stationery 0.9.0 → 0.10.1

data/lib/stationery/text/breaks.rb

@@ -0,0 +1,123 @@
+# frozen_string_literal: true
+
+module Stationery
+  module Text
+    # Where a line may break inside a word: between two ideographic
+    # characters (UAX #14 class ID: CJK ideographs, kana, Hangul, fullwidth
+    # forms), never before a closing mark or a character that must not start
+    # a line (、。」ー and the small kana: classes CL, CP, EX and NS), never
+    # after an opening bracket (「（: class OP). A word is cut into units at
+    # those points; Latin text stays one unit, so its breaking is unchanged.
+    module Breaks
+      ZERO_WIDTH_SPACE = "​"
+
+      IDEOGRAPHIC = [
+        0x1100..0x11FF,   # Hangul Jamo
+        0x2E80..0x2FDF,   # CJK radicals
+        0x3005..0x3007,   # 々 〆 〇
+        0x3021..0x3029,   # Hangzhou numerals
+        0x3031..0x3035,   # kana repeat marks
+        0x3038..0x303C,
+        0x3041..0x309A,   # Hiragana (the small kana are taken out below)
+        0x30A1..0x30FA,   # Katakana
+        0x3100..0x312F,   # Bopomofo
+        0x3130..0x318F,   # Hangul compatibility Jamo
+        0x3190..0x31FF,   # Kanbun, Bopomofo extended, CJK strokes, Katakana phonetic extensions
+        0x3200..0x33FF,   # enclosed CJK, CJK compatibility
+        0x3400..0x4DBF,   # CJK extension A
+        0x4E00..0x9FFF,   # CJK unified ideographs
+        0xA000..0xA4CF,   # Yi
+        0xA960..0xA97F,   # Hangul Jamo extended A
+        0xAC00..0xD7AF,   # Hangul syllables
+        0xF900..0xFAFF,   # CJK compatibility ideographs
+        0xFE30..0xFE4F,   # CJK compatibility forms
+        0xFF10..0xFF19,   # fullwidth digits
+        0xFF21..0xFF3A, 0xFF41..0xFF5A, # fullwidth Latin
+        0xFF66..0xFF6F, 0xFF71..0xFF9D, # halfwidth Katakana (ｰ and the small ones are NS)
+        0x1B000..0x1B16F, # Kana supplement and extended
+        0x20000..0x3134F  # CJK extensions B–G
+      ].freeze
+
+      # Never at the start of a line: closing brackets and quotes, sentence
+      # and clause punctuation, prolonged sound and iteration marks, small kana.
+      CLOSING = [
+        0x3001..0x3003, 0x3009, 0x300B, 0x300D, 0x300F, 0x3011, 0x3015, 0x3017, 0x3019, 0x301B, 0x301C,
+        0x301F, 0x303B, 0x309B..0x309E, 0x30A0, 0x30FB..0x30FE,
+        0x3041, 0x3043, 0x3045, 0x3047, 0x3049, 0x3063, 0x3083, 0x3085, 0x3087, 0x308E, 0x3095, 0x3096,
+        0x30A1, 0x30A3, 0x30A5, 0x30A7, 0x30A9, 0x30C3, 0x30E3, 0x30E5, 0x30E7, 0x30EE, 0x30F5, 0x30F6,
+        0x31F0..0x31FF,
+        0xFE50..0xFE52, 0xFE54..0xFE57, 0xFE5A, 0xFE5C, 0xFE5E,
+        0xFF01, 0xFF09, 0xFF0C, 0xFF0E, 0xFF1A, 0xFF1B, 0xFF1F, 0xFF3D, 0xFF5D, 0xFF60, 0xFF61, 0xFF63,
+        0xFF64, 0xFF65, 0xFF67..0xFF6F, 0xFF70, 0xFF9E, 0xFF9F,
+        0x2025, 0x2026
+      ].freeze
+
+      # Never at the end of a line: opening brackets and quotes.
+      OPENING = [
+        0x3008, 0x300A, 0x300C, 0x300E, 0x3010, 0x3014, 0x3016, 0x3018, 0x301A, 0x301D,
+        0xFE59, 0xFE5B, 0xFE5D,
+        0xFF08, 0xFF3B, 0xFF5B, 0xFF5F, 0xFF62
+      ].freeze
+
+      CLOSING_CODES = CLOSING.flat_map { |entry| Array(entry) }.to_h { |code| [code, true] }.freeze
+      OPENING_CODES = OPENING.flat_map { |entry| Array(entry) }.to_h { |code| [code, true] }.freeze
+
+      module_function
+
+      def ideographic?(char)
+        code = char.ord
+        !CLOSING_CODES.key?(code) && IDEOGRAPHIC.any? { |range| range.cover?(code) }
+      end
+
+      def closing?(char) = CLOSING_CODES.key?(char.ord)
+      def opening?(char) = OPENING_CODES.key?(char.ord)
+
+      # Whether `text` holds any character these rules know about, the cheap
+      # test that keeps Latin text on the fast path.
+      def cjk?(text) = text.match?(CJK)
+
+      CJK = /[ᄀ-ᇿ⺀-⿟、-ヿ㄀-鿿ꀀ-꓏ꥠ-꥿가-힯豈-﫿︰-﹞！-ﾟ\u{1B000}-\u{1B16F}\u{20000}-\u{3134F}‥…]/
+
+      # The word's break units in order, each `[text, glued]`: `glued` is true
+      # for a unit that must stay with the unit before it (it starts with a
+      # closing mark), so the caller joins it to whatever came before, even
+      # across styled runs. A word without CJK characters is one unit.
+      def units(word)
+        return [[word, false]] unless cjk?(word)
+
+        units = []
+        current = +""
+        glued = false
+        open = false # current holds opening brackets only, waiting for what they open
+        word.each_char do |char|
+          if closing?(char)
+            glued = true if current.empty?
+          elsif starts_unit?(char, current, open)
+            units << [current.freeze, glued] unless current.empty?
+            current = +""
+            glued = false
+          end
+          current << char
+          open = opening?(char) && (open || current.length == 1)
+        end
+        units << [current.freeze, glued] unless current.empty?
+        units
+      end
+
+      # Whether `char` opens a new unit after `current`: an opening bracket or
+      # an ideographic character does unless `current` is only opening
+      # brackets; Latin text does after a unit that ends in a CJK character
+      # (an opening bracket or Latin followed by a closing mark stay glued).
+      def starts_unit?(char, current, open)
+        return false if current.empty? || open
+        return true if opening?(char) || ideographic?(char)
+
+        cjk?(current[-1]) && !(closing?(current[-1]) && !ideographic_unit?(current))
+      end
+
+      # A unit that carries an ideographic character (as opposed to Latin text
+      # followed by a closing mark, such as "abc。").
+      def ideographic_unit?(text) = text.each_char.any? { |char| ideographic?(char) }
+    end
+  end
+end

data/lib/stationery/text/paragraph.rb

@@ -34,6 +34,14 @@ module Stationery
         [with_lines(@lines.first(count)), with_lines(@lines.drop(count))]
       end
 
+      # [first `count` lines, the rest]; nil for an empty side.
+      def split_at(count)
+        return [nil, self] if count <= 0
+        return [self, nil] if count >= @lines.size
+
+        [with_lines(@lines.first(count)), with_lines(@lines.drop(count))]
+      end
+
       # A paragraph that fits `max_height`: truncated to whole lines, shrunk
       # (never below 4pt) or left as is for :visible.
       def fit(max_height, overflow:)

data/lib/stationery/text/runs_builder.rb

@@ -46,6 +46,10 @@ module Stationery
       def size(value, text = nil, &) = styled({ size: value }, text, &)
       def font(family, text = nil, &) = styled({ family: family.to_s }, text, &)
       def link(url, text = nil, &) = styled({ link: url.to_s }, text, &)
+      # The current size multiplied: `scale(1.5) { … }` is one and a half times the text around it.
+      def scale(factor, text = nil, &) = styled({ size: @styles.last.size * factor }, text, &)
+      # Any style fields at once, also to switch a mark off: `with({ weight: :regular }) { … }`.
+      def with(overrides, text = nil, &) = styled(overrides, text, &)
 
