YunShu/pkg/mdprint/parse.go

package mdprint

import (
	"strings"
)

func parseBlocks(content string) []Node {
	lines := strings.Split(content, "\n")
	var blocks []Node
	i := 0

	for i < len(lines) {
		line := lines[i]
		trim := strings.TrimSpace(line)

		if trim == "" {
			i++
			continue
		}

		if level := headingLevel(trim); level > 0 {
			rest := strings.TrimSpace(trim[level:])
			blocks = append(blocks, Heading{Level: level, Content: parseInline(rest)})
			i++
			continue
		}

		if trim == "---" || trim == "***" || trim == "___" {
			blocks = append(blocks, ThematicBreak{})
			i++
			continue
		}

		if isFence(trim) {
			b, lang, consumed := parseFence(lines, i)
			blocks = append(blocks, CodeBlock{Lang: lang, Body: b})
			i += consumed
			continue
		}

		if strings.HasPrefix(trim, ">") {
			q, consumed := parseQuote(lines, i)
			blocks = append(blocks, q)
			i += consumed
			continue
		}

		if mark, ordered, ok := listMarker(trim); ok {
			items, consumed := parseListItems(lines, i, mark)
			blocks = append(blocks, List{Ordered: ordered, Items: items})
			i += consumed
			continue
		}

		if strings.HasPrefix(trim, "|") && i+1 < len(lines) && isTableSep(strings.TrimSpace(lines[i+1])) {
			t, consumed := parseTable(lines, i)
			blocks = append(blocks, t)
			i += consumed
			continue
		}

		p, consumed := parseParagraph(lines, i)
		blocks = append(blocks, p)
		i += consumed
	}

	return blocks
}

func headingLevel(s string) int {
	n := 0
	for _, c := range s {
		if c == '#' {
			n++
		} else {
			break
		}
	}
	if n > 0 && n <= 6 && len(s) > n && s[n] == ' ' {
		return n
	}
	return 0
}

func isFence(s string) bool {
	count := 0
	for _, c := range s {
		if c == '`' || c == '~' {
			count++
		} else {
			break
		}
	}
	return count >= 3
}

func fenceChar(s string) rune {
	for _, c := range s {
		return c
	}
	return '`'
}

func parseFence(lines []string, start int) (body, lang string, consumed int) {
	first := strings.TrimSpace(lines[start])
	ch := fenceChar(first)
	lang = strings.TrimSpace(first[3:])
	var buf strings.Builder
	consumed = 1
	for i := start + 1; i < len(lines); i++ {
		consumed++
		if strings.TrimSpace(lines[i]) == strings.Repeat(string(ch), 3) {
			break
		}
		if buf.Len() > 0 {
			buf.WriteString("\n")
		}
		buf.WriteString(lines[i])
	}
	body = buf.String()
	return
}

func parseQuote(lines []string, start int) (Blockquote, int) {
	var contents []Node
	var buf []string
	flush := func() {
		if len(buf) > 0 {
			text := strings.Join(buf, "\n")
			contents = append(contents, Paragraph{Content: parseInline(text)})
			buf = nil
		}
	}

	i := start
	for i < len(lines) {
		trim := strings.TrimSpace(lines[i])
		if !strings.HasPrefix(trim, ">") {
			break
		}
		rest := trim[1:]
		if strings.HasPrefix(rest, " ") {
			rest = rest[1:]
		}
		if rest == "" {
			flush()
		} else {
			buf = append(buf, rest)
		}
		i++
	}
	flush()
	return Blockquote{Children: contents}, i - start
}

func listMarker(s string) (mark string, ordered bool, ok bool) {
	trim := strings.TrimSpace(s)
	if len(trim) >= 2 && (trim[0] == '-' || trim[0] == '*') && trim[1] == ' ' {
		return string(trim[0]), false, true
	}
	if len(trim) >= 3 && trim[0] >= '1' && trim[0] <= '9' && trim[1] == '.' && trim[2] == ' ' {
		return ".", true, true
	}
	return "", false, false
}

func parseListItems(lines []string, start int, mark string) ([]ListItem, int) {
	var items []ListItem
	consumed := 0
	for i := start; i < len(lines); i++ {
		trim := strings.TrimSpace(lines[i])
		m, _, ok := listMarker(trim)
		if !ok || m != mark {
			break
		}
		consumed++
		rest := strings.TrimSpace(trim[len(mark)+1:])

		var checked *bool
		if strings.HasPrefix(rest, "[ ] ") {
			f := false
			checked = &f
			rest = rest[4:]
		} else if strings.HasPrefix(rest, "[x] ") {
			t := true
			checked = &t
			rest = rest[4:]
		} else if strings.HasPrefix(rest, "[X] ") {
			t := true
			checked = &t
			rest = rest[4:]
		}

		items = append(items, ListItem{
			Checked: checked,
			Content: parseInline(rest),
		})
	}
	return items, consumed
}

func isTableSep(s string) bool {
	if !strings.HasPrefix(s, "|") || !strings.HasSuffix(s, "|") {
		return false
	}
	s = s[1 : len(s)-1]
	cells := strings.Split(s, "|")
	if len(cells) == 0 {
		return false
	}
	for _, c := range cells {
		c = strings.TrimSpace(c)
		if c == "" {
			return false
		}
		for _, ch := range c {
			if ch != '-' && ch != ':' {
				return false
			}
		}
	}
	return true
}

func parseTable(lines []string, start int) (Table, int) {
	headerLine := strings.TrimSpace(lines[start])
	headers := splitTableRow(headerLine)

	consumed := 2
	var rows [][]string
	for i := start + 2; i < len(lines); i++ {
		trim := strings.TrimSpace(lines[i])
		if !strings.HasPrefix(trim, "|") {
			break
		}
		rows = append(rows, splitTableRow(trim))
		consumed++
	}

	return Table{Headers: headers, Rows: rows}, consumed
}

func splitTableRow(s string) []string {
	s = strings.TrimSpace(s)
	if strings.HasPrefix(s, "|") {
		s = s[1:]
	}
	if strings.HasSuffix(s, "|") {
		s = s[:len(s)-1]
	}
	cells := strings.Split(s, "|")
	result := make([]string, len(cells))
	for i, c := range cells {
		result[i] = strings.TrimSpace(c)
	}
	return result
}

func parseParagraph(lines []string, start int) (Paragraph, int) {
	var contentLines []string
	consumed := 0
	for i := start; i < len(lines); i++ {
		trim := strings.TrimSpace(lines[i])
		if trim == "" {
			break
		}
		if headingLevel(trim) > 0 {
			break
		}
		if trim == "---" || trim == "***" || trim == "___" {
			break
		}
		if isFence(trim) {
			break
		}
		if strings.HasPrefix(trim, ">") {
			break
		}
		if _, _, ok := listMarker(trim); ok {
			break
		}
		contentLines = append(contentLines, lines[i])
		consumed++
	}
	text := strings.Join(contentLines, "\n")
	return Paragraph{Content: parseInline(text)}, consumed
}