🎨 Improve formula clipping siyuan-note/siyuan#11640

Author: 88250

h2m.go

@@ -166,7 +166,7 @@ func (lute *Lute) genASTByDOM(n *html.Node, tree *parse.Tree) {
 		if lute.ParseOptions.ProtyleWYSIWYG {
 			node.Tokens = lex.EscapeProtyleMarkers(node.Tokens)
 		} else {
-			node.Tokens = lex.EscapeMarkers(node.Tokens)
+			node.Tokens = lex.EscapeCommonMarkers(node.Tokens)
 			if lute.parentIs(n, atom.Table) {
 				node.Tokens = bytes.ReplaceAll(node.Tokens, []byte("\\|"), []byte("|"))
 				node.Tokens = bytes.ReplaceAll(node.Tokens, []byte("|"), []byte("\\|"))

javascript/lute.min.js

@@ -414,7 +414,7 @@ func RepeatBackslashBeforePipe(content string) string {
 	return buf.String()
 }
 
-func EscapeMarkers(tokens []byte) []byte {
+func EscapeCommonMarkers(tokens []byte) []byte {
 	for i := 0; i < len(tokens); i++ {
 		if IsCommonInlineMarker(tokens[i]) {
 			remains := append([]byte{ItemBackslash}, tokens[i:]...)

javascript/lute.min.js.map

@@ -18,6 +18,7 @@ import (
 
 var html2MdTests = []parseTest{
 
+	{"140", "<span class=\"texhtml\">{<i>f</i><sub><i>k</i></sub>}<sub><i>k</i> ∈ <b>N</b></sub></span>", "$\\{f_k\\}_k ∈ N$\n"},
 	{"139", "<div id=\"readability-page-1\" class=\"page\">\n    <div class=\"tablet:max-xl:w-mid-content xl:w-content flex\" id=\"post-content\">\n        <div class=\" post-body text-gray-text xl:col-start-1 xl:order-4 order-4 text-lg\">\n            <p><img src=\"https://test.com/foo.png\"\n                    width=\"640\" height=\"366\" alt=\"bar\"></p>\n            <p>baz\n            </p>\n        </div>\n    </div>\n</div>", "![bar](https://test.com/foo.png)\n\nbaz\n"},
 	{"138", "<code class=\"samp docutils literal notranslate\"><em><span class=\"pre\">package_name</span></em><span class=\"pre\">-</span><em><span class=\"pre\">version</span></em><span class=\"pre\">-</span><em><span class=\"pre\">python_tag</span></em><span class=\"pre\">-</span><em><span class=\"pre\">abi_tag</span></em><span class=\"pre\">-</span><em><span class=\"pre\">platform_tag</span></em><span class=\"pre\">.whl</span></code>", "`package_name-version-python_tag-abi_tag-platform_tag.whl`\n"},
 	{"137", "<div><figure>foo<figcaption>bar</figcaption></figure><figure>foo1<figcaption>bar1</figcaption></figure><p>baz</p></div>", "foo\nbar\n\nfoo1\nbar1\n\n\nbaz\n"},

lex/token.go

@@ -141,6 +141,47 @@ func DomHTML(n *html.Node) []byte {
 	return bytes.ReplaceAll(buf.Bytes(), []byte(editor.Zwsp), nil)
 }
 
+func DomTexhtml(n *html.Node) string {
+	buf := &bytes.Buffer{}
+	if html.TextNode == n.Type {
+		buf.WriteString(n.Data)
+		return buf.String()
+	}
+	for child := n.FirstChild; nil != child; child = child.NextSibling {
+		domTexhtml0(child, buf)
+	}
+	return buf.String()
+}
+
+func domTexhtml0(n *html.Node, buffer *bytes.Buffer) {
+	if nil == n {
+		return
+	}
+
+	switch n.DataAtom {
+	case 0:
+		buffer.WriteString(escapeMathSymbol(n.Data))
+		return
+	case atom.Sup:
+		buffer.WriteString("^")
+	case atom.Sub:
+		buffer.WriteString("_")
+	}
+
+	for child := n.FirstChild; nil != child; child = child.NextSibling {
+		domTexhtml0(child, buffer)
+	}
+}
+
+func escapeMathSymbol(s string) string {
+	// 转义 Tex 公式中的符号，比如 _ ^ { }
+	s = strings.ReplaceAll(s, "_", "\\_")
+	s = strings.ReplaceAll(s, "^", "\\^")
+	s = strings.ReplaceAll(s, "{", "\\{")
+	s = strings.ReplaceAll(s, "}", "\\}")
+	return s
+}
+
 func DomText(n *html.Node) string {
 	buf := &bytes.Buffer{}
 	if html.TextNode == n.Type {

test/h2m_test.go

@@ -249,7 +249,7 @@ func (lute *Lute) adjustMath(n *html.Node) {
 	}
 
 	if strings.Contains(class, "texhtml") {
-		if mathContent := util.DomText(n); "" != mathContent {
+		if mathContent := util.DomTexhtml(n); "" != mathContent {
 			util.SetDomAttrValue(n, "data-tex", mathContent)
 			return
 		}