refactor passthrough as XML preprocessing, process formats=all: metan…

…orma/metanorma-standoc#827
metanorma · Nov 13, 2023 · 54b65b4 · 54b65b4
1 parent 2afe2bd
commit 54b65b4
Show file tree

Hide file tree

Showing 7 changed files with 163 additions and 24 deletions.
diff --git a/lib/isodoc/convert.rb b/lib/isodoc/convert.rb
@@ -114,7 +114,7 @@ def preprocess_xslt(docxml)
         docxml = Nokogiri::XSLT(x).transform(docxml)
       end
       docxml
-    rescue
+    rescue ::Error => e
       require "debug"; binding.b
     end
 

diff --git a/lib/isodoc/presentation_function/bibdata.rb b/lib/isodoc/presentation_function/bibdata.rb
@@ -17,17 +17,48 @@ def bibdata(docxml)
 
     def extension_insert(docxml, path = [])
       ins = docxml.at(ns("//metanorma-extension")) ||
-        docxml.at(ns("//bibdata")).after("<metanorma-extension/>").next_element
+        docxml.at(ns("//bibdata"))&.after("<metanorma-extension/>")&.next_element ||
+        docxml.root.elements.first.before("<metanorma-extension/>").previous_element
       path.each do |n|
         ins = ins.at(ns("./#{n}")) || ins.add_child("<#{n}/>").first
       end
       ins
     end
 
     def preprocess_xslt_insert(docxml)
-      content = preprocess_xslt_read or return
+      content = ""
+      p = passthrough_xslt and content += p
+      p = preprocess_xslt_read and content += File.read(p)
+      content.empty? and return
       ins = extension_insert(docxml, %w(render))
-      ins << File.read(content)
+      ins << content
+    end
+
+    def passthrough_xslt
+      @output_formats.nil? and return nil
+      @output_formats.empty? and return nil
+      @output_formats.each_key.with_object([]) do |k, m|
+        m << <<~XSLT
+<preprocess-xslt format="#{k}">
+  <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns="http://www.w3.org/1999/xhtml" version="1.0">
+    <xsl:output method="xml" version="1.0" encoding="UTF-8" indent="no"/>
+    <xsl:strip-space elements="*"/>
+    <xsl:template match="@* | node()">
+    <xsl:copy>
+      <xsl:apply-templates select="@* | node()"/>
+    </xsl:copy>
+  </xsl:template>
+    <xsl:template match="*[local-name() = 'passthrough']">
+      <xsl:if test="contains(@formats,',#{k},')"> <!-- delimited -->
+    <xsl:copy>
+        <xsl:apply-templates select="@* | node()"/>
+    </xsl:copy>
+      </xsl:if>
+    </xsl:template>
+  </xsl:stylesheet>
+</preprocess-xslt>
+        XSLT
+      end.join("\n")
     end
 
     # read in from file, but with `<preprocess-xslt @format="">` wrapper
@@ -45,7 +76,7 @@ def toc_metadata(docxml)
         ins << "<toc type='table'><title>#{@i18n.toc_tables}</title></toc>"
       @tocfigures and
         ins << "<toc type='recommendation'><title>#{@i18n.toc_recommendations}" \
-               "</title></toc>"
+        "</title></toc>"
     end
 
     def address_precompose(bib)
@@ -152,8 +183,8 @@ def i8n_name(hash, pref)
       when Array
         hash.reject { |a| blank?(a) }.each_with_object([])
           .with_index do |(v1, g), i|
-          i8n_name(v1, "#{i18n_safe(k)}.#{i}").each { |x| g << x }
-        end
+            i8n_name(v1, "#{i18n_safe(k)}.#{i}").each { |x| g << x }
+          end
       else [i18n_tag(pref, hash)]
       end
     end

diff --git a/lib/isodoc/presentation_function/inline.rb b/lib/isodoc/presentation_function/inline.rb
@@ -111,14 +111,29 @@ def custom_charset(docxml)
       end
     end
 
