kenaito-media-picker/internal/scanner/scanner.go

// Package scanner 负责递归扫描文件、读取 MIME 并按类型过滤。
package scanner

import (
	"fmt"
	"os"
	"path/filepath"
	"strings"
	"time"

	"kenaito-media-picker/internal/probe"

	"github.com/gabriel-vasile/mimetype"
)

// File 扫描到的文件信息
type File struct {
	Path string // 完整路径
	Name string // 文件名（含扩展名）
	Ext  string // 扩展名（含点）
	Mime string // MIME 类型
}

// Result 过滤结果
type Result struct {
	Files  []File // 全部文件
	Images []File // 图片文件
	Media  []File // 待 ffprobe 确认的媒体候选
}

// Options 扫描选项
type Options struct {
	RootDir      string   // 扫描根目录
	SkipDir      string   // 白名单输出目录，整目录跳过
	SkipDirNames []string // 按目录名跳过（如 KMP已整理），任意位置命中即跳过
	SkipNames    []string // 按文件名跳过（数据库、ffprobe 等）
	OnProgress   func(done, total int)
}

// Scan 递归扫描目录并读取每个文件的 MIME 信息
func Scan(opts Options) (*Result, error) {
	res := &Result{}
	// 遍历阶段尚不知道总文件数，先按「已发现文件数」实时上报（total=0），
	// 并节流上报频率，避免目录很大时界面长时间无动静或推送过密。
	walkCount := 0
	lastReport := time.Now()
	err := filepath.Walk(opts.RootDir, func(currentPath string, info os.FileInfo, err error) error {
		if err != nil {
			// 单个路径访问失败（权限不足/损坏）不中止整体扫描，仅告警后跳过该子树
			fmt.Printf("=== 扫描跳过(无法访问): %s, 错误: %v\n", currentPath, err)
			return nil
		}
		if info.IsDir() {
			if opts.SkipDir != "" && currentPath == opts.SkipDir {
				return filepath.SkipDir
			}
			for _, n := range opts.SkipDirNames {
				if info.Name() == n {
					return filepath.SkipDir
				}
			}
			return nil
		}
		for _, n := range opts.SkipNames {
			if info.Name() == n {
				return nil
			}
		}
		if info.Size() == 0 {
			// 空文件不可能是媒体，直接跳过，避免无谓的 MIME/ffprobe 探测
			return nil
		}
		res.Files = append(res.Files, File{
			Path: currentPath,
			Name: info.Name(),
			Ext:  filepath.Ext(currentPath),
		})
		walkCount++
		if opts.OnProgress != nil && time.Since(lastReport) >= 100*time.Millisecond {
			lastReport = time.Now()
			opts.OnProgress(walkCount, 0)
		}
		return nil
	})
	if err != nil {
		return nil, err
	}

	total := len(res.Files)
	fmt.Printf("=== 文件总数: %d \n", total)
	// 遍历结束、开始 MIME 探测前，上报一次完整总数
	if opts.OnProgress != nil {
		opts.OnProgress(total, total)
	}
	for i := range res.Files {
		res.Files[i].Mime = detectMime(res.Files[i].Path)
		if opts.OnProgress != nil {
			opts.OnProgress(i+1, total)
		}
	}
	return res, nil
}

// Filter 将文件分为图片与媒体候选
func (r *Result) Filter() {
	for _, f := range r.Files {
		if strings.Contains(f.Mime, "image/") || probe.IsSupportImage(f.Ext) || strings.EqualFold(f.Ext, ".psd") {
			r.Images = append(r.Images, f)
			continue
		}
		if isPotentialMedia(f.Mime) {
			r.Media = append(r.Media, f)
			continue
		}
		// 明显的非媒体文件（代码/文档/压缩包等）直接不处理
	}
}

// detectMime 读取文件的 MIME 类型
func detectMime(filePath string) string {
	mt, err := mimetype.DetectFile(filePath)
	if err != nil {
		return ""
	}
	return mt.String()
}

// IsExplicitMediaMime 判断 MIME 是否明确属于视频/音频（区别于 octet-stream/空 等无法确定的情况）
func IsExplicitMediaMime(fileMime string) bool {
	return strings.HasPrefix(fileMime, "video/") || strings.HasPrefix(fileMime, "audio/")
}

// isPotentialMedia 判断是否可能是媒体文件（交由 ffprobe 进一步确认）
func isPotentialMedia(fileMime string) bool {
	if fileMime == "" || fileMime == "application/octet-stream" {
		return true
	}
	if strings.HasPrefix(fileMime, "text/") {
		return false
	}
	if strings.HasPrefix(fileMime, "video/") || strings.HasPrefix(fileMime, "audio/") || strings.HasPrefix(fileMime, "image/") {
		return true
	}
	switch fileMime {
	case "application/zip", "application/gzip", "application/x-gzip", "application/pdf",
		"application/json", "application/xml", "application/x-httpd-php", "application/x-sh",
		"application/javascript", "application/x-7z-compressed", "application/x-rar-compressed",
		"application/x-tar", "application/x-bzip2", "application/x-archive", "application/vnd.sqlite3",
		"application/vnd.microsoft.portable-executable", "application/x-dosexec", "application/x-executable",
		"application/vnd.openxmlformats-officedocument.wordprocessingml.document",
		"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
		"application/vnd.openxmlformats-officedocument.presentationml.presentation",
		"application/msword", "application/vnd.ms-excel", "application/vnd.ms-powerpoint":
		return false
	}
	return true
}