Add support for latex fragments

README.org

@@ -12,9 +12,6 @@ in general, have a look at the Makefile - it's short enough.
 * not yet implemented
 ** deadlines and scheduling
 see https://orgmode.org/manual/Deadlines-and-scheduling.html
-** latex fragments
-see https://orgmode.org/manual/LaTeX-fragments.html
-+ mathjax for html_writer ?
 ** more types of links
 see https://orgmode.org/manual/External-links.html & https://orgmode.org/manual/Internal-links.html
 - radio target <<<MyTarget>>>

org/html_writer.go

@@ -247,6 +247,12 @@ func (w *HTMLWriter) WriteEmphasis(e Emphasis) {
 	w.WriteString(tags[1])
 }
 
+func (w *HTMLWriter) WriteLatexFragment(l LatexFragment) {
+	w.WriteString(l.OpeningPair)
+	WriteNodes(w, l.Content...)
+	w.WriteString(l.ClosingPair)
+}
+
 func (w *HTMLWriter) WriteStatisticToken(s StatisticToken) {
 	w.WriteString(fmt.Sprintf(`<code class="statistic">[%s]</code>`, s.Content))
 }

org/inline.go

@@ -30,6 +30,12 @@ type Emphasis struct {
 	Content []Node
 }
 
+type LatexFragment struct {
+	OpeningPair string
+	ClosingPair string
+	Content     []Node
+}
+
 type FootnoteLink struct {
 	Name       string
 	Definition *FootnoteDefinition
@@ -51,10 +57,17 @@ var subScriptSuperScriptRegexp = regexp.MustCompile(`^([_^]){([^{}]+?)}`)
 var timestampRegexp = regexp.MustCompile(`^<(\d{4}-\d{2}-\d{2})( [A-Za-z]+)?( \d{2}:\d{2})?( \+\d+[dwmy])?>`)
 var footnoteRegexp = regexp.MustCompile(`^\[fn:([\w-]*?)(:(.*?))?\]`)
 var statisticsTokenRegexp = regexp.MustCompile(`^\[(\d+/\d+|\d+%)\]`)
+var latexFragmentRegexp = regexp.MustCompile(`(?s)^\\begin{(\w+)}(.*)\\end{(\w+)}`)
 
 var timestampFormat = "2006-01-02 Mon 15:04"
 var datestampFormat = "2006-01-02 Mon"
 
+var latexFragmentPairs = map[string]string{
+	`\(`: `\)`,
+	`\[`: `\]`,
+	`$$`: `$$`,
+}
+
 func (d *Document) parseInline(input string) (nodes []Node) {
 	previous, current := 0, 0
 	for current < len(input) {
@@ -73,7 +86,9 @@ func (d *Document) parseInline(input string) (nodes []Node) {
 		case '<':
 			consumed, node = d.parseTimestamp(input, current)
 		case '\\':
-			consumed, node = d.parseExplicitLineBreak(input, current)
+			consumed, node = d.parseExplicitLineBreakOrLatexFragment(input, current)
+		case '$':
+			consumed, node = d.parseLatexFragment(input, current)
 		case '\n':
 			consumed, node = d.parseLineBreak(input, current)
 		case ':':
@@ -128,14 +143,38 @@ func (d *Document) parseLineBreak(input string, start int) (int, Node) {
 	return i - start, LineBreak{i - start}
 }
 
-func (d *Document) parseExplicitLineBreak(input string, start int) (int, Node) {
+func (d *Document) parseExplicitLineBreakOrLatexFragment(input string, start int) (int, Node) {
-	if start == 0 || input[start-1] == '\n' || start+2 >= len(input) || input[start+1] != '\\' {
+	switch {
+	case start+2 >= len(input):
+	case input[start+1] == '\\' && start != 0 && input[start-1] != '\n':
+		for i := start + 2; unicode.IsSpace(rune(input[i])); i++ {
+			if i >= len(input) || input[i] == '\n' {
+				return i + 1 - start, ExplicitLineBreak{}
+			}
+		}
+	case input[start+1] == '(' || input[start+1] == '[':
+		return d.parseLatexFragment(input, start)
+	case strings.Index(input[start:], `\begin{`) == 0:
+		if m := latexFragmentRegexp.FindStringSubmatch(input[start:]); m != nil {
+			if open, content, close := m[1], m[2], m[3]; open == close {
+				openingPair, closingPair := `\begin{`+open+`}`, `\end{`+close+`}`
+				i := strings.Index(input[start:], closingPair)
+				return i + len(closingPair), LatexFragment{openingPair, closingPair, d.parseRawInline(content)}
+			}
+		}
+	}
+	return 0, nil
+}
+
+func (d *Document) parseLatexFragment(input string, start int) (int, Node) {
+	if start+2 >= len(input) {
 		return 0, nil
 	}
-	for i := start + 2; unicode.IsSpace(rune(input[i])); i++ {
+	openingPair := input[start : start+2]
-		if i >= len(input) || input[i] == '\n' {
+	closingPair := latexFragmentPairs[openingPair]
-			return i + 1 - start, ExplicitLineBreak{}
+	if i := strings.Index(input[start+2:], closingPair); i != -1 {
-		}
+		content := d.parseRawInline(input[start+2 : start+2+i])
+		return i + 2 + 2, LatexFragment{openingPair, closingPair, content}
 	}
 	return 0, nil
 }
@@ -312,6 +351,7 @@ func (n LineBreak) String() string         { return orgWriter.nodesAsString(n) }
 func (n ExplicitLineBreak) String() string { return orgWriter.nodesAsString(n) }
 func (n StatisticToken) String() string    { return orgWriter.nodesAsString(n) }
 func (n Emphasis) String() string          { return orgWriter.nodesAsString(n) }
+func (n LatexFragment) String() string     { return orgWriter.nodesAsString(n) }
 func (n FootnoteLink) String() string      { return orgWriter.nodesAsString(n) }
 func (n RegularLink) String() string       { return orgWriter.nodesAsString(n) }
 func (n Timestamp) String() string         { return orgWriter.nodesAsString(n) }