       private
 

data/lib/stationery/text/wrapper.rb

@@ -3,7 +3,8 @@
 module Stationery
   module Text
     # Greedy line breaking over styled runs. Breaks at spaces (the space is
-    # dropped at the break) and after hyphens; a word that does not fit is
+    # dropped at the break), after hyphens, at a zero-width space (U+200B,
+    # HTML's <wbr>) and between ideographic characters (see Breaks); a word that does not fit is
     # hyphenated when it may be (a soft hyphen U+00AD names the break points,
     # else the style's `hyphenate` language), and a word wider than the line
     # is broken between characters. A word that changes style midway
@@ -12,6 +13,7 @@ module Stationery
     class Wrapper
       TOKEN = /\n|[ \t]+|[^ \t\n-]*-+|[^ \t\n-]+/
       SOFT_HYPHEN = "­"
+      ZERO_WIDTH_SPACE = Breaks::ZERO_WIDTH_SPACE
       EPSILON = 0.0001
 
       Segment = Data.define(:text, :style)
@@ -34,22 +36,67 @@ module Stationery
 
       private
 
-      # [:word, segments] | [:space, segments] | [:newline]
+      # [:word, segments] | [:space, segments] | [:newline]. A word that runs
+      # on into the next styled run stays one word ("<b>Tot</b>al"); a
+      # zero-width space is an empty :space, a break that draws nothing; a
+      # word with CJK characters is cut into its break units (Breaks).
       def items(runs)
         items = []
+        @join = false # whether the next token may continue the last word
         runs.each do |run|
-          run.text.scan(TOKEN) do |token|
-            if token == "\n" then items << [:newline]
-            elsif token.match?(/\A[ \t]/) then items << [:space, [Segment.new(token, run.style)]]
-            elsif items.last&.first == :word && !items.last.last.last.text.end_with?("-")
-              items.last.last << Segment.new(token, run.style)
-            else items << [:word, [Segment.new(token, run.style)]]
+          if run.text.include?(ZERO_WIDTH_SPACE)
+            run.text.split(ZERO_WIDTH_SPACE, -1).each_with_index do |part, index|
+              if index.positive?
+                items << [:space, []]
+                @join = false
+              end
+              tokens(items, part, run.style)
             end
+          else
+            tokens(items, run.text, run.style)
           end
         end
         items
       end
 
+      def tokens(items, text, style)
+        text.scan(TOKEN) do |token|
+          if token == "\n"
+            items << [:newline]
+            @join = false
+          elsif token.match?(/\A[ \t]/)
+            items << [:space, [Segment.new(token, style)]]
+            @join = false
+          else
+            word_units(items, token, style)
+          end
+        end
+      end
+
+      # A Latin word is one unit and may run on into the next styled run
+      # unless it ends in a hyphen; a word with CJK characters is cut into
+      # its break units, none of which the next token continues. Latin words
+      # skip the unit split, so they cost no allocation of their own.
+      def word_units(items, token, style)
+        if Breaks.cjk?(token)
+          Breaks.units(token).each do |text, glued|
+            add_unit(items, text, style, glued)
+            @join = false
+          end
+        else
+          add_unit(items, token, style, false)
+          @join = !token.end_with?("-")
+        end
+      end
+
+      def add_unit(items, text, style, glued)
+        if (@join || glued) && items.last&.first == :word
+          items.last.last << Segment.new(text, style)
+        else
+          items << [:word, [Segment.new(text, style)]]
+        end
+      end
+
       def place((kind, segments))
         @ended_with_newline = kind == :newline
         case kind
@@ -115,6 +162,8 @@ module Stationery
         return (1...chars.length).select { |index| chars[index - 1].text == SOFT_HYPHEN } if @explicit
 
         text = chars.map(&:text).join
+        return [] if Breaks.cjk?(text)
+
         language = chars.first.style.hyphenate or return []
         core = text[/\p{L}+/] or return []
         offset = text.index(core)

data/lib/stationery/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Stationery
-  VERSION = "0.9.0"
+  VERSION = "0.10.1"
 end

data/lib/stationery/warnings.rb

@@ -47,6 +47,15 @@ module Stationery
       def message = "content nested #{depth} levels deep was flattened below level #{limit}"
     end
 
+    UnsupportedCss = Data.define(:properties, :selectors) do
+      def message
+        parts = []
+        parts << "properties #{properties.join(", ")}" if properties.any?
+        parts << "selectors #{selectors.join(", ")}" if selectors.any?
+        "html styles not read: #{parts.join("; ")}"
+      end
+    end
+
     DuplicateAnchor = Data.define(:name, :page) do
       def message = %(anchor "#{name}" on page #{page} is already defined)
     end

data/lib/stationery.rb

@@ -68,6 +68,7 @@ require_relative "stationery/pdf/conformance"
 require_relative "stationery/pdf/factur_x"
 require_relative "stationery/pdf/signature"
 require_relative "stationery/pdf/signature/cms"
+require_relative "stationery/pdf/signature/timestamp"
 require_relative "stationery/forms/metrics"
 require_relative "stationery/forms/typeface"
 require_relative "stationery/forms/field"
@@ -82,6 +83,7 @@ require_relative "stationery/text/entities"
 require_relative "stationery/text/markup"
 require_relative "stationery/text/runs_builder"
 require_relative "stationery/text/line"
+require_relative "stationery/text/breaks"
 require_relative "stationery/text/wrapper"
 require_relative "stationery/text/paragraph"
 require_relative "stationery/layout/node"
@@ -106,8 +108,7 @@ require_relative "stationery/svg/path_data"
 require_relative "stationery/svg/arc"
 require_relative "stationery/svg/parser"
 require_relative "stationery/svg/transform"
-require_relative "stationery/svg/selector"
-require_relative "stationery/svg/stylesheet"
+require_relative "stationery/css"
 require_relative "stationery/svg/style"
 require_relative "stationery/svg/shapes"
 require_relative "stationery/svg/bounds"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: stationery
 version: !ruby/object:Gem::Version
-  version: 0.9.0
+  version: 0.10.1
 platform: ruby
 authors:
 - Mikael Henriksson
@@ -40,6 +40,9 @@ files:
 - lib/stationery/cli/render.rb
 - lib/stationery/color.rb
 - lib/stationery/component.rb
+- lib/stationery/css.rb
+- lib/stationery/css/selector.rb
+- lib/stationery/css/stylesheet.rb
 - lib/stationery/document.rb
 - lib/stationery/elements.rb
 - lib/stationery/elements/forms.rb
@@ -90,6 +93,7 @@ files:
 - lib/stationery/forms/metrics.rb
 - lib/stationery/forms/typeface.rb
 - lib/stationery/geometry.rb
+- lib/stationery/html/css.rb
 - lib/stationery/html/document.rb
 - lib/stationery/html/tokenizer.rb
 - lib/stationery/html/tree_builder.rb
@@ -161,6 +165,7 @@ files:
 - lib/stationery/pdf/serializer.rb
 - lib/stationery/pdf/signature.rb
 - lib/stationery/pdf/signature/cms.rb
+- lib/stationery/pdf/signature/timestamp.rb
 - lib/stationery/pdf/stream.rb
 - lib/stationery/pdf/types.rb
 - lib/stationery/pdf/writer.rb
@@ -174,6 +179,7 @@ files:
 - lib/stationery/rich/nesting.rb
 - lib/stationery/rich/nodes.rb
 - lib/stationery/rich/renderer.rb
+- lib/stationery/rich/renderer/boxes.rb
 - lib/stationery/rich/renderer/indents.rb
 - lib/stationery/rich/renderer/inlines.rb
 - lib/stationery/rich/renderer/links.rb
@@ -189,11 +195,9 @@ files:
 - lib/stationery/svg/parser.rb
 - lib/stationery/svg/path_data.rb
 - lib/stationery/svg/region.rb
-- lib/stationery/svg/selector.rb
 - lib/stationery/svg/shading.rb
 - lib/stationery/svg/shapes.rb
 - lib/stationery/svg/style.rb
-- lib/stationery/svg/stylesheet.rb
 - lib/stationery/svg/text.rb
 - lib/stationery/svg/transform.rb
 - lib/stationery/svg/viewport.rb
@@ -205,6 +209,7 @@ files:
 - lib/stationery/testing/marked_text.rb
 - lib/stationery/testing/matchers.rb
 - lib/stationery/testing/structure_reader.rb
+- lib/stationery/text/breaks.rb
 - lib/stationery/text/entities.rb
 - lib/stationery/text/entities/html4.rb
 - lib/stationery/text/line.rb