markup/asciidoc: Add support for .TableOfContents

Fill the .TableOfContents template variable when writing Asciidoc content. This is done by letting Asciidoc render its TOC as HTML, then extract this HTML rendered TOC, parse it into a tableofcontents.Root and finally remove it from the HTML content. This aims to stay in the logic that the Asciidoc parsing is entirely done by the external helper. See #1687
author: Nicolas Piganeau <nicolas.piganeau@ndp-systemes.fr> 2020-08-24 20:23:16 +0300
committer: Bjørn Erik Pedersen <bjorn.erik.pedersen@gmail.com> 2020-09-06 12:34:32 +0300
commit: 3ba7c92530a80f2f04fe57705ab05c247a6e8437 (patch)
tree: 1e3354fcef9ad233632006957a94a50155bc36e6 /markup
parent: 19ef27b98edca53c4138b01c0f7c7bfbeb5ffcf1 (diff)
2 files changed, 162 insertions, 2 deletions
diff --git a/markup/asciidocext/convert.go b/markup/asciidocext/convert.go
index 7dd3e3533..73ed85daa 100644
--- a/markup/asciidocext/convert.go
+++ b/markup/asciidocext/convert.go
@@ -17,6 +17,7 @@
 package asciidocext
 
 import (
+	"bytes"
 	"os/exec"
 	"path/filepath"
 
@@ -24,6 +25,8 @@ import (
 	"github.com/gohugoio/hugo/markup/asciidocext/asciidocext_config"
 	"github.com/gohugoio/hugo/markup/converter"
 	"github.com/gohugoio/hugo/markup/internal"
+	"github.com/gohugoio/hugo/markup/tableofcontents"
+	"golang.org/x/net/html"
 )
 
 /* ToDo: RelPermalink patch for svg posts not working*/
@@ -45,16 +48,32 @@ func (p provider) New(cfg converter.ProviderConfig) (converter.Provider, error)
 	}), nil
 }
 
+type asciidocResult struct {
+	converter.Result
+	toc tableofcontents.Root
+}
+
+func (r asciidocResult) TableOfContents() tableofcontents.Root {
+	return r.toc
+}
+
 type asciidocConverter struct {
 	ctx converter.DocumentContext
 	cfg converter.ProviderConfig
 }
 
 func (a *asciidocConverter) Convert(ctx converter.RenderContext) (converter.Result, error) {
-	return converter.Bytes(a.getAsciidocContent(ctx.Src, a.ctx)), nil
+	content, toc, err := extractTOC(a.getAsciidocContent(ctx.Src, a.ctx))
+	if err != nil {
+		return nil, err
+	}
+	return asciidocResult{
+		Result: converter.Bytes(content),
+		toc:    toc,
+	}, nil
 }
 
-func (c *asciidocConverter) Supports(feature identity.Identity) bool {
+func (a *asciidocConverter) Supports(_ identity.Identity) bool {
 	return false
 }
 