+    def passthrough(docxml)
+      formats = @output_formats.keys
+      docxml.xpath(ns("//passthrough")).each do |p|
+        passthrough1(p, formats)
+      end
+    end
+
+    def passthrough1(elem, formats)
+      (elem["formats"] && !elem["formats"].empty?) or elem["formats"] = "all"
+      f = elem["formats"].split(",")
+      (f - formats).size == f.size or elem["formats"] = "all"
+      elem["formats"] == "all" and elem["formats"] = formats.join(",")
+      elem["formats"] = ",#{elem['formats']},"
+    end
+
     private
 
     def extract_custom_charsets(docxml)
       docxml.xpath(ns("//presentation-metadata/custom-charset-font")).
         each_with_object({}) do |x, m|
-        kv = x.text.split(":", 2)
-        m[kv[0]] = kv[1]
-      end
+          kv = x.text.split(":", 2)
+          m[kv[0]] = kv[1]
+        end
     end
 
     def found_matching_variant_sibling(node)

diff --git a/lib/isodoc/presentation_xml_convert.rb b/lib/isodoc/presentation_xml_convert.rb
@@ -89,6 +89,7 @@ def inline(docxml)
       variant docxml
       identifier docxml
       date docxml
+      passthrough docxml
       inline_format docxml
     end
 

diff --git a/spec/isodoc/blocks_spec.rb b/spec/isodoc/blocks_spec.rb
@@ -1812,13 +1812,17 @@
 
   it "processes passthrough with compatible format" do
     FileUtils.rm_f "test.html"
-    IsoDoc::HtmlConvert.new({}).convert("test", <<~INPUT, false)
+    input = <<~INPUT
       <iso-standard xmlns="http://riboseinc.com/isoxml">
-      <preface><foreword displayorder="1">
+      <preface><foreword>
       <passthrough format="html,rfc">&lt;A&gt;</passthrough><em>Hello</em><passthrough format="html,rfc">&lt;/A&gt;</passthrough>
       </foreword></preface>
       </iso-standard>
     INPUT
+    presxml = IsoDoc::PresentationXMLConvert
+      .new(presxml_options.merge(output_formats: { html: "html", rfc: "rfc" }))
+      .convert("test", input, true)
+    IsoDoc::HtmlConvert.new({}).convert("test", presxml, false)
     expect(xmlpp(File.read("test.html")
       .gsub(%r{^.*<h1 class="ForewordTitle">Foreword</h1>}m, "")
       .gsub(%r{</div>.*}m, ""))).to be_equivalent_to xmlpp(<<~OUTPUT)
@@ -1832,14 +1836,17 @@
     begin
       input = <<~INPUT
         <iso-standard xmlns="http://riboseinc.com/isoxml">
-        <preface><foreword displayorder="1">
+        <preface><foreword>
         <passthrough format="html,rfc">&lt;A&gt;</passthrough><em>Hello</em>
         </foreword></preface>
         </iso-standard>
       INPUT
+      presxml = IsoDoc::PresentationXMLConvert
+      .new(presxml_options.merge(output_formats: { html: "html", rfc: "rfc" }))
+      .convert("test", input, true)
       expect do
         IsoDoc::HtmlConvert.new({})
-          .convert("test", input, false)
+          .convert("test", presxml, false)
       end.to raise_error(SystemExit)
     rescue SystemExit
     end
@@ -1868,8 +1875,11 @@
             </body>
           </html>
     OUTPUT
+    presxml = IsoDoc::PresentationXMLConvert
+      .new(presxml_options.merge(output_formats: { html: "html", rfc: "rfc" }))
+      .convert("test", input, true)
     expect(xmlpp(strip_guid(IsoDoc::HtmlConvert.new({})
-     .convert("test", input, true)))).to be_equivalent_to xmlpp(output)
+     .convert("test", presxml, true)))).to be_equivalent_to xmlpp(output)
   end
 
   it "ignores columnbreak" do
@@ -1881,7 +1891,6 @@
            <columnbreak/>
           </clause>
         <foreword displayorder="2">
-      <passthrough format="doc,rfc">&lt;A&gt;</passthrough>
       </foreword></preface>
       </iso-standard>
     INPUT

diff --git a/spec/isodoc/figures_spec.rb b/spec/isodoc/figures_spec.rb
@@ -561,20 +561,22 @@
       </html>
     DOC
 