org/org_writer.go

@@ -266,6 +266,12 @@ func (w *OrgWriter) WriteEmphasis(e Emphasis) {
 	w.WriteString(borders[1])
 }
 
+func (w *OrgWriter) WriteLatexFragment(l LatexFragment) {
+	w.WriteString(l.OpeningPair)
+	WriteNodes(w, l.Content...)
+	w.WriteString(l.ClosingPair)
+}
+
 func (w *OrgWriter) WriteStatisticToken(s StatisticToken) {
 	w.WriteString(fmt.Sprintf("[%s]", s.Content))
 }

org/testdata/latex.html

@@ -0,0 +1,45 @@
+<p>
+without latex delimiters the <code class="verbatim">_{i=1}</code> and <code class="verbatim">_n</code> in <code class="verbatim">\sum_{i=1}^n a_n</code> are interpreted as subscripts.
+we support <code class="verbatim">\(...\)</code>, <code class="verbatim">\[...\]</code>, <code class="verbatim">$$...$$</code> and <code class="verbatim">\begin{$env}...\end{$env}</code> as latex fragment delimiters.
+</p>
+<ul>
+<li>
+<p>
+∑<sub>i=1</sub>^n a_n (without latex delimiter)
+</p>
+</li>
+<li>
+<p>
+\(\sum_{i=1}^n a_n\)
+</p>
+</li>
+<li>
+<p>
+\[\sum_{i=1}^n a_n\]
+</p>
+</li>
+<li>
+<p>
+$$\sum_{i=1}^n a_n$$
+</p>
+</li>
+<li>
+<p>
+\begin{xyz}\sum_{i=1}^n a_n\end{xyz}
+</p>
+</li>
+<li>
+<p>
+\begin{xyz}
+\sum_{i=1}^n a_n
+\end{xyz}
+</p>
+</li>
+<li>
+<p>
+\begin{xyz}
+foobar
+\end{xyz}
+</p>
+</li>
+</ul>

org/testdata/latex.org

@@ -0,0 +1,15 @@
+without latex delimiters the =_{i=1}= and =_n= in =\sum_{i=1}^n a_n= are interpreted as subscripts.
+we support =\(...\)=, =\[...\]=, =$$...$$= and =\begin{$env}...\end{$env}= as latex fragment delimiters.
+
+- \sum_{i=1}^n a_n (without latex delimiter)
+- \(\sum_{i=1}^n a_n\)
+- \[\sum_{i=1}^n a_n\]
+- $$\sum_{i=1}^n a_n$$
+- \begin{xyz}\sum_{i=1}^n a_n\end{xyz}
+- \begin{xyz}
+  \sum_{i=1}^n a_n
+  \end{xyz}
+
+- \begin{xyz}
+  foobar
+  \end{xyz}

org/testdata/latex.pretty_org

@@ -0,0 +1,15 @@
+without latex delimiters the =_{i=1}= and =_n= in =\sum_{i=1}^n a_n= are interpreted as subscripts.
+we support =\(...\)=, =\[...\]=, =$$...$$= and =\begin{$env}...\end{$env}= as latex fragment delimiters.
+
+- \sum_{i=1}^n a_n (without latex delimiter)
+- \(\sum_{i=1}^n a_n\)
+- \[\sum_{i=1}^n a_n\]
+- $$\sum_{i=1}^n a_n$$
+- \begin{xyz}\sum_{i=1}^n a_n\end{xyz}
+- \begin{xyz}
+  \sum_{i=1}^n a_n
+  \end{xyz}
+
+- \begin{xyz}
+  foobar
+  \end{xyz}

org/writer.go

@@ -25,6 +25,7 @@ type Writer interface {
 	WriteParagraph(Paragraph)
 	WriteText(Text)
 	WriteEmphasis(Emphasis)
+	WriteLatexFragment(LatexFragment)
 	WriteStatisticToken(StatisticToken)
 	WriteExplicitLineBreak(ExplicitLineBreak)
 	WriteLineBreak(LineBreak)
@@ -71,6 +72,8 @@ func WriteNodes(w Writer, nodes ...Node) {
 			w.WriteText(n)
 		case Emphasis:
 			w.WriteEmphasis(n)
+		case LatexFragment:
+			w.WriteLatexFragment(n)
 		case StatisticToken:
 			w.WriteStatisticToken(n)
 		case ExplicitLineBreak: