mopac-pdf/internal/markdown/markdown.go

// Package markdown parses the markdown subset mopac-pdf typesets: headings,
// paragraphs, lists (ordered/unordered, one nesting level), GFM pipe tables,
// fenced code blocks, chart blocks, blockquotes, horizontal rules, images and
// inline bold/italic/code/link. It is intentionally small: the input currency
// of this stack is Redmine notes, Discourse posts and briefing output, all of
// which are plain, readable markdown.
package markdown

import (
	"strconv"
	"strings"
)

// Block kinds emitted by Parse.
type Kind int

const (
	KindParagraph Kind = iota
	KindHeading
	KindList
	KindTable
	KindCode
	KindChart
	KindQuote
	KindRule
	KindImage
)

// Block is one top-level block of the document.
type Block struct {
	Kind     Kind
	Level    int      // heading level 1..6
	Text     string   // paragraph / heading / quote text (raw inline)
	Ordered  bool     // list
	Items    []string // list item text (raw inline)
	Header   []string // table: column labels (raw inline)
	Rows     [][]string
	Align    []string // table: per-column "left"|"center"|"right"
	Language string   // code fence info string
	Lines    []string // code body lines (verbatim)
	Chart    *Chart   // chart block
	Alt      string   // image alt/caption
	Path     string   // image path
}

// Chart is a parsed ```chart fenced data block. Labels/Values are parallel
// slices in input order.
type Chart struct {
	Type       string
	Title      string
	Unit       string
	Labels     []string
	Values     []float64
	WidthPct   int // embed width in percent of text column (default 78)
	OtherPairs map[string]string
}

// Parse converts markdown source into blocks. Front-matter must already be
// stripped by the caller. Unknown constructs degrade to paragraphs.
func Parse(src string) []Block {
	lines := strings.Split(strings.ReplaceAll(src, "\r\n", "\n"), "\n")
	var blocks []Block
	i := 0
	for i < len(lines) {
		line := lines[i]
		trimmed := strings.TrimSpace(line)
		switch {
		case trimmed == "":
			i++
		case strings.HasPrefix(trimmed, "```") || strings.HasPrefix(trimmed, "~~~"):
			block, next := parseFence(lines, i, trimmed)
			blocks = append(blocks, block)
			i = next
		case isHeading(trimmed):
			level := len(trimmed) - len(strings.TrimLeft(trimmed, "#"))
			text := strings.TrimSpace(strings.Trim(trimmed[level:], "#"))
			text = strings.TrimSpace(text)
			blocks = append(blocks, Block{Kind: KindHeading, Level: level, Text: text})
			i++
		case trimmed == "---" || trimmed == "***" || trimmed == "___":
			blocks = append(blocks, Block{Kind: KindRule})
			i++
		case isTableLine(trimmed) && i+1 < len(lines) && isTableSeparator(lines[i+1]):
			block, next := parseTable(lines, i)
			blocks = append(blocks, block)
			i = next
		case isListLine(trimmed):
			block, next := parseList(lines, i)
			blocks = append(blocks, block)
			i = next
		case strings.HasPrefix(trimmed, ">"):
			var parts []string
			for i < len(lines) {
				t := strings.TrimSpace(lines[i])
				if !strings.HasPrefix(t, ">") {
					break
				}
				parts = append(parts, strings.TrimSpace(strings.TrimPrefix(t, ">")))
				i++
			}
			blocks = append(blocks, Block{Kind: KindQuote, Text: strings.Join(parts, " ")})
		case isImageLine(trimmed):
			alt, path := parseImage(trimmed)
			blocks = append(blocks, Block{Kind: KindImage, Alt: alt, Path: path})
			i++
		default:
			var parts []string
			for i < len(lines) {
				t := strings.TrimSpace(lines[i])
				if t == "" || isHeading(t) || isListLine(t) || isTableLine(t) ||
					strings.HasPrefix(t, "```") || strings.HasPrefix(t, ">") ||
					t == "---" || isImageLine(t) {
					break
				}
				parts = append(parts, t)
				i++
			}
			blocks = append(blocks, Block{Kind: KindParagraph, Text: strings.Join(parts, " ")})
		}
	}
	return blocks
}

func isHeading(t string) bool {
	if !strings.HasPrefix(t, "#") {
		return false
	}
	level := len(t) - len(strings.TrimLeft(t, "#"))
	return level >= 1 && level <= 6 && len(t) > level && t[level] == ' '
}

func isListLine(t string) bool {
	if strings.HasPrefix(t, "- ") || strings.HasPrefix(t, "* ") || strings.HasPrefix(t, "+ ") {
		return true
	}
	label, rest, ok := strings.Cut(t, ". ")
	return ok && rest != "" && isOrderedLabel(label)
}

func isOrderedLabel(s string) bool {
	if s == "" {
		return false
	}
	for _, r := range s {
		if r < '0' || r > '9' {
			return false
		}
	}
	return true
}

func isTableLine(t string) bool {
	return strings.HasPrefix(t, "|") && strings.HasSuffix(t, "|") && strings.Count(t, "|") >= 2
}

func isTableSeparator(t string) bool {
	t = strings.TrimSpace(t)
	if !isTableLine(t) {
		return false
	}
	cells := splitRow(t)
	for _, c := range cells {
		c = strings.TrimSpace(c)
		if c == "" {
			return false
		}
		if !strings.ContainsAny(c, "-") {
			return false
		}
		for _, r := range c {
			if r != '-' && r != ':' && r != ' ' {
				return false
			}
		}
	}
	return true
}

func isImageLine(t string) bool {
	return strings.HasPrefix(t, "![") && strings.Contains(t, "](") && strings.HasSuffix(t, ")")
}

func parseImage(t string) (alt, path string) {
	inner := strings.TrimSuffix(strings.TrimPrefix(t, "!["), ")")
	alt, path, _ = strings.Cut(inner, "](")
	return alt, path
}

func parseFence(lines []string, i int, fence string) (Block, int) {
	info := strings.TrimSpace(strings.Trim(fence, "`~ "))
	var body []string
	j := i + 1
	for j < len(lines) {
		t := strings.TrimSpace(lines[j])
		if isClosingFence(t, fence) {
			break
		}
		body = append(body, lines[j])
		j++
	}
	if info == "chart" {
		if chart := parseChart(body); chart != nil {
			return Block{Kind: KindChart, Chart: chart}, j + 1
		}
	}
	return Block{Kind: KindCode, Language: info, Lines: body}, j + 1
}

func isClosingFence(t, fence string) bool {
	if t == fence {
		return true
	}
	marker := "`"
	if strings.HasPrefix(fence, "~") {
		marker = "~"
	}
	if strings.Trim(fence, marker) == "" {
		return strings.Trim(t, marker) == "" && len(t) >= len(fence)
	}
	// Opening fence carried an info string (e.g. ```chart): the closer is a
	// bare run of the marker at least as long as the opening run.
	openLen := len(fence) - len(strings.TrimLeft(fence, marker))
	return strings.Trim(t, marker) == "" && len(t) >= openLen
}

func parseChart(body []string) *Chart {
	c := &Chart{Type: "bar", WidthPct: 78, OtherPairs: map[string]string{}}
	for _, raw := range body {
		line := strings.TrimSpace(raw)
		if line == "" || strings.HasPrefix(line, "#") {
			continue
		}
		key, value, ok := strings.Cut(line, ":")
		if !ok {
			continue
		}
		key = strings.TrimSpace(key)
		value = strings.TrimSpace(value)
		switch strings.ToLower(key) {
		case "type":
			c.Type = strings.ToLower(value)
		case "title":
			c.Title = value
		case "unit":
			c.Unit = value
		case "width":
			if n, err := strconv.Atoi(strings.TrimSuffix(value, "%")); err == nil && n > 10 && n <= 100 {
				c.WidthPct = n
			}
		default:
			if num, err := strconv.ParseFloat(value, 64); err == nil && key != "" {
				c.Labels = append(c.Labels, key)
				c.Values = append(c.Values, num)
			} else {
				c.OtherPairs[key] = value
			}
		}
	}
	if len(c.Labels) == 0 {
		return nil
	}
	return c
}

func parseList(lines []string, i int) (Block, int) {
	block := Block{Kind: KindList}
	first := strings.TrimSpace(lines[i])
	_, rest, _ := strings.Cut(first, ". ")
	if strings.HasPrefix(first, "- ") || strings.HasPrefix(first, "* ") || strings.HasPrefix(first, "+ ") {
		rest = first[2:]
	} else {
		block.Ordered = true
	}
	block.Items = append(block.Items, rest)
	i++
	for i < len(lines) {
		t := strings.TrimSpace(lines[i])
		if t == "" {
			break
		}
		if !isListLine(t) {
			break
		}
		_, item, _ := strings.Cut(t, ". ")
		if strings.HasPrefix(t, "- ") || strings.HasPrefix(t, "* ") || strings.HasPrefix(t, "+ ") {
			item = t[2:]
		}
		block.Items = append(block.Items, item)
		i++
	}
	return block, i
}

func splitRow(t string) []string {
	t = strings.TrimSpace(t)
	t = strings.TrimPrefix(t, "|")
	t = strings.TrimSuffix(t, "|")
	raw := strings.Split(t, "|")
	cells := make([]string, len(raw))
	for i, c := range raw {
		cells[i] = strings.TrimSpace(c)
	}
	return cells
}

func parseTable(lines []string, i int) (Block, int) {
	header := splitRow(lines[i])
	sep := splitRow(lines[i+1])
	align := make([]string, len(sep))
	for col, s := range sep {
		left := strings.HasPrefix(s, ":")
		right := strings.HasSuffix(s, ":")
		switch {
		case left && right:
			align[col] = "center"
		case right:
			align[col] = "right"
		default:
			align[col] = "left"
		}
	}
	block := Block{Kind: KindTable, Header: header, Align: align}
	j := i + 2
	for j < len(lines) {
		t := strings.TrimSpace(lines[j])
		if !isTableLine(t) {
			break
		}
		row := splitRow(t)
		for len(row) < len(header) {
			row = append(row, "")
		}
		row = row[:len(header)]
		block.Rows = append(block.Rows, row)
		j++
	}
	return block, j
}