vaultik/internal/snapshot/scanner.go

package snapshot

import (
	"context"
	"database/sql"
	"errors"
	"fmt"
	"os"
	"path/filepath"
	"strings"
	"sync"
	"time"

	"git.eeqj.de/sneak/vaultik/internal/blob"
	"git.eeqj.de/sneak/vaultik/internal/chunker"
	"git.eeqj.de/sneak/vaultik/internal/database"
	"git.eeqj.de/sneak/vaultik/internal/log"
	"git.eeqj.de/sneak/vaultik/internal/storage"
	"git.eeqj.de/sneak/vaultik/internal/types"
	"github.com/dustin/go-humanize"
	"github.com/gobwas/glob"
	"github.com/spf13/afero"
)

// FileToProcess holds information about a file that needs processing
type FileToProcess struct {
	Path     string
	FileInfo os.FileInfo
	File     *database.File
}

// pendingFileData holds all data needed to commit a file to the database
type pendingFileData struct {
	file       *database.File
	fileChunks []database.FileChunk
	chunkFiles []database.ChunkFile
}

// compiledPattern holds a compiled glob pattern and whether it's anchored
type compiledPattern struct {
	pattern  glob.Glob
	anchored bool // If true, only matches from root of source dir
	original string
}

// Scanner scans directories and populates the database with file and chunk information
type Scanner struct {
	fs                afero.Fs
	chunker           *chunker.Chunker
	packer            *blob.Packer
	repos             *database.Repositories
	storage           storage.Storer
	maxBlobSize       int64
	compressionLevel  int
	ageRecipient      string
	snapshotID        string            // Current snapshot being processed
	currentSourcePath string            // Current source directory being scanned (for restore path stripping)
	exclude           []string          // Glob patterns for files/directories to exclude
	compiledExclude   []compiledPattern // Compiled glob patterns
	progress          *ProgressReporter
	skipErrors        bool // Skip file read errors (log loudly but continue)

	// In-memory cache of known chunk hashes for fast existence checks
	knownChunks   map[string]struct{}
	knownChunksMu sync.RWMutex

	// Pending chunk hashes - chunks that have been added to packer but not yet committed to DB
	// When a blob finalizes, the committed chunks are removed from this set
	pendingChunkHashes   map[string]struct{}
	pendingChunkHashesMu sync.Mutex

	// Pending file data buffer for batch insertion
	// Files are flushed when all their chunks have been committed to DB
	pendingFiles   []pendingFileData
	pendingFilesMu sync.Mutex

	// Mutex for coordinating blob creation
	packerMu sync.Mutex // Blocks chunk production during blob creation

	// Context for cancellation
	scanCtx context.Context
}

// ScannerConfig contains configuration for the scanner
type ScannerConfig struct {
	FS               afero.Fs
	ChunkSize        int64
	Repositories     *database.Repositories
	Storage          storage.Storer
	MaxBlobSize      int64
	CompressionLevel int
	AgeRecipients    []string // Optional, empty means no encryption
	EnableProgress   bool     // Enable progress reporting
	Exclude          []string // Glob patterns for files/directories to exclude
	SkipErrors       bool     // Skip file read errors (log loudly but continue)
}

// ScanResult contains the results of a scan operation
type ScanResult struct {
	FilesScanned  int
	FilesSkipped  int
	FilesDeleted  int
	BytesScanned  int64
	BytesSkipped  int64
	BytesDeleted  int64
	ChunksCreated int
	BlobsCreated  int
	StartTime     time.Time
	EndTime       time.Time
}

// NewScanner creates a new scanner instance
func NewScanner(cfg ScannerConfig) *Scanner {
	// Create encryptor (required for blob packing)
	if len(cfg.AgeRecipients) == 0 {
		log.Error("No age recipients configured - encryption is required")
		return nil
	}

	// Create blob packer with encryption
	packerCfg := blob.PackerConfig{
		MaxBlobSize:      cfg.MaxBlobSize,
		CompressionLevel: cfg.CompressionLