streamrip-go/internal/naming/naming.go

package naming

import (
	"fmt"
	"regexp"
	"strconv"
	"strings"
	"unicode"
	"unicode/utf8"
)

type Config struct {
	RestrictCharacters bool
	TruncateTo         int
}

var tokenRe = regexp.MustCompile(`\{([a-z_]+)(?::0?(\d+))?\}`)
var invalidPathRe = regexp.MustCompile(`[<>:"/\\|?*\x00-\x1F\x7F]`)

func FormatTemplate(template string, values map[string]string) string {
	return tokenRe.ReplaceAllStringFunc(template, func(m string) string {
		groups := tokenRe.FindStringSubmatch(m)
		if len(groups) < 2 {
			return m
		}
		key := groups[1]
		val := values[key]
		if len(groups) >= 3 && groups[2] != "" {
			if n, err := strconv.Atoi(val); err == nil {
				if width, widthErr := strconv.Atoi(groups[2]); widthErr == nil {
					return fmt.Sprintf("%0*d", width, n)
				}
			}
		}
		return val
	})
}

// maxNameBytes is the per-component limit on ext4, APFS and most other modern
// filesystems. TruncateTo counts runes, so a 120-rune CJK title is ~360 bytes
// and would fail with ENAMETOOLONG without this second, byte-based cap.
const maxNameBytes = 255

// windowsReserved are device names that cannot be used as a path component on
// Windows, with or without an extension.
var windowsReserved = map[string]struct{}{
	"CON": {}, "PRN": {}, "AUX": {}, "NUL": {},
	"COM1": {}, "COM2": {}, "COM3": {}, "COM4": {}, "COM5": {},
	"COM6": {}, "COM7": {}, "COM8": {}, "COM9": {},
	"LPT1": {}, "LPT2": {}, "LPT3": {}, "LPT4": {}, "LPT5": {},
	"LPT6": {}, "LPT7": {}, "LPT8": {}, "LPT9": {},
}

// CleanName turns a remote title into a single safe path component.
//
// Everything it receives is attacker-controlled: album, playlist and track
// titles come straight from provider APIs, and on SoundCloud and Tidal they are
// user-generated. The result is joined onto the download root by the caller, so
// it must never be "." or ".." (which would escape or collapse into the parent
// directory) and must never contain a separator.
func CleanName(in string, cfg Config) string {
	s := strings.TrimSpace(in)
	s = invalidPathRe.ReplaceAllString(s, "_")
	if cfg.RestrictCharacters {
		r := make([]rune, 0, len(s))
		for _, ch := range s {
			if ch >= 32 && ch <= 126 {
				r = append(r, ch)
			}
		}
		s = string(r)
	}
	if cfg.TruncateTo > 0 {
		runes := []rune(s)
		if len(runes) > cfg.TruncateTo {
			s = string(runes[:cfg.TruncateTo])
		}
	}
	s = truncateBytes(s, maxNameBytes)

	// Trailing dots and spaces are stripped by Windows, which turns "foo." into
	// "foo" and makes ".." survive as a traversal component.
	s = strings.TrimRight(s, ". ")
	s = strings.TrimSpace(s)

	if s == "" || isAllDots(s) {
		return "Unknown"
	}
	if _, bad := windowsReserved[strings.ToUpper(s)]; bad {
		return s + "_"
	}
	return s
}

// truncateBytes shortens s to at most max bytes without splitting a rune.
func truncateBytes(s string, max int) string {
	if len(s) <= max {
		return s
	}
	cut := s[:max]
	for len(cut) > 0 && !utf8.ValidString(cut) {
		cut = cut[:len(cut)-1]
	}
	return cut
}

// isAllDots reports whether s consists only of "." characters, i.e. "." or
// ".." or any longer run that some filesystems also treat as relative.
func isAllDots(s string) bool {
	for _, ch := range s {
		if ch != '.' {
			return false
		}
	}
	return len(s) > 0
}

func YearFromDate(date string) string {
	if len(date) >= 4 {
		prefix := date[:4]
		for _, ch := range prefix {
			if !unicode.IsDigit(ch) {
				return "Unknown"
			}
		}
		return prefix
	}
	return "Unknown"
}