Initial push to gitea

internal/output/spotify.go

@@ -0,0 +1,154 @@
+// Package output renders summaries to user-visible formats. Markdown is
+// passed through; Spotify HTML uses the small tag subset that Spotify for
+// Podcasters' show-notes editor accepts (b, i, a, ul/ol/li, paragraphs).
+package output
+
+import (
+	"regexp"
+	"strings"
+)
+
+var (
+	reBoldStar    = regexp.MustCompile(`\*\*([^*\n]+)\*\*`)
+	reBoldUnder   = regexp.MustCompile(`__([^_\n]+)__`)
+	reItalicStar  = regexp.MustCompile(`\*([^*\n]+)\*`)
+	reItalicUnder = regexp.MustCompile(`(^|[\s(])_([^_\n]+)_($|[\s).,!?;:])`)
+	reLink        = regexp.MustCompile(`\[([^\]]+)\]\(([^)\s]+)\)`)
+	reInlineCode  = regexp.MustCompile("`([^`\n]+)`")
+)
+
+// MarkdownToSpotifyHTML converts a markdown summary into the limited HTML
+// subset Spotify for Podcasters renders. Unknown markdown structures degrade
+// to plain text rather than producing rejected tags.
+func MarkdownToSpotifyHTML(md string) string {
+	lines := strings.Split(strings.ReplaceAll(md, "\r\n", "\n"), "\n")
+
+	var out strings.Builder
+	listKind := "" // "ul" or "ol" while we're inside a list
+	flushList := func() {
+		if listKind != "" {
+			out.WriteString("</" + listKind + ">\n")
+			listKind = ""
+		}
+	}
+	openList := func(kind string) {
+		if listKind != kind {
+			flushList()
+			out.WriteString("<" + kind + ">\n")
+			listKind = kind
+		}
+	}
+
+	paragraph := []string{}
+	flushPara := func() {
+		if len(paragraph) == 0 {
+			return
+		}
+		text := strings.Join(paragraph, " ")
+		out.WriteString("<p>" + inline(text) + "</p>\n")
+		paragraph = paragraph[:0]
+	}
+
+	for _, raw := range lines {
+		line := strings.TrimRight(raw, " \t")
+		trim := strings.TrimSpace(line)
+
+		// Blank line: end current paragraph/list block.
+		if trim == "" {
+			flushPara()
+			flushList()
+			continue
+		}
+
+		// Horizontal rule.
+		if trim == "---" || trim == "***" || trim == "___" {
+			flushPara()
+			flushList()
+			continue
+		}
+
+		// Heading -> bold paragraph.
+		if h := headingText(trim); h != "" {
+			flushPara()
+			flushList()
+			out.WriteString("<p><b>" + inline(h) + "</b></p>\n")
+			continue
+		}
+
+		// Blockquote -> italic paragraph.
+		if strings.HasPrefix(trim, "> ") {
+			flushPara()
+			flushList()
+			out.WriteString("<p><i>" + inline(strings.TrimPrefix(trim, "> ")) + "</i></p>\n")
+			continue
+		}
+
+		// Unordered list item.
+		if strings.HasPrefix(trim, "- ") || strings.HasPrefix(trim, "* ") || strings.HasPrefix(trim, "+ ") {
+			flushPara()
+			openList("ul")
+			out.WriteString("  <li>" + inline(trim[2:]) + "</li>\n")
+			continue
+		}
+
+		// Ordered list item like "1. text".
+		if item, ok := orderedItem(trim); ok {
+			flushPara()
+			openList("ol")
+			out.WriteString("  <li>" + inline(item) + "</li>\n")
+			continue
+		}
+
+		// Anything else: append to current paragraph.
+		flushList()
+		paragraph = append(paragraph, trim)
+	}
+
+	flushPara()
+	flushList()
+
+	return strings.TrimRight(out.String(), "\n")
+}
+
+func headingText(s string) string {
+	// Up to 6 leading '#' followed by a space.
+	hashes := 0
+	for hashes < len(s) && s[hashes] == '#' {
+		hashes++
+	}
+	if hashes == 0 || hashes > 6 || hashes >= len(s) || s[hashes] != ' ' {
+		return ""
+	}
+	return strings.TrimSpace(s[hashes+1:])
+}
+
+func orderedItem(s string) (string, bool) {
+	i := 0
+	for i < len(s) && s[i] >= '0' && s[i] <= '9' {
+		i++
+	}
+	if i == 0 || i+1 >= len(s) || s[i] != '.' || s[i+1] != ' ' {
+		return "", false
+	}
+	return strings.TrimSpace(s[i+2:]), true
+}
+
+func inline(s string) string {
+	s = escapeHTML(s)
+	s = reInlineCode.ReplaceAllString(s, "$1")
+	s = reBoldStar.ReplaceAllString(s, "<b>$1</b>")
+	s = reBoldUnder.ReplaceAllString(s, "<b>$1</b>")
+	s = reItalicStar.ReplaceAllString(s, "<i>$1</i>")
+	s = reItalicUnder.ReplaceAllString(s, "$1<i>$2</i>$3")
+	s = reLink.ReplaceAllString(s, `<a href="$2">$1</a>`)
+	return s
+}
+
+func escapeHTML(s string) string {
+	r := strings.NewReplacer(
+		"&", "&amp;",
+		"<", "&lt;",
+		">", "&gt;",
+	)
+	return r.Replace(s)
+}