// Command chdump 从 PVF 里取出 etc/channel_info.chn.str 并按需过滤，
// 用来回答「某个频道在游戏里显示的中文名是什么」。
//
// 用法:
//
//	go run ./cmd/scratch/chdump -pvf <Script.pvf> [-path etc/channel_info.chn.str] [-grep dname_0]
package main

import (
	"encoding/binary"
	"flag"
	"fmt"
	"os"
	"strings"
	"unicode/utf16"

	"longheng.io/server/internal/platform/pvf"
)

func main() {
	pvfPath := flag.String("pvf", "", "Script.pvf 路径")
	relPath := flag.String("path", "etc/channel_info.chn.str", "PVF 内的相对路径")
	grep := flag.String("grep", "", "只打印包含该子串的行")
	limit := flag.Int("limit", 200, "最多打印多少行")
	flag.Parse()
	if *pvfPath == "" {
		fmt.Fprintln(os.Stderr, "需要 -pvf")
		os.Exit(2)
	}
	archive, err := pvf.Open(*pvfPath)
	if err != nil {
		fmt.Fprintln(os.Stderr, "打开 PVF 失败:", err)
		os.Exit(1)
	}
	raw, err := archive.ReadRaw(*relPath)
	if err != nil {
		fmt.Fprintln(os.Stderr, "读取失败:", err)
		os.Exit(1)
	}
	fmt.Printf("路径=%s 原始大小=%d 字节\n", *relPath, len(raw))
	if len(raw) >= 16 {
		fmt.Printf("前 16 字节: % X\n", raw[:16])
	}
	text := decodeBest(raw)
	lines := strings.Split(text, "\n")
	printed := 0
	for i, line := range lines {
		line = strings.TrimRight(line, "\r")
		if line == "" {
			continue
		}
		if *grep != "" && !strings.Contains(line, *grep) {
			continue
		}
		fmt.Printf("%4d: %s\n", i+1, line)
		printed++
		if printed >= *limit {
			fmt.Println("... (达到 limit 上限)")
			break
		}
	}
	if printed == 0 {
		fmt.Println("(没有匹配的行)")
	}
}

// decodeBest 依次尝试 UTF-8、UTF-16LE，选能解出更多可打印字符的那个。
func decodeBest(raw []byte) string {
	utf8Text := string(raw)
	utf16Text := decodeUTF16LE(raw)
	if score(utf16Text) > score(utf8Text) {
		return utf16Text
	}
	return utf8Text
}

func decodeUTF16LE(raw []byte) string {
	if len(raw) < 2 {
		return ""
	}
	units := make([]uint16, 0, len(raw)/2)
	for i := 0; i+1 < len(raw); i += 2 {
		units = append(units, binary.LittleEndian.Uint16(raw[i:i+2]))
	}
	return string(utf16.Decode(units))
}

// score 粗略衡量文本「像不像正常文本」：可打印/常见字符越多分越高。
func score(text string) int {
	total := 0
	good := 0
	for _, r := range text {
		total++
		switch {
		case r == '\n' || r == '\r' || r == '\t':
			good++
		case r >= 0x20 && r <= 0x7E:
			good++
		case r >= 0x4E00 && r <= 0x9FFF: // CJK
			good++
		case r >= 0x3000 && r <= 0x303F: // CJK 标点
		case r >= 0xFF00 && r <= 0xFFEF: // 全角
			good++
		}
	}
	if total == 0 {
		return 0
	}
	return good * 1000 / total
}