@@ -182,6 +201,112 @@ func getAsciidoctorExecPath() string {
 	return path
 }
 
+// extractTOC extracts the toc from the given src html.
+// It returns the html without the TOC, and the TOC data
+func extractTOC(src []byte) ([]byte, tableofcontents.Root, error) {
+	var buf bytes.Buffer
+	buf.Write(src)
+	node, err := html.Parse(&buf)
+	if err != nil {
+		return nil, tableofcontents.Root{}, err
+	}
+	var (
+		f       func(*html.Node) bool
+		toc     tableofcontents.Root
+		toVisit []*html.Node
+	)
+	f = func(n *html.Node) bool {
+		if n.Type == html.ElementNode && n.Data == "div" {
+			for _, a := range n.Attr {
+				if a.Key == "id" && a.Val == "toc" {
+					toc, err = parseTOC(n)
+					if err != nil {
+						return false
+					}
+					n.Parent.RemoveChild(n)
+					return true
+				}
+			}
+		}
+		if n.FirstChild != nil {
+			toVisit = append(toVisit, n.FirstChild)
+		}
+		if n.NextSibling != nil {
+			if ok := f(n.NextSibling); ok {
+				return true
+			}
+		}
+		for len(toVisit) > 0 {
+			nv := toVisit[0]
+			toVisit = toVisit[1:]
+			if ok := f(nv); ok {
+				return true
+			}
+		}
+		return false
+	}
+	f(node)
+	if err != nil {
+		return nil, tableofcontents.Root{}, err
+	}
+	buf.Reset()
+	err = html.Render(&buf, node)
+	if err != nil {
+		return nil, tableofcontents.Root{}, err
+	}
+	// ltrim <html><head></head><body> and rtrim </body></html> which are added by html.Render
+	res := buf.Bytes()[25:]
+	res = res[:len(res)-14]
+	return res, toc, nil
+}
+
+// parseTOC returns a TOC root from the given toc Node
+func parseTOC(doc *html.Node) (tableofcontents.Root, error) {
+	var (
+		toc tableofcontents.Root
+		f   func(*html.Node, int, int)
+	)
+	f = func(n *html.Node, parent, level int) {
+		if n.Type == html.ElementNode {
+			switch n.Data {
+			case "ul":
+				if level == 0 {
+					parent += 1
+				}
+				level += 1
+				f(n.FirstChild, parent, level)
+			case "li":
+				for c := n.FirstChild; c != nil; c = c.NextSibling {
+					if c.Type != html.ElementNode || c.Data != "a" {
+						continue
+					}
+					var href string
+					for _, a := range c.Attr {
+						if a.Key == "href" {
+							href = a.Val[1:]
+							break
+						}
+					}
+					for d := c.FirstChild; d != nil; d = d.NextSibling {
+						if d.Type == html.TextNode {
+							toc.AddAt(tableofcontents.Header{
+								Text: d.Data,
+								ID:   href,
+							}, parent, level)
+						}
+					}
+				}
+				f(n.FirstChild, parent, level)
+			}
+		}
+		if n.NextSibling != nil {
+			f(n.NextSibling, parent, level)
+		}
+	}
+	f(doc.FirstChild, 0, 0)
+	return toc, nil
+}
+
 // Supports returns whether Asciidoctor is installed on this computer.
 func Supports() bool {
 	return getAsciidoctorExecPath() != ""
diff --git a/markup/asciidocext/convert_test.go b/markup/asciidocext/convert_test.go
index 22ac8e557..1c4b77bb9 100644
--- a/markup/asciidocext/convert_test.go
+++ b/markup/asciidocext/convert_test.go
@@ -270,3 +270,38 @@ func TestConvert(t *testing.T) {
 	c.Assert(err, qt.IsNil)
 	c.Assert(string(b.Bytes()), qt.Equals, "<div class=\"paragraph\">\n<p>testContent</p>\n</div>\n")
 }
+
+func TestTableOfContents(t *testing.T) {
+	if !Supports() {
+		t.Skip("asciidoc/asciidoctor not installed")
+	}
+	c := qt.New(t)
+	p, err := Provider.New(converter.ProviderConfig{Logger: loggers.NewErrorLogger()})
+	c.Assert(err, qt.IsNil)
+	conv, err := p.New(converter.DocumentContext{})
+	c.Assert(err, qt.IsNil)
+	b, err := conv.Convert(converter.RenderContext{Src: []byte(`:toc: macro
+:toclevels: 4
+toc::[]
+
+=== Introduction
+
+== Section 1
+
+=== Section 1.1
+
+==== Section 1.1.1
+
+=== Section 1.2
+
+testContent
+
+== Section 2
+`)})
+	c.Assert(err, qt.IsNil)
+	toc, ok := b.(converter.TableOfContentsProvider)
+	c.Assert(ok, qt.Equals, true)
+	root := toc.TableOfContents()
+	c.Assert(root.ToHTML(2, 4, false), qt.Equals, "<nav id=\"TableOfContents\">\n  <ul>\n    <li><a href=\"#_introduction\">Introduction</a></li>\n    <li><a href=\"#_section_1\">Section 1</a>\n      <ul>\n        <li><a href=\"#_section_1_1\">Section 1.1</a>\n          <ul>\n            <li><a href=\"#_section_1_1_1\">Section 1.1.1</a></li>\n          </ul>\n        </li>\n        <li><a href=\"#_section_1_2\">Section 1.2</a></li>\n      </ul>\n    </li>\n    <li><a href=\"#_section_2\">Section 2</a></li>\n  </ul>\n</nav>")
+	c.Assert(root.ToHTML(2, 3, false), qt.Equals, "<nav id=\"TableOfContents\">\n  <ul>\n    <li><a href=\"#_introduction\">Introduction</a></li>\n    <li><a href=\"#_section_1\">Section 1</a>\n      <ul>\n        <li><a href=\"#_section_1_1\">Section 1.1</a></li>\n        <li><a href=\"#_section_1_2\">Section 1.2</a></li>\n      </ul>\n    </li>\n    <li><a href=\"#_section_2\">Section 2</a></li>\n  </ul>\n</nav>")
+}
author	Nicolas Piganeau <nicolas.piganeau@ndp-systemes.fr>	2020-08-24 20:23:16 +0300
committer	Bjørn Erik Pedersen <bjorn.erik.pedersen@gmail.com>	2020-09-06 12:34:32 +0300
commit	3ba7c92530a80f2f04fe57705ab05c247a6e8437 (patch)
tree	1e3354fcef9ad233632006957a94a50155bc36e6 /markup
parent	19ef27b98edca53c4138b01c0f7c7bfbeb5ffcf1 (diff)