analysis

type AnalysisResult struct {
	RootDir        string                 `json:"root_dir"`
	ScanStartTime  time.Time              `json:"scan_start_time"`
	ScanEndTime    time.Time              `json:"scan_end_time"`
	TotalFiles     int                    `json:"total_files"`
	TotalDirs      int                    `json:"total_dirs"`
	TotalSize      int64                  `json:"total_size"`
	FileEntries    []*FileEntry           `json:"file_entries,omitempty"` // May be omitted in summary-only reports
	TypeSummary    map[string]*TypeStats  `json:"type_summary"`           // Keyed by type label
	DuplicateSets  []*DuplicateSet        `json:"duplicate_sets,omitempty"`
	MonthlySummary map[string]*MonthStats `json:"monthly_summary"` // Keyed by "YYYY-MM"
	ToolStatus     map[string]*ToolInfo   `json:"tool_status"`
	OverallErrors  []string               `json:"overall_errors,omitempty"`
	Config         *config.Config         `json:"config_used"` // Include the config used for this run
}

type Analyzer interface {
	// Name returns a unique identifier for the analyzer (e.g., "MagikaTypeAnalyzer").
	Name() string

	// Type returns whether this is a file-level or aggregate analyzer.
	Type() AnalyzerType

	// DependsOn returns names of analyzers that must run before this one.
	DependsOn() []string

	// Analyze performs the analysis.
	// For file-level analyzers (Type() == FileAnalyzer), it operates on a single FileEntry.
	// For aggregate analyzers (Type() == AggregateAnalyzer), the entry parameter is ignored.
	// Both types update the AnalysisResult as needed.
	// The ToolProvider gives access to configured external tool runners.
	Analyze(ctx context.Context, cfg *config.Config, toolProvider ToolProvider, result *AnalysisResult, entry *FileEntry) error
}

type AnalyzerType int

type DuplicateSet struct {
	ID          string   `json:"id"`           // Hash or external tool ID
	FilePaths   []string `json:"file_paths"`   // Relative paths
	Size        int64    `json:"size"`         // Size of one file
	Count       int      `json:"count"`        // Number of files in the set
	WastedSpace int64    `json:"wasted_space"` // (Count-1) * Size
}

type FileEntry struct {
	Path     string      `json:"path"` // Relative to TargetDir
	FullPath string      `json:"-"`    // Absolute path
	IsDir    bool        `json:"is_dir"`
	Size     int64       `json:"size"`
	ModTime  time.Time   `json:"mod_time"`
	Mode     fs.FileMode `json:"-"`

	TypeInfo map[string]any    `json:"type_info,omitempty"` // Source ("magika", "file"), Label, Group, MIME
	Metadata map[string]any    `json:"metadata,omitempty"`  // Source ("exiftool"), Key-Value pairs
	Hashes   map[string]string `json:"hashes,omitempty"`    // Algorithm ("md5", "sha256") -> Hash string
	Tags     []string          `json:"tags,omitempty"`      // "LargeFile", "RecentFile", "DuplicateSetID:xyz"
	Error    string            `json:"error,omitempty"`     // Record file-specific processing errors
}

type GlobMatcher struct {
	// contains filtered or unexported fields
}

type MagikaTypeAnalyzer struct {
	// contains filtered or unexported fields
}

type MonthStats struct {
	YearMonth string `json:"year_month"` // YYYY-MM
	Count     int    `json:"count"`
	Size      int64  `json:"size"`
}

type PathMatcher interface {
	Match(path string) bool
}

type Registry struct {
	// contains filtered or unexported fields
}

type Runner struct {
	// contains filtered or unexported fields
}

type ToolInfo struct {
	Name      string `json:"name"`
	Path      string `json:"path,omitempty"`
	Available bool   `json:"available"`
	Error     string `json:"error,omitempty"`
}

type ToolProvider interface {
	GetToolRunner(name string) (tools.Runner, bool)
}

type TypeStats struct {
	Label string   `json:"label"`
	Count int      `json:"count"`
	Size  int64    `json:"size"`
	Paths []string `json:"-"` // Temporary storage during aggregation
}