dfa2ee8f917ebae2235827e
-  data.tar.gz: 69878d4330ef0d4947f1d95ed165720c0a59b8c0d54793d2e28aad8f5fea40754f95afacc1fc1a29c5b5b9f03e02399d9502924c7c580e7bc953b674f6292644
+  metadata.gz: c7926f0d00a1b322a1e76b8a37c36010b356d7cb3cde97662378476e02787ca8ea09227b64a496dd778ef630c70e80f7cc6db69d39f7c84c165b185a2b367627
+  data.tar.gz: 2cddf8a0b594755ae959935f8df954a7543d3208ab4c7ad55ab95b5fe28bb9f802ec140f11dcde9b728cc924e65d077afa5248734098c1013612ab3083e6ed28

data/README.md

@@ -229,15 +229,6 @@ The callable should return a Hash with symbol keys. The supported values are:
 - `data` (optional, default `{}`)
   - an arbitrary data-bag (Hash) of entity data
 
-#### `:is_semantic_markup:`
-
-Defaults to `true`.
-
-By setting to `false`, the user is stating they want to treat `div` tags as semantic,
-block-level tags. In some markup (emails, for example), there are no semantic tags
-(read, no `p` tags), so the only indications of whitespace and structure come from
-`div` tags. This flag will flush content wrapped in a `div` as a DraftJS block. 
-
 ## Development
 
 After checking out the repo, run `bin/setup` to install dependencies. Then, run

data/lib/draftjs_html/from_html/char_list.rb

@@ -60,6 +60,10 @@ module DraftjsHtml
         size > 0
       end
 
+      def more_than_whitespace?
+        !text.match?(/^[[:space:]]*$/)
+      end
+
       def atomic?
         @chars.any? && @chars.all?(&:atomic?)
       end

data/lib/draftjs_html/from_html/depth_stack.rb

@@ -1,16 +1,16 @@
 module DraftjsHtml
   class FromHtml < Nokogiri::XML::SAX::Document
     class DepthStack
-      def initialize(is_semantic_markup: true)
+      def initialize(options)
         @stack = []
         @nodes = []
         @list_depth = -1
         @active_styles = []
-        @is_semantic_markup = is_semantic_markup
+        @options = options
       end
 
       def push(tagname, attrs)
-        @stack << PendingBlock.from_tag(tagname, attrs, @nodes.dup, @list_depth, is_semantic_markup: @is_semantic_markup)
+        @stack << PendingBlock.from_tag(tagname, attrs, @nodes.dup, @list_depth, options: @options)
         track_block_node(tagname)
       end
 

data/lib/draftjs_html/from_html/elements.rb

@@ -1,8 +1,9 @@
 module DraftjsHtml
   class FromHtml < Nokogiri::XML::SAX::Document
     INLINE_STYLE_ELEMENTS = HtmlDefaults::HTML_STYLE_TAGS_TO_STYLE.keys.freeze
-    LIST_PARENT_ELEMENTS = %w[ol ul table].freeze
-    INLINE_NON_STYLE_ELEMENTS = %w[a abbr cite font img output q samp span thead tbody td time var].freeze
+    LIST_PARENT_ELEMENTS = %w[ol ul].freeze
+    INLINE_NON_STYLE_ELEMENTS = %w[a abbr cite font img output q samp span table thead tbody td time var].freeze
     BLOCK_CONTENT_ELEMENTS = %w[p dl h1 h2 h3 h4 h5 h6].freeze
+    FLUSH_BOUNDARIES = %w[OPENING div ol ul li tr].freeze
   end
 end

data/lib/draftjs_html/from_html/pending_block.rb

@@ -1,7 +1,7 @@
 module DraftjsHtml
   class FromHtml < Nokogiri::XML::SAX::Document
-    PendingBlock = Struct.new(:tagname, :attrs, :chars, :entities, :pending_entities, :parent_tagnames, :depth, :is_semantic_markup, keyword_init: true) do
-      def self.from_tag(name, attrs, parent_tagnames, depth, is_semantic_markup: true)
+    PendingBlock = Struct.new(:tagname, :attrs, :chars, :entities, :pending_entities, :parent_tagnames, :depth, :options, keyword_init: true) do
+      def self.from_tag(name, attrs, parent_tagnames, depth, options: {})
         self.new(
           tagname: name,
           attrs: attrs,
@@ -10,7 +10,7 @@ module DraftjsHtml
           pending_entities: [],
           depth: depth,
           parent_tagnames: parent_tagnames,
-          is_semantic_markup: is_semantic_markup,
+          options: options
         )
       end
 
