slaw 0.13.0 → 0.14.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 60aacb892dc41e5b6f92e698df123bd34c46a585
-  data.tar.gz: 4aa1590df1aac33962cb38780235bd45a7e522be
+  metadata.gz: f461d44372208bc42f6cedb31c28b4cf8953c58d
+  data.tar.gz: cc3e56f02927e80e96aacfb67b19b3be88c64836
 SHA512:
-  metadata.gz: ebd4dfc35e3f15518236ddea7bfd961482e768a39703bb4e123cb04695825938cd426600773ef3323592a616ea07be9eeaf678222cd7243c6fccf263a9cf1065
-  data.tar.gz: 1fbfd3883cb51575e7318f2bf1e1935ad6e5dc3ff1bed9f9dd6d71e19eb9c7bf895b2b7829a3402bc529675d80b34aeb921853c179074d77dc72437a8bc1c6ea
+  metadata.gz: 17ba42c10bd5fb2efaf5bec9e9570efb0cbd840c974d0ccf69509114358a9bc4370385d07d4ec18d685c39f96c61899790fdd101d30df69596e634875ebcac9b
+  data.tar.gz: 8c47909bbe42bd51b6fb0ebb46893b4808340e8259aa79721ed1516f8dfd9740e7873a5ba65451245505d8890be26a61bc10aec7ce9f29af2766d1ef74314d27

data/README.md

@@ -218,6 +218,11 @@ Akoma Ntoso `component` elements at the end of the XML document, with a name of
 
 ## Changelog
 
+### 0.14.0
+
+* Support inline image tags, using Markdown syntax: \![alt text](image url)
+* Smarter un-break lines
+
 ### 0.13.0
 
 * FIX allow Schedule, Part and other headings at the start of blocklist and subsections

data/lib/slaw/parse/cleanser.rb

@@ -47,7 +47,7 @@ module Slaw
       def scrub(s)
         # we often get this unicode codepoint in the string, nuke it
         s.gsub([65532].pack('U*'), '')\
-         .gsub("", '')
+         .gsub(/\n*/, '')
       end
 
       # change weird quotes to normal ones
@@ -135,8 +135,12 @@ module Slaw
       def unbreak_lines(s)
         lines = s.split(/\n/)
         output = []
-        start_re = /^\s*[a-z]/
-        end_re   = /[a-z0-9]\s*$/
+
+        # set of regex matcher pairs, one for the prev line, one for the current line
+        matchers = [
+          [/[a-z0-9]$/, /^\s*[a-z]/],  # line ends with and starst with lowercase
+          [/;$/, /^\s*(and|or)/],      # ends with ; then and/or on new line
+        ]
 
         prev = nil
         lines.each_with_index do |line, i|
@@ -144,8 +148,16 @@ module Slaw
             output << line
           else
             prev = output[-1]
+            unbreak = false
+
+            for prev_re, curr_re in matchers
+              if prev =~ prev_re and line =~ curr_re
+                unbreak = true
+                break
+              end
+            end
 
-            if line =~ start_re and prev =~ end_re
+            if unbreak
               output[-1] = prev + ' ' + line
             else
               output << line

data/lib/slaw/version.rb

@@ -1,3 +1,3 @@
 module Slaw
-  VERSION = "0.13.0"
+  VERSION = "0.14.0"
 end

data/lib/slaw/za/act.treetop

@@ -210,7 +210,7 @@ module Slaw
 
       # one or more words, allowing inline elements
       rule clauses
-        (remark / ref / [^\n])+
+        (remark / image / ref / [^\n])+
         <Clauses>
       end
 
@@ -219,6 +219,15 @@ module Slaw
         <Remark>
       end
 
+      rule image
+        # images like markdown
+        # eg. ![title text](image url)
+        #
+        # the title text is optional, but the enclosing square brackets aren't
+        '![' content:(!'](' [^\n])* '](' href:([^)\n]+) ')'
+        <Image>
+      end
+
       rule ref
         # links like markdown
         # eg. [link text](link url)

data/lib/slaw/za/act_nodes.rb

@@ -312,6 +312,14 @@ module Slaw
         end
       end
 
+      class Image < Treetop::Runtime::SyntaxNode
+        def to_xml(b, idprefix)
+          attrs = {src: href.text_value}
+          attrs[:alt] = content.text_value unless content.text_value.empty?
+          b.img(attrs)
+        end
+      end
+
       class Ref < Treetop::Runtime::SyntaxNode
         def to_xml(b, idprefix)
           b.ref(content.text_value, href: href.text_value)

data/spec/za/{act_spec.rb → act_block_spec.rb}

@@ -842,257 +842,6 @@ EOS
     end
   end
 
-  #-------------------------------------------------------------------------------
-  # Remarks
-
-  describe 'remark' do
-    it 'should handle a plain remark' do
-      node = parse :block_paragraphs, <<EOS
-      [[Section 2 amended by Act 23 of 2004]]
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>
-      <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark>
-    </p>
-  </content>
-</paragraph>'
-    end
-
-    it 'should handle an inline remark at the end of a sentence' do
-      node = parse :block_paragraphs, <<EOS
-      This statement has an inline remark. [[Section 2 amended by Act 23 of 2004]]
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>This statement has an inline remark. <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark></p>
-  </content>
-</paragraph>'
-    end
-
-    it 'should handle an inline remark mid-way through' do
-      node = parse :subsection, <<EOS
-      (1) This statement has an inline remark. [[Section 2 amended by Act 23 of 2004]] And now some more.
-EOS
-      to_xml(node, "", 1).should == '<subsection id="1">
-  <num>(1)</num>
-  <content>
-    <p>This statement has an inline remark. <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark> And now some more.</p>
-  </content>
-</subsection>'
-    end
-
-    it 'should handle many inline remarks' do
-      node = parse :block_paragraphs, <<EOS
-      This statement has an inline remark. [[Section 2 amended by Act 23 of 2004]]. And now some more. [[Another remark]] [[and another]]
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>This statement has an inline remark. <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark>. And now some more. <remark status="editorial">[Another remark]</remark> <remark status="editorial">[and another]</remark></p>
-  </content>
-</paragraph>'
-    end
-
-    it 'should handle a remark in a section' do
-      node = parse :section, <<EOS
-      1. Section title
-      Some text is a long line.
-
-      [[Section 1 amended by Act 23 of 2004]]
-EOS
-      to_xml(node).should == '<section id="section-1">
-  <num>1.</num>
-  <heading>Section title</heading>
-  <paragraph id="section-1.paragraph-0">
-    <content>
-      <p>Some text is a long line.</p>
-      <p>
-        <remark status="editorial">[Section 1 amended by Act 23 of 2004]</remark>
-      </p>
-    </content>
-  </paragraph>
-</section>'
-    end
-
-    it 'should handle a remark in a blocklist' do
-      node = parse :section, <<EOS
-      1. Section title
-      Some text is a long line.
-
-      (1) something
-      (a) with a remark [[Section 1 amended by Act 23 of 2004]]
-EOS
-      to_xml(node).should == '<section id="section-1">
-  <num>1.</num>
-  <heading>Section title</heading>
-  <paragraph id="section-1.paragraph-0">
-    <content>
-      <p>Some text is a long line.</p>
-    </content>
-  </paragraph>
-  <subsection id="section-1.1">
-    <num>(1)</num>
-    <content>
-      <p>something</p>
-      <blockList id="section-1.1.list1">
-        <item id="section-1.1.list1.a">
-          <num>(a)</num>
-          <p>with a remark <remark status="editorial">[Section 1 amended by Act 23 of 2004]</remark></p>
-        </item>
-      </blockList>
-    </content>
-  </subsection>
-</section>'
-    end
-
-    it 'should handle a remark in a schedule' do
-      node = parse :schedule, <<EOS
-      Schedule 1
-      A Title
-
-      [[Schedule 1 added by Act 23 of 2004]]
-
-      Some content
-EOS
-
-      today = Time.now.strftime('%Y-%m-%d')
-      to_xml(node, "").should == '<component id="component-schedule1">
-  <doc name="schedule1">
-    <meta>
-      <identification source="#slaw">
-        <FRBRWork>
-          <FRBRthis value="/za/act/1980/01/schedule1"/>
-          <FRBRuri value="/za/act/1980/01"/>
-          <FRBRalias value="Schedule 1"/>
-          <FRBRdate date="1980-01-01" name="Generation"/>
-          <FRBRauthor href="#council"/>
-          <FRBRcountry value="za"/>
-        </FRBRWork>
-        <FRBRExpression>
-          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
-          <FRBRuri value="/za/act/1980/01/eng@"/>
-          <FRBRdate date="1980-01-01" name="Generation"/>
-          <FRBRauthor href="#council"/>
-          <FRBRlanguage language="eng"/>
-        </FRBRExpression>
-        <FRBRManifestation>
-          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
-          <FRBRuri value="/za/act/1980/01/eng@"/>
-          <FRBRdate date="' + today + '" name="Generation"/>
-          <FRBRauthor href="#slaw"/>
-        </FRBRManifestation>
-      </identification>
-    </meta>
-    <mainBody>
-      <article id="schedule1">
-        <heading>A Title</heading>
-        <paragraph id="schedule1.paragraph-0">
-          <content>
-            <p>
-              <remark status="editorial">[Schedule 1 added by Act 23 of 2004]</remark>
-            </p>
-            <p>Some content</p>
-          </content>
-        </paragraph>
-      </article>
-    </mainBody>
-  </doc>
-</component>'
-    end
-  end
-
-  #-------------------------------------------------------------------------------
-  # Refs
-
-  describe 'ref' do
-    it 'should handle a plain ref' do
-      node = parse :block_paragraphs, <<EOS
-      Hello [there](/za/act/123) friend.
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>Hello <ref href="/za/act/123">there</ref> friend.</p>
-  </content>
-</paragraph>'
-    end
-
-    it 'should work many on a line' do
-      node = parse :block_paragraphs, <<EOS
-      Hello [there](/za/act/123) friend [and](http://foo.bar.com/with space) you too.
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>Hello <ref href="/za/act/123">there</ref> friend <ref href="http://foo.bar.com/with space">and</ref> you too.</p>
-  </content>
-</paragraph>'
-    end
-
-    it 'should handle brackets' do
-      node = parse :block_paragraphs, <<EOS
-      Hello ([there](/za/act/123)).
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>Hello (<ref href="/za/act/123">there</ref>).</p>
-  </content>
-</paragraph>'
-    end
-
-    it 'should handle many clauses on a line' do
-      node = parse :block_paragraphs, <<EOS
-      Hello [there](/za/act/123)[[remark one]] my[friend](/za) [[remark 2]][end](/foo).
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>Hello <ref href="/za/act/123">there</ref><remark status="editorial">[remark one]</remark> my<ref href="/za">friend</ref> <remark status="editorial">[remark 2]</remark><ref href="/foo">end</ref>.</p>
-  </content>
-</paragraph>'
-    end
-
-    it 'text should not cross end of line' do
-      node = parse :block_paragraphs, <<EOS
-      Hello [there
-      
-      my](/za/act/123) friend.
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>Hello [there</p>
-    <p>my](/za/act/123) friend.</p>
-  </content>
-</paragraph>'
-    end
-
-    it 'href should not cross end of line' do
-      node = parse :block_paragraphs, <<EOS
-      Hello [there](/za/act
-      /123) friend.
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <p>Hello [there](/za/act</p>
-    <p>/123) friend.</p>
-  </content>
-</paragraph>'
-    end
-
-    it 'href should handle refs in a list' do
-      node = parse :block_paragraphs, <<EOS
-      2.18.1 a traffic officer appointed in terms of section 3 of the Road Traffic [Act, No. 29 of 1989](/za/act/1989/29) or section 3A of the National Road Traffic [Act No. 93 of 1996](/za/act/1996/93) as the case may be;
-EOS
-      to_xml(node, "").should == '<paragraph id="paragraph-0">
-  <content>
-    <blockList id="paragraph-0.list0">
-      <item id="paragraph-0.list0.2.18.1">
-        <num>2.18.1</num>
-        <p>a traffic officer appointed in terms of section 3 of the Road Traffic <ref href="/za/act/1989/29">Act, No. 29 of 1989</ref> or section 3A of the National Road Traffic <ref href="/za/act/1996/93">Act No. 93 of 1996</ref> as the case may be;</p>
-      </item>
-    </blockList>
-  </content>
-</paragraph>'
-    end
-  end
-
   #-------------------------------------------------------------------------------
   # Numbered statements
 

data/spec/za/act_inline_spec.rb

@@ -0,0 +1,308 @@
+# encoding: UTF-8
+
+require 'builder'
+
+require 'slaw'
+
+describe Slaw::ActGenerator do
+  def parse(rule, s)
+    subject.builder.text_to_syntax_tree(s, {root: rule})
+  end
+
+  def should_parse(rule, s)
+    s << "\n" unless s.end_with?("\n")
+    tree = subject.builder.text_to_syntax_tree(s, {root: rule})
+
+    if not tree
+      raise Exception.new(subject.failure_reason || "Couldn't match to grammar") if tree.nil?
+    else
+      # count an assertion
+      tree.should_not be_nil
+    end
+  end
+
+  def to_xml(node, *args)
+    b = ::Nokogiri::XML::Builder.new
+    node.to_xml(b, *args)
+    b.doc.root.to_xml(encoding: 'UTF-8')
+  end
+
+  #-------------------------------------------------------------------------------
+  # Remarks
+
+  describe 'remark' do
+    it 'should handle a plain remark' do
+      node = parse :block_paragraphs, <<EOS
+      [[Section 2 amended by Act 23 of 2004]]
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>
+      <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark>
+    </p>
+  </content>
+</paragraph>'
+    end
+
+    it 'should handle an inline remark at the end of a sentence' do
+      node = parse :block_paragraphs, <<EOS
+      This statement has an inline remark. [[Section 2 amended by Act 23 of 2004]]
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>This statement has an inline remark. <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark></p>
+  </content>
+</paragraph>'
+    end
+
+    it 'should handle an inline remark mid-way through' do
+      node = parse :subsection, <<EOS
+      (1) This statement has an inline remark. [[Section 2 amended by Act 23 of 2004]] And now some more.
+EOS
+      to_xml(node, "", 1).should == '<subsection id="1">
+  <num>(1)</num>
+  <content>
+    <p>This statement has an inline remark. <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark> And now some more.</p>
+  </content>
+</subsection>'
+    end
+
+    it 'should handle many inline remarks' do
+      node = parse :block_paragraphs, <<EOS
+      This statement has an inline remark. [[Section 2 amended by Act 23 of 2004]]. And now some more. [[Another remark]] [[and another]]
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>This statement has an inline remark. <remark status="editorial">[Section 2 amended by Act 23 of 2004]</remark>. And now some more. <remark status="editorial">[Another remark]</remark> <remark status="editorial">[and another]</remark></p>
+  </content>
+</paragraph>'
+    end
+
+    it 'should handle a remark in a section' do
+      node = parse :section, <<EOS
+      1. Section title
+      Some text is a long line.
+
+      [[Section 1 amended by Act 23 of 2004]]
+EOS
+      to_xml(node).should == '<section id="section-1">
+  <num>1.</num>
+  <heading>Section title</heading>
+  <paragraph id="section-1.paragraph-0">
+    <content>
+      <p>Some text is a long line.</p>
+      <p>
+        <remark status="editorial">[Section 1 amended by Act 23 of 2004]</remark>
+      </p>
+    </content>
+  </paragraph>
+</section>'
+    end
+
+    it 'should handle a remark in a blocklist' do
+      node = parse :section, <<EOS
+      1. Section title
+      Some text is a long line.
+
+      (1) something
+      (a) with a remark [[Section 1 amended by Act 23 of 2004]]
+EOS
+      to_xml(node).should == '<section id="section-1">
+  <num>1.</num>
+  <heading>Section title</heading>
+  <paragraph id="section-1.paragraph-0">
+    <content>
+      <p>Some text is a long line.</p>
+    </content>
+  </paragraph>
+  <subsection id="section-1.1">
+    <num>(1)</num>
+    <content>
+      <p>something</p>
+      <blockList id="section-1.1.list1">
+        <item id="section-1.1.list1.a">
+          <num>(a)</num>
+          <p>with a remark <remark status="editorial">[Section 1 amended by Act 23 of 2004]</remark></p>
+        </item>
+      </blockList>
+    </content>
+  </subsection>
+</section>'
+    end
+
+    it 'should handle a remark in a schedule' do
+      node = parse :schedule, <<EOS
+      Schedule 1
+      A Title
+
+      [[Schedule 1 added by Act 23 of 2004]]
+
+      Some content
+EOS
+
+      today = Time.now.strftime('%Y-%m-%d')
+      to_xml(node, "").should == '<component id="component-schedule1">
+  <doc name="schedule1">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/schedule1"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="Schedule 1"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="' + today + '" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="schedule1">
+        <heading>A Title</heading>
+        <paragraph id="schedule1.paragraph-0">
+          <content>
+            <p>
+              <remark status="editorial">[Schedule 1 added by Act 23 of 2004]</remark>
+            </p>
+            <p>Some content</p>
+          </content>
+        </paragraph>
+      </article>
+    </mainBody>
+  </doc>
+</component>'
+    end
+  end
+
+  #-------------------------------------------------------------------------------
+  # Refs
+
+  describe 'ref' do
+    it 'should handle a plain ref' do
+      node = parse :block_paragraphs, <<EOS
+      Hello [there](/za/act/123) friend.
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello <ref href="/za/act/123">there</ref> friend.</p>
+  </content>
+</paragraph>'
+    end
+
+    it 'should work many on a line' do
+      node = parse :block_paragraphs, <<EOS
+      Hello [there](/za/act/123) friend [and](http://foo.bar.com/with space) you too.
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello <ref href="/za/act/123">there</ref> friend <ref href="http://foo.bar.com/with space">and</ref> you too.</p>
+  </content>
+</paragraph>'
+    end
+
+    it 'should handle brackets' do
+      node = parse :block_paragraphs, <<EOS
+      Hello ([there](/za/act/123)).
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello (<ref href="/za/act/123">there</ref>).</p>
+  </content>
+</paragraph>'
+    end
+
+    it 'should handle many clauses on a line' do
+      node = parse :block_paragraphs, <<EOS
+      Hello [there](/za/act/123)[[remark one]] my[friend](/za) [[remark 2]][end](/foo).
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello <ref href="/za/act/123">there</ref><remark status="editorial">[remark one]</remark> my<ref href="/za">friend</ref> <remark status="editorial">[remark 2]</remark><ref href="/foo">end</ref>.</p>
+  </content>
+</paragraph>'
+    end
+
+    it 'text should not cross end of line' do
+      node = parse :block_paragraphs, <<EOS
+      Hello [there
+      
+      my](/za/act/123) friend.
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello [there</p>
+    <p>my](/za/act/123) friend.</p>
+  </content>
+</paragraph>'
+    end
+
+    it 'href should not cross end of line' do
+      node = parse :block_paragraphs, <<EOS
+      Hello [there](/za/act
+      /123) friend.
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello [there](/za/act</p>
+    <p>/123) friend.</p>
+  </content>
+</paragraph>'
+    end
+
+    it 'href should handle refs in a list' do
+      node = parse :block_paragraphs, <<EOS
+      2.18.1 a traffic officer appointed in terms of section 3 of the Road Traffic [Act, No. 29 of 1989](/za/act/1989/29) or section 3A of the National Road Traffic [Act No. 93 of 1996](/za/act/1996/93) as the case may be;
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <blockList id="paragraph-0.list0">
+      <item id="paragraph-0.list0.2.18.1">
+        <num>2.18.1</num>
+        <p>a traffic officer appointed in terms of section 3 of the Road Traffic <ref href="/za/act/1989/29">Act, No. 29 of 1989</ref> or section 3A of the National Road Traffic <ref href="/za/act/1996/93">Act No. 93 of 1996</ref> as the case may be;</p>
+      </item>
+    </blockList>
+  </content>
+</paragraph>'
+    end
+  end
+
+  #-------------------------------------------------------------------------------
+  # images
+
+  describe 'images' do
+    it 'should handle a simple image' do
+      node = parse :block_paragraphs, <<EOS
+      Hello ![title](media/foo.png) friend.
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello <img src="media/foo.png" alt="title"/> friend.</p>
+  </content>
+</paragraph>'
+    end
+
+    it 'should work many on a line' do
+      node = parse :block_paragraphs, <<EOS
+      Hello ![title](media/foo.png) friend and ![](media/bar.png) a second.
+EOS
+      to_xml(node, "").should == '<paragraph id="paragraph-0">
+  <content>
+    <p>Hello <img src="media/foo.png" alt="title"/> friend and <img src="media/bar.png"/> a second.</p>
+  </content>
+</paragraph>'
+    end
+  end
+
+end

data/spec/za/act_schedules_spec.rb

@@ -0,0 +1,576 @@
+# encoding: UTF-8
+
+require 'builder'
+
+require 'slaw'
+
+describe Slaw::ActGenerator do
+  def parse(rule, s)
+    subject.builder.text_to_syntax_tree(s, {root: rule})
+  end
+
+  def should_parse(rule, s)
+    s << "\n" unless s.end_with?("\n")
+    tree = subject.builder.text_to_syntax_tree(s, {root: rule})
+
+    if not tree
+      raise Exception.new(subject.failure_reason || "Couldn't match to grammar") if tree.nil?
+    else
+      # count an assertion
+      tree.should_not be_nil
+    end
+  end
+
+  def to_xml(node, *args)
+    b = ::Nokogiri::XML::Builder.new
+    node.to_xml(b, *args)
+    b.doc.root.to_xml(encoding: 'UTF-8')
+  end
+
+  #-------------------------------------------------------------------------------
+  # schedules
+
+  context 'schedules' do
+    it 'should handle a simple schedule' do
+      node = parse :schedules, <<EOS
+Schedule
+
+Subject to approval in terms of this By-Law, the erection:
+1. Foo
+2. Bar
+EOS
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == '<component id="component-schedule">
+  <doc name="schedule">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/schedule"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="Schedule"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="' + today + '" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="schedule">
+        <paragraph id="schedule.paragraph-0">
+          <content>
+            <p>Subject to approval in terms of this By-Law, the erection:</p>
+          </content>
+        </paragraph>
+        <section id="section-1">
+          <num>1.</num>
+          <heading>Foo</heading>
+        </section>
+        <section id="section-2">
+          <num>2.</num>
+          <heading>Bar</heading>
+        </section>
+      </article>
+    </mainBody>
+  </doc>
+</component>'
+    end
+
+    it 'should serialise many schedules correctly' do
+      node = parse :schedules_container, <<EOS
+Schedule "2"
+A Title
+1. Foo
+2. Bar
+Schedule 3
+Another Title
+Baz
+Boom
+EOS
+
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == <<EOS
+<components>
+  <component id="component-schedule2">
+    <doc name="schedule2">
+      <meta>
+        <identification source="#slaw">
+          <FRBRWork>
+            <FRBRthis value="/za/act/1980/01/schedule2"/>
+            <FRBRuri value="/za/act/1980/01"/>
+            <FRBRalias value="Schedule 2"/>
+            <FRBRdate date="1980-01-01" name="Generation"/>
+            <FRBRauthor href="#council"/>
+            <FRBRcountry value="za"/>
+          </FRBRWork>
+          <FRBRExpression>
+            <FRBRthis value="/za/act/1980/01/eng@/schedule2"/>
+            <FRBRuri value="/za/act/1980/01/eng@"/>
+            <FRBRdate date="1980-01-01" name="Generation"/>
+            <FRBRauthor href="#council"/>
+            <FRBRlanguage language="eng"/>
+          </FRBRExpression>
+          <FRBRManifestation>
+            <FRBRthis value="/za/act/1980/01/eng@/schedule2"/>
+            <FRBRuri value="/za/act/1980/01/eng@"/>
+            <FRBRdate date="#{today}" name="Generation"/>
+            <FRBRauthor href="#slaw"/>
+          </FRBRManifestation>
+        </identification>
+      </meta>
+      <mainBody>
+        <article id="schedule2">
+          <heading>A Title</heading>
+          <section id="section-1">
+            <num>1.</num>
+            <heading>Foo</heading>
+          </section>
+          <section id="section-2">
+            <num>2.</num>
+            <heading>Bar</heading>
+          </section>
+        </article>
+      </mainBody>
+    </doc>
+  </component>
+  <component id="component-schedule3">
+    <doc name="schedule3">
+      <meta>
+        <identification source="#slaw">
+          <FRBRWork>
+            <FRBRthis value="/za/act/1980/01/schedule3"/>
+            <FRBRuri value="/za/act/1980/01"/>
+            <FRBRalias value="Schedule 3"/>
+            <FRBRdate date="1980-01-01" name="Generation"/>
+            <FRBRauthor href="#council"/>
+            <FRBRcountry value="za"/>
+          </FRBRWork>
+          <FRBRExpression>
+            <FRBRthis value="/za/act/1980/01/eng@/schedule3"/>
+            <FRBRuri value="/za/act/1980/01/eng@"/>
+            <FRBRdate date="1980-01-01" name="Generation"/>
+            <FRBRauthor href="#council"/>
+            <FRBRlanguage language="eng"/>
+          </FRBRExpression>
+          <FRBRManifestation>
+            <FRBRthis value="/za/act/1980/01/eng@/schedule3"/>
+            <FRBRuri value="/za/act/1980/01/eng@"/>
+            <FRBRdate date="#{today}" name="Generation"/>
+            <FRBRauthor href="#slaw"/>
+          </FRBRManifestation>
+        </identification>
+      </meta>
+      <mainBody>
+        <article id="schedule3">
+          <heading>Another Title</heading>
+          <paragraph id="schedule3.paragraph-0">
+            <content>
+              <p>Baz</p>
+              <p>Boom</p>
+            </content>
+          </paragraph>
+        </article>
+      </mainBody>
+    </doc>
+  </component>
+</components>
+EOS
+    .strip
+
+    end
+
+    it 'should handle a schedule with a title and a number' do
+      node = parse :schedules, <<EOS
+Schedule 1 - First Schedule
+Schedule Heading
+
+Subject to approval in terms of this By-Law, the erection:
+1. Foo
+2. Bar
+EOS
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == '<component id="component-schedule1">
+  <doc name="schedule1">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/schedule1"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="First Schedule"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="' + today + '" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="schedule1">
+        <heading>Schedule Heading</heading>
+        <paragraph id="schedule1.paragraph-0">
+          <content>
+            <p>Subject to approval in terms of this By-Law, the erection:</p>
+          </content>
+        </paragraph>
+        <section id="section-1">
+          <num>1.</num>
+          <heading>Foo</heading>
+        </section>
+        <section id="section-2">
+          <num>2.</num>
+          <heading>Bar</heading>
+        </section>
+      </article>
+    </mainBody>
+  </doc>
+</component>'
+    end
+
+    it 'should handle a schedule with dot in the number' do
+      node = parse :schedules, <<EOS
+Schedule 1. First Schedule
+Schedule Heading
+
+Subject to approval in terms of this By-Law, the erection:
+1. Foo
+2. Bar
+EOS
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == '<component id="component-schedule1">
+  <doc name="schedule1">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/schedule1"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="First Schedule"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="' + today + '" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="schedule1">
+        <heading>Schedule Heading</heading>
+        <paragraph id="schedule1.paragraph-0">
+          <content>
+            <p>Subject to approval in terms of this By-Law, the erection:</p>
+          </content>
+        </paragraph>
+        <section id="section-1">
+          <num>1.</num>
+          <heading>Foo</heading>
+        </section>
+        <section id="section-2">
+          <num>2.</num>
+          <heading>Bar</heading>
+        </section>
+      </article>
+    </mainBody>
+  </doc>
+</component>'
+    end
+
+    it 'should handle a schedule with a title' do
+      node = parse :schedules, <<EOS
+Schedule - First Schedule
+
+Subject to approval in terms of this By-Law, the erection:
+1. Foo
+2. Bar
+EOS
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == '<component id="component-firstschedule">
+  <doc name="firstschedule">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/firstschedule"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="First Schedule"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/firstschedule"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/firstschedule"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="' + today + '" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="firstschedule">
+        <paragraph id="firstschedule.paragraph-0">
+          <content>
+            <p>Subject to approval in terms of this By-Law, the erection:</p>
+          </content>
+        </paragraph>
+        <section id="section-1">
+          <num>1.</num>
+          <heading>Foo</heading>
+        </section>
+        <section id="section-2">
+          <num>2.</num>
+          <heading>Bar</heading>
+        </section>
+      </article>
+    </mainBody>
+  </doc>
+</component>'
+    end
+
+    it 'should serialise a single schedule without a heading' do
+      node = parse :schedules, <<EOS
+Schedule "1"
+
+Other than as is set out hereinbelow, no signs other than locality bound signs, temporary signs including loose portable sign, estate agents signs, newspaper headline posters and posters (the erection of which must comply with the appropriate schedules pertinent thereto) shall be erected on Municipal owned land.
+1. Foo
+2. Bar
+EOS
+
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == <<EOS
+<component id="component-schedule1">
+  <doc name="schedule1">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/schedule1"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="Schedule 1"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="#{today}" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="schedule1">
+        <paragraph id="schedule1.paragraph-0">
+          <content>
+            <p>Other than as is set out hereinbelow, no signs other than locality bound signs, temporary signs including loose portable sign, estate agents signs, newspaper headline posters and posters (the erection of which must comply with the appropriate schedules pertinent thereto) shall be erected on Municipal owned land.</p>
+          </content>
+        </paragraph>
+        <section id="section-1">
+          <num>1.</num>
+          <heading>Foo</heading>
+        </section>
+        <section id="section-2">
+          <num>2.</num>
+          <heading>Bar</heading>
+        </section>
+      </article>
+    </mainBody>
+  </doc>
+</component>
+EOS
+      .strip
+    end
+
+    it 'should support rich parts, chapters and sections in a schedule' do
+      node = parse :schedules, <<EOS
+Schedule 1
+Forms
+
+Part I
+Form of authentication statement
+
+This printed impression has been carefully compared by me with the bill which was passed by Parliament and found by me to be a true copy of the bill.
+
+Part II
+Form of statement of the President’s assent.
+
+I signify my assent to the bill and a whole bunch of other stuff.
+EOS
+
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == <<EOS
+<component id="component-schedule1">
+  <doc name="schedule1">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/schedule1"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="Schedule 1"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule1"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="#{today}" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="schedule1">
+        <heading>Forms</heading>
+        <part id="part-I">
+          <num>I</num>
+          <heading>Form of authentication statement</heading>
+          <paragraph id="part-I.paragraph-0">
+            <content>
+              <p>This printed impression has been carefully compared by me with the bill which was passed by Parliament and found by me to be a true copy of the bill.</p>
+            </content>
+          </paragraph>
+        </part>
+        <part id="part-II">
+          <num>II</num>
+          <heading>Form of statement of the President’s assent.</heading>
+          <paragraph id="part-II.paragraph-0">
+            <content>
+              <p>I signify my assent to the bill and a whole bunch of other stuff.</p>
+            </content>
+          </paragraph>
+        </part>
+      </article>
+    </mainBody>
+  </doc>
+</component>
+EOS
+      .strip
+    end
+
+    it 'should not get confused by schedule headings in TOC' do
+      parse :schedules_container, <<EOS
+Schedule 1. Summoning and procedure of Joint Sittings of Senate and House of Assembly.
+Schedule 2. Oaths.
+Schedule 3. Matters which shall continue to be regulated by Swazi Law and Custom.
+Schedule 4. Specially entrenched provisions and entrenched provisions.
+EOS
+    end
+
+    it 'should handle escaped schedules' do
+      node = parse :schedules, <<EOS
+Schedule
+
+Subject to approval in terms of this By-Law.
+
+\\Schedule another
+
+More stuff
+EOS
+      s = to_xml(node)
+      today = Time.now.strftime('%Y-%m-%d')
+      s.should == '<component id="component-schedule">
+  <doc name="schedule">
+    <meta>
+      <identification source="#slaw">
+        <FRBRWork>
+          <FRBRthis value="/za/act/1980/01/schedule"/>
+          <FRBRuri value="/za/act/1980/01"/>
+          <FRBRalias value="Schedule"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRcountry value="za"/>
+        </FRBRWork>
+        <FRBRExpression>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="1980-01-01" name="Generation"/>
+          <FRBRauthor href="#council"/>
+          <FRBRlanguage language="eng"/>
+        </FRBRExpression>
+        <FRBRManifestation>
+          <FRBRthis value="/za/act/1980/01/eng@/schedule"/>
+          <FRBRuri value="/za/act/1980/01/eng@"/>
+          <FRBRdate date="' + today + '" name="Generation"/>
+          <FRBRauthor href="#slaw"/>
+        </FRBRManifestation>
+      </identification>
+    </meta>
+    <mainBody>
+      <article id="schedule">
+        <paragraph id="schedule.paragraph-0">
+          <content>
+            <p>Subject to approval in terms of this By-Law.</p>
+            <p>Schedule another</p>
+            <p>More stuff</p>
+          </content>
+        </paragraph>
+      </article>
+    </mainBody>
+  </doc>
+</component>'
+    end
+
+  end
+end
+

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: slaw
 version: !ruby/object:Gem::Version
-  version: 0.13.0
+  version: 0.14.0
 platform: ruby
 authors:
 - Greg Kempe
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-05-16 00:00:00.000000000 Z
+date: 2017-10-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -216,7 +216,9 @@ files:
 - spec/parse/cleanser_spec.rb
 - spec/spec_helper.rb
 - spec/xml_helpers.rb
-- spec/za/act_spec.rb
+- spec/za/act_block_spec.rb
+- spec/za/act_inline_spec.rb
+- spec/za/act_schedules_spec.rb
 homepage: ''
 licenses:
 - MIT
@@ -237,7 +239,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.6
+rubygems_version: 2.6.12
 signing_key: 
 specification_version: 4
 summary: A lightweight library for using Akoma Ntoso acts in Ruby.
@@ -251,4 +253,6 @@ test_files:
 - spec/parse/cleanser_spec.rb
 - spec/spec_helper.rb
 - spec/xml_helpers.rb
-- spec/za/act_spec.rb
+- spec/za/act_block_spec.rb
+- spec/za/act_inline_spec.rb
+- spec/za/act_schedules_spec.rb