-    expect(strip_guid(xmlpp(IsoDoc::PresentationXMLConvert
+    output = IsoDoc::PresentationXMLConvert
       .new(presxml_options.merge(output_formats: { html: "html", doc: "doc" }))
       .convert("test", input, true)
+    expect(strip_guid(xmlpp(output
       .gsub(/&lt;/, "&#x3c;")
+      .sub(%r{<metanorma-extension>.*</metanorma-extension}m, "")
       .gsub(%r{data:image/emf;base64,[^"']+}, "data:image/emf;base64"))))
       .to be_equivalent_to xmlpp(presxml
          .gsub(%r{data:image/emf;base64,[^"']+}, "data:image/emf;base64"))
     expect(xmlpp(strip_guid(IsoDoc::HtmlConvert.new({})
-      .convert("test", presxml, true)))).to be_equivalent_to xmlpp(html)
+      .convert("test", output, true)))).to be_equivalent_to strip_guid(xmlpp(html))
     expect(xmlpp(strip_guid(IsoDoc::WordConvert.new({})
-      .convert("test", presxml, true)
+      .convert("test", output, true)
       .gsub(/['"][^'".]+(?<!odf1)(?<!odf)\.emf['"]/, "'_.emf'")
       .gsub(/['"][^'".]+\.(gif|xml)['"]/, "'_.\\1'"))))
-      .to be_equivalent_to xmlpp(doc)
+      .to be_equivalent_to strip_guid(xmlpp(doc))
   end
 
   it "converts SVG (Word)" do
@@ -619,7 +621,7 @@
         </preface>
       </iso-standard>
     OUTPUT
-    output = <<~OUTPUT
+    word = <<~OUTPUT
           <html xmlns:epub='http://www.idpf.org/2007/ops' lang='en'>
       <head>
       <style>
@@ -662,19 +664,21 @@
         </body>
       </html>
     OUTPUT
-    expect(strip_guid(xmlpp(IsoDoc::PresentationXMLConvert
+    output = IsoDoc::PresentationXMLConvert
       .new(presxml_options.merge(output_formats: { html: "html", doc: "doc" }))
      .convert("test", input, true)
+    expect(strip_guid(xmlpp(output
+      .sub(%r{<metanorma-extension>.*</metanorma-extension}m, "")
      .gsub(/&lt;/, "&#x3c;"))
           .gsub(%r{data:image/emf;base64,[^"']+}, "data:image/emf;base64")))
       .to be_equivalent_to xmlpp(presxml
         .gsub(%r{data:image/emf;base64,[^"']+}, "data:image/emf;base64"))
     expect(xmlpp(strip_guid(IsoDoc::WordConvert.new({})
-      .convert("test", presxml, true)
+      .convert("test", output, true)
       .gsub(/['"][^'".]+(?<!odf1)(?<!odf)\.emf['"]/, "'_.emf'")
       .gsub(/['"][^'".]+\.(gif|xml)['"]/, "'_.\\1'")
       .gsub(/mso-bookmark:_Ref\d+/, "mso-bookmark:_Ref"))))
-      .to be_equivalent_to xmlpp(output)
+      .to be_equivalent_to strip_guid(xmlpp(word))
   end
 
   context "disable inkscape" do

diff --git a/spec/isodoc/presentation_xml_spec.rb b/spec/isodoc/presentation_xml_spec.rb
@@ -732,6 +732,7 @@
       .new(presxml_options.merge(output_formats: { html: "html", doc: "doc" }))
       .convert("test", input, true)
       .sub(%r{<localized-strings>.*</localized-strings>}m, "")
+      .sub(%r{<metanorma-extension>.*</metanorma-extension>}m, "")
       .gsub(%r{"data:image/emf;base64,[^"]+"},
             '"data:image/emf;base64"')
       .gsub(%r{"data:application/x-msmetafile;base64,[^"]+"},
@@ -789,6 +790,7 @@
       .new(presxml_options.merge(output_formats: { html: "html" }))
       .convert("test", input, true)
       .sub(%r{<localized-strings>.*</localized-strings>}m, "")
+      .sub(%r{<metanorma-extension>.*</metanorma-extension>}m, "")
       .gsub(%r{"data:image/emf;base64,[^"]+"},
             '"data:image/emf;base64"')
       .gsub(%r{"data:application/x-msmetafile;base64,[^"]+"},
@@ -1866,6 +1868,83 @@
       .to be_equivalent_to (presxml)
   end
 
+  it "supplies formats in passthrough" do
+    input = <<~INPUT
+      <standard-document xmlns="https://www.metanorma.org/ns/standoc" type="semantic">
+        <preface>
+          <foreword id="A">
+            <p>
+            <passthrough formats="html">A</passthrough>
+            <passthrough formats="word,html">A</passthrough>
+            <passthrough formats="word,html,pdf">A</passthrough>
+            <passthrough>A</passthrough>
+            <passthrough formats="all">A</passthrough>
+            </p>
+       </foreword></preface></standard-document>
+    INPUT
+    presxml = <<~OUTPUT
+      <standard-document xmlns="https://www.metanorma.org/ns/standoc" type="presentation">
+        <metanorma-extension>
+           <render>
+             <preprocess-xslt format="html">
+               <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns="http://www.w3.org/1999/xhtml" version="1.0">
+                 <xsl:output method="xml" version="1.0" encoding="UTF-8" indent="no"/>
+                 <xsl:strip-space elements="*"/>
+                 <xsl:template match="@* | node()">
+                   <xsl:copy>
+                     <xsl:apply-templates select="@* | node()"/>
+                   </xsl:copy>
+                 </xsl:template>
+                 <xsl:template match="*[local-name() = 'passthrough']">
+                   <xsl:if test="contains(@formats,',html,')">
+                     <!-- delimited -->
+                     <xsl:copy>
+                     <xsl:apply-templates select="@* | node()"/>
+                     </xsl:copy>
+                   </xsl:if>
+                 </xsl:template>
+               </xsl:stylesheet>
+             </preprocess-xslt>
+             <preprocess-xslt format="doc">
+               <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns="http://www.w3.org/1999/xhtml" version="1.0">
+                 <xsl:output method="xml" version="1.0" encoding="UTF-8" indent="no"/>
+                 <xsl:strip-space elements="*"/>
+                 <xsl:template match="@* | node()">
+                   <xsl:copy>
+                     <xsl:apply-templates select="@* | node()"/>
+                   </xsl:copy>
+                 </xsl:template>
+                 <xsl:template match="*[local-name() = 'passthrough']">
+                   <xsl:if test="contains(@formats,',doc,')">
+                     <!-- delimited -->
+                     <xsl:copy>
+                     <xsl:apply-templates select="@* | node()"/>
+                     </xsl:copy>
+                   </xsl:if>
+                 </xsl:template>
+               </xsl:stylesheet>
+             </preprocess-xslt>
+           </render>
+         </metanorma-extension>
+       <preface><clause type="toc" id="_" displayorder="1"><title depth="1">Table of contents</title></clause>
+          <foreword id="A" displayorder="2">
+             <p>
+               <passthrough formats=",html,">A</passthrough>
+               <passthrough formats=",word,html,">A</passthrough>
+               <passthrough formats=",word,html,pdf,">A</passthrough>
+               <passthrough formats=",html,doc,">A</passthrough>
+               <passthrough formats=",html,doc,">A</passthrough>
+             </p>
+           </foreword>
+         </preface>
+       </standard-document>
+    OUTPUT
+    expect(xmlpp(strip_guid(IsoDoc::PresentationXMLConvert
+      .new(presxml_options.merge(output_formats: { html: "html", doc: "doc" }))
+      .convert("test", input, true))
+       .sub(%r{<localized-strings>.*</localized-strings>}m, "")))
+      .to be_equivalent_to xmlpp(presxml)
+  end
 
   private
-Original file line number
+Diff line change
@@ Expand Up / @@ -114,7 +114,7 @@ def preprocess_xslt(docxml) @@
             docxml = Nokogiri::XSLT(x).transform(docxml)
           end
           docxml
-        rescue
+        rescue ::Error => e
           require "debug"; binding.b
         end
@@ Expand Down @@