@@ -23,9 +23,7 @@ module DraftjsHtml
       end
 
       def flushable?
-        %w[OPENING ol ul li table].include?(parent_tagnames.last) ||
-          (parent_tagnames.last == 'div' && tagname != 'div') ||
-          (!is_semantic_markup && tagname == 'div')
+        FLUSH_BOUNDARIES.include?(parent_tagnames.last)
       end
 
       def consume(other_pending_block)
@@ -35,21 +33,23 @@ module DraftjsHtml
       end
 
       def flush_to(draftjs)
-        if text_buffer.any?
-          text_buffer.each_line do |line|
-            block_type = line.atomic? ? 'atomic' : block_name
-            draftjs.typed_block(block_type, line.text, depth: [depth, 0].max)
+        text_buffer.each_line do |line|
+          block_type = line.atomic? ? 'atomic' : block_name
+          next unless should_flush_line?(line)
 
-            line.entity_ranges.each do |entity_range|
-              entity = entity_range.entity
-              draftjs.apply_entity entity[:type], entity_range.range, data: entity[:data], mutability: entity.fetch(:mutability, 'IMMUTABLE')
-            end
+          draftjs.typed_block(block_type, line.text, depth: [depth, 0].max)
 
-            line.style_ranges.each do |style_range|
-              draftjs.inline_style(style_range.style, style_range.range)
-            end
+          line.entity_ranges.each do |entity_range|
+            entity = entity_range.entity
+            draftjs.apply_entity entity[:type], entity_range.range, data: entity[:data], mutability: entity.fetch(:mutability, 'IMMUTABLE')
+          end
+
+          line.style_ranges.each do |style_range|
+            draftjs.inline_style(style_range.style, style_range.range)
           end
         end
+
+        self.text_buffer = CharList.new
       end
 
       def block_name
@@ -65,6 +65,13 @@ module DraftjsHtml
       def text_buffer=(other)
         self[:chars] = other
       end
+
+      def should_flush_line?(chars)
+        return true if chars.atomic?
+        return true unless options[:squeeze_whitespace_blocks]
+
+        options[:squeeze_whitespace_blocks] && chars.more_than_whitespace?
+      end
     end
   end
 end

data/lib/draftjs_html/from_html.rb

@@ -11,8 +11,8 @@ module DraftjsHtml
     def initialize(options = {})
       @draftjs = Draftjs::RawBuilder.new
       @parser = Nokogiri::HTML4::SAX::Parser.new(self)
-      @depth_stack = DepthStack.new(is_semantic_markup: options.fetch(:is_semantic_markup, true))
       @options = ensure_options!(options.dup)
+      @depth_stack = DepthStack.new(@options)
     end
 
     def convert(raw_html)
@@ -56,6 +56,7 @@ module DraftjsHtml
       when *FromHtml::LIST_PARENT_ELEMENTS
         @depth_stack.push_parent(name, attrs)
       else
+        @depth_stack.flush_to(@draftjs) if FromHtml::FLUSH_BOUNDARIES.include?(name)
         @depth_stack.push(name, attributes)
       end
 
@@ -101,6 +102,7 @@ module DraftjsHtml
         when 'img' then { type: 'IMAGE', mutability: 'IMMUTABLE', atomic: true, data: attrs }
         end
       }
+      opts[:squeeze_whitespace_blocks] ||= false
       opts
     end
   end

data/lib/draftjs_html/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module DraftjsHtml
-  VERSION = "0.24.0"
+  VERSION = "0.25.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: draftjs_html
 version: !ruby/object:Gem::Version
-  version: 0.24.0
+  version: 0.25.0
 platform: ruby
 authors:
 - TJ Taylor
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2022-12-19 00:00:00.000000000 Z
+date: 2022-12-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri