slaw 2.2.0 → 3.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: aa23761ec680b927115914fd421ee7492984b78f
-  data.tar.gz: 80da01e30b41401f66b5784302f92002432535f2
+  metadata.gz: 655dd4179f692952514dc6341d332b92bb5ed09f
+  data.tar.gz: f94a6f5d09da791132c901343cdc7c0f228822ec
 SHA512:
-  metadata.gz: 28c4c706a2dd46256ea6a669d242576dc970195060b42e45e1c940c34c53b72520781a7d8e9f6171f4df7c6265040462626072ebc3bbd44c6c824e6e50c24de5
-  data.tar.gz: 9df4834b9134ad84f78f43bebc1a46956987876d6f6ffcf54b643d6f1b7addc7bff25273e859fb78fff428e1150211525b985ed37b368753bb578514f064d5ba
+  metadata.gz: dae4cc911be85f416e81252489f7914d8da142aa0eacf6e92a15c4f314f7569394eb718cad620ccabe098c352c87f02c4a5ae0bdb8cc7430523a4db1f65add54
+  data.tar.gz: ea1bc757d9a29b0f998366f36bd5c5baec6b2b0196705369a06a9ded88aafb1740ac27e364a9ccd1b5320a14a00cb57179911294cd858fd94f5106a2ff5f2d18

data/README.md

@@ -84,6 +84,16 @@ You can create your own grammar by creating a gem that provides these files and
 
 ## Changelog
 
+### 3.0.0 (28 March 2019)
+
+* Inline bold and italics
+* Support for CROSSHEADING elements using an empty hcontainer until we support AKN 3.0
+* Support for LONGTITLE in PREFACE
+* Remarks and references support nested inline elements
+* BREAKING: `clauses` rule renamed to `inline_elements` so as not to clash with real AKN clauses
+* BREAKING: `block_paragraphs` rule renamed to `generic_container` and adjusted to be singular to be simpler to understand
+* BREAKING: un-numbered paragraph elements have new ids, that should not clash with numbered paragraphs from other grammars
+
 ### 2.2.0 (18 March 2019)
 
 * Schedules use hcontainer, not article

data/bin/slaw

@@ -100,3 +100,8 @@ class SlawCLI < Thor
 end
 
 SlawCLI.start(ARGV)
+# Nokogiri can take 5-10 seconds to clean up and we don't care, just
+# exit now
+STDOUT.flush
+STDERR.flush
+exit! 0

data/lib/slaw/grammars/inlines.treetop

@@ -10,18 +10,23 @@ module Slaw
       # inline content
 
       rule inline_statement
-        space? '\\'? clauses eol
+        space? '\\'? inline_items eol
         <NakedStatement>
       end
 
       # one or more words, allowing inline elements
-      rule clauses
-        (remark / image / ref / [^\n])+
-        <Clauses>
+      # TODO: this is badly named, a better name would be inline_items
+      rule inline_items
+        inline_item+ <InlineItems>
+      end
+
+      rule inline_item
+        remark / image / ref / bold / italics / [^\n]
+        <InlineItem>
       end
 
       rule remark
-        '[[' content:(ref / (!']]' .))+ ']]'
+        '[[' content:(!']]' inline_item)+ ']]'
         <Remark>
       end
 
@@ -34,12 +39,25 @@ module Slaw
         <Image>
       end
 
+      rule bold
+        # **foo**
+        '**' content:(!'**' inline_item)+ '**'
+        <Bold>
+      end
+
+      rule italics
+        # //foo//
+        '//' content:(!'//' inline_item)+ '//'
+        <Italics>
+      end
+
       rule ref
         # links like markdown
         # eg. [link text](link url)
-        '[' content:(!'](' [^\n])+ '](' href:([^)\n]+) ')'
+        '[' content:(!'](' inline_item)+ '](' href:([^)\n]+) ')'
         <Ref>
       end
+
     end
   end
 end

data/lib/slaw/grammars/inlines_nodes.rb

@@ -3,22 +3,18 @@ module Slaw
     module Inlines
       class NakedStatement < Treetop::Runtime::SyntaxNode
         def to_xml(b, idprefix, i=0)
-          b.p { |b| clauses.to_xml(b, idprefix) } if clauses
+          b.p { |b| inline_items.to_xml(b, idprefix) } if inline_items
         end
 
         def content
-          clauses
+          inline_items
         end
       end
 
-      class Clauses < Treetop::Runtime::SyntaxNode
+      class InlineItems < Treetop::Runtime::SyntaxNode
         def to_xml(b, idprefix=nil)
           for e in elements
-            if e.respond_to? :to_xml
-              e.to_xml(b, idprefix)
-            else
-              b.text(e.text_value)
-            end
+            e.to_xml(b, idprefix)
           end
         end
       end
@@ -28,11 +24,7 @@ module Slaw
           b.remark(status: 'editorial') do |b|
             b.text('[')
             for e in content.elements
-              if e.respond_to? :to_xml
-                e.to_xml(b, idprefix)
-              else
-                b.text(e.text_value)
-              end
+              e.inline_item.to_xml(b, idprefix)
             end
             b.text(']')
           end
@@ -47,9 +39,39 @@ module Slaw
         end
       end
 
+      class InlineItem < Treetop::Runtime::SyntaxNode
+        def to_xml(b, idprefix)
+          b.text(text_value)
+        end
+      end
+
       class Ref < Treetop::Runtime::SyntaxNode
         def to_xml(b, idprefix)
-          b.ref(content.text_value, href: href.text_value)
+          b.ref(href: href.text_value) { |b|
+            for e in content.elements
+              e.inline_item.to_xml(b, idprefix)
+            end
+          }
+        end
+      end
+
+      class Bold < Treetop::Runtime::SyntaxNode
+        def to_xml(b, idprefix)
+          b.b { |b|
+            for e in content.elements
+              e.inline_item.to_xml(b, idprefix)
+            end
+          }
+        end
+      end
+
+      class Italics < Treetop::Runtime::SyntaxNode
+        def to_xml(b, idprefix)
+          b.i { |b|
+            for e in content.elements
+              e.inline_item.to_xml(b, idprefix)
+            end
+          }
         end
       end
 

data/lib/slaw/grammars/schedules.treetop

@@ -20,8 +20,8 @@ module Slaw
       end
 
       rule schedule_title
-        space? schedule_title_prefix space? "\""? num:alphanums? "\""? [ \t:.-]* title:clauses?
-        subheading:(newline space? clauses)?
+        space? schedule_title_prefix space? "\""? num:alphanums? "\""? [ \t:.-]* title:inline_items?
+        subheading:(newline space? inline_items)?
         eol
       end
 

data/lib/slaw/grammars/schedules_nodes.rb

@@ -27,7 +27,7 @@ module Slaw
         def alias
           if not schedule_title.title.text_value.blank?
             # plain-text elements only
-            schedule_title.title.elements.select { |x| !x.respond_to? :to_xml }.map { |x| x.text_value }.join('').strip
+            schedule_title.title.elements.select { |x| x.instance_of? ::Slaw::Grammars::Inlines::InlineItem }.map { |x| x.text_value }.join('').strip
           elsif num
             "Schedule #{num}"
           else
@@ -45,7 +45,7 @@ module Slaw
 
         def subheading
           if not schedule_title.subheading.text_value.blank?
-            schedule_title.subheading.clauses
+            schedule_title.subheading.inline_items
           else
             nil
           end
@@ -113,7 +113,7 @@ module Slaw
 
       class ScheduleStatement < Treetop::Runtime::SyntaxNode
         def to_xml(b, idprefix)
-          b.p { |b| clauses.to_xml(b, idprefix) } if clauses
+          b.p { |b| inline_items.to_xml(b, idprefix) } if inline_items
         end
       end
     end

data/lib/slaw/grammars/tables.treetop

@@ -38,7 +38,7 @@ module Slaw
       end
 
       rule table_line
-        clauses:clauses? eol
+        inline_items:inline_items? eol
         <TableLine>
       end
 

data/lib/slaw/grammars/tables_nodes.rb

@@ -59,12 +59,12 @@ module Slaw
       class TableLine < Treetop::Runtime::SyntaxNode
         # line of table content
         def to_xml(b, i, tail)
-          clauses.to_xml(b) unless clauses.empty?
+          inline_items.to_xml(b) unless inline_items.empty?
 
           # add trailing newlines.
           #   for the first line, eat whitespace at the start
           #   for the last line, eat whitespace at the end
-          if not tail and (i > 0 or not clauses.empty?)
+          if not tail and (i > 0 or not inline_items.empty?)
             eol.text_value.count("\n").times { b.eol }
           end
         end

data/lib/slaw/grammars/za/act.treetop

@@ -16,6 +16,8 @@ module Slaw
 
         ########
         # major containers
+        #
+        # These are AKN's heirarchical containers which wrap actual content.
 
         rule act
           empty_line*
@@ -28,33 +30,33 @@ module Slaw
         rule preface
           !'PREAMBLE'
           ('PREFACE'i space? eol)?
-          statements:(!'PREAMBLE' pre_body_statement)* <Preface>
+          statements:(!'PREAMBLE' preface_statement)* <Preface>
         end
 
         rule preamble
           'PREAMBLE'i space? eol
-          statements:pre_body_statement* <Preamble>
+          statements:preamble_statement* <Preamble>
         end
 
         rule body
-          children:(chapter / part / section / subsection / block_paragraphs)+ <Body>
+          children:(chapter / part / section / subsection / generic_container)+ <Body>
         end
 
         rule chapter
           heading:chapter_heading
-          children:(part / section / subsection / block_paragraphs)*
+          children:(part / section / subsection / generic_container)*
           <Chapter>
         end
 
         rule part
           heading:part_heading
-          children:(section / subsection / block_paragraphs)*
+          children:(section / subsection / generic_container)*
           <Part>
         end
 
         rule section
           section_title
-          children:(subsection / block_paragraphs)* <Section>
+          children:(subsection / generic_container)* <Section>
         end
 
         rule subsection
@@ -66,6 +68,22 @@ module Slaw
           children:block_element* <Subsection>
         end
 
+        rule generic_container
+          crossheading / block_elements
+        end
+
+        rule crossheading
+          'CROSSHEADING ' inline_items:inline_items eol
+          <Crossheading>
+        end
+
+        # Consecutive non-structured content. We allow many elements
+        # here so that we wrap consecutive block elements in one <content> tag,
+        # rather than multiple containers for each.
+        rule block_elements
+          block_element+ <BlockElements>
+        end
+
         ##########
         # group elements
         # 
@@ -129,10 +147,6 @@ module Slaw
         ##########
         # blocks of content inside containers
 
-        rule block_paragraphs
-          block_element+ <BlockParagraph>
-        end
-
         rule block_element
           (table / blocklist / naked_statement)
         end
@@ -149,7 +163,7 @@ module Slaw
 
         rule blocklist_item
           # TODO: this whitespace should probably be space, to allow empty blocklist items followed by plain text
-          space? blocklist_item_prefix whitespace item_content:(!blocklist_item_prefix clauses:clauses? eol)? eol?
+          space? blocklist_item_prefix whitespace item_content:(!blocklist_item_prefix inline_items:inline_items? eol)? eol?
           <BlocklistItem>
         end
 
@@ -164,15 +178,26 @@ module Slaw
         # and is ignored. This allows escaping of section headings, etc.
 
         rule naked_statement
-          space? !(chapter_heading / part_heading / section_title / schedule_title / subsection_prefix) '\\'? clauses eol
+          space? !(chapter_heading / part_heading / section_title / schedule_title / subsection_prefix / crossheading) '\\'? inline_items eol
           <NakedStatement>
         end
 
-        rule pre_body_statement
-          space? !(chapter_heading / part_heading / section_title / schedule_title) '\\'? clauses eol
+        rule preface_statement
+          space? !(chapter_heading / part_heading / section_title / schedule_title)
+          content:(longtitle / ('\\'? inline_items:inline_items eol))
+          <PrefaceStatement>
+        end
+
+        rule preamble_statement
+          space? !(chapter_heading / part_heading / section_title / schedule_title) '\\'? inline_items eol
           <NakedStatement>
         end
 
+        rule longtitle
+          'LONGTITLE ' inline_items:inline_items eol
+          <LongTitle>
+        end
+
         ##########
         # prefixes
 

data/lib/slaw/grammars/za/act_nodes.rb

@@ -91,6 +91,32 @@ module Slaw
           end
         end
 
+        class PrefaceStatement < Treetop::Runtime::SyntaxNode
+          def to_xml(b, idprefix, i=0)
+            if longtitle
+              longtitle.to_xml(b, idprefix)
+            else
+              b.p { |b| inline_items.to_xml(b, idprefix) }
+            end
+          end
+
+          def longtitle
+            self.content if self.content.is_a? LongTitle
+          end
+
+          def inline_items
+            content.inline_items if content.respond_to? :inline_items
+          end
+        end
+
+        class LongTitle < Treetop::Runtime::SyntaxNode
+          def to_xml(b, idprefix, i=0)
+            b.longTitle { |b|
+              b.p { |b| inline_items.to_xml(b, idprefix) }
+            }
+          end
+        end
+
         class Preamble < Treetop::Runtime::SyntaxNode
           def to_xml(b, *args)
             if text_value != ""
@@ -232,9 +258,17 @@ module Slaw
           end
         end
 
-        class BlockParagraph < Treetop::Runtime::SyntaxNode
+        class BlockElements < Treetop::Runtime::SyntaxNode
+          @@counters = {}
+
+          def self.counters
+            @@counters
+          end
+
           def to_xml(b, idprefix='', i=0)
-            id = "#{idprefix}paragraph-0"
+            @@counters[idprefix] ||= -1
+            @@counters[idprefix] += 1
+            id = "#{idprefix}paragraph#{@@counters[idprefix]}"
             idprefix = "#{id}."
 
             b.paragraph(id: id) { |b|
@@ -295,12 +329,31 @@ module Slaw
             b.item(id: idprefix + num.gsub(/[()]/, '')) { |b|
               b.num(num)
               b.p { |b|
-                item_content.clauses.to_xml(b, idprefix) if respond_to? :item_content and item_content.respond_to? :clauses
+                item_content.inline_items.to_xml(b, idprefix) if respond_to? :item_content and item_content.respond_to? :inline_items
               }
             }
           end
         end
 
+        class Crossheading < Treetop::Runtime::SyntaxNode
+          @@counters = {}
+
+          def self.counters
+            @@counters
+          end
+
+          def to_xml(b, idprefix, i=0)
+            @@counters[idprefix] ||= -1
+            @@counters[idprefix] += 1
+            id = "#{idprefix}crossheading-#{@@counters[idprefix]}"
+
+            b.hcontainer(id: id, name: 'crossheading') { |b|
+              b.heading { |b|
+                inline_items.to_xml(b, idprefix)
+              }
+            }
+          end
+        end
       end
     end
   end

data/lib/slaw/grammars/za/act_text.xsl

@@ -88,6 +88,20 @@
     <xsl:apply-templates select="./*[not(self::a:num) and not(self::a:heading)]" />
   </xsl:template>
 
+  <!-- crossheadings -->
+  <xsl:template match="a:hcontainer[@name='crossheading']">
+    <xsl:text>CROSSHEADING </xsl:text>
+    <xsl:apply-templates select="a:heading" />
+    <xsl:text>&#10;&#10;</xsl:text>
+  </xsl:template>
+
+  <!-- longtitle -->
+  <xsl:template match="a:longTitle">
+    <xsl:text>LONGTITLE </xsl:text>
+    <xsl:apply-templates />
+    <xsl:text>&#10;&#10;</xsl:text>
+  </xsl:template>
+
   <!-- p tags must end with a blank line -->
   <xsl:template match="a:p">
     <xsl:apply-templates/>
@@ -127,7 +141,7 @@
   <!-- components/schedules -->
   <!-- new-style schedules, "article" elements -->
   <xsl:template match="a:hcontainer[@name='schedule']">
-    <xsl:text>Schedule - </xsl:text>
+    <xsl:text>SCHEDULE - </xsl:text>
     <xsl:apply-templates select="a:heading" />
     <xsl:text>&#10;</xsl:text>
 
@@ -143,16 +157,17 @@
 
   <!-- old-style schedules, "article" elements -->
   <xsl:template match="a:doc/a:mainBody/a:article">
-    <xsl:text>Schedule - </xsl:text>
+    <xsl:text>SCHEDULE - </xsl:text>
     <xsl:value-of select="../../a:meta/a:identification/a:FRBRWork/a:FRBRalias/@value" />
     <xsl:text>&#10;</xsl:text>
 
-    <xsl:if test="not(a:mainBody/a:article/a:heading)">
-      <!-- ensure an extra blank line if there is no heading -->
+    <xsl:if test="a:heading">
+      <xsl:apply-templates select="a:heading" />
       <xsl:text>&#10;</xsl:text>
     </xsl:if>
 
-    <xsl:apply-templates select="a:mainBody" />
+    <xsl:text>&#10;</xsl:text>
+    <xsl:apply-templates select="./*[not(self::a:heading)]"/>
   </xsl:template>
 
 
@@ -238,6 +253,18 @@
     <xsl:text>)</xsl:text>
   </xsl:template>
 
+  <xsl:template match="a:i">
+    <xsl:text>//</xsl:text>
+    <xsl:apply-templates />
+    <xsl:text>//</xsl:text>
+  </xsl:template>
+
+  <xsl:template match="a:b">
+    <xsl:text>**</xsl:text>
+    <xsl:apply-templates />
+    <xsl:text>**</xsl:text>
+  </xsl:template>
+
   <xsl:template match="a:eol">
     <xsl:text>&#10;</xsl:text>
   </xsl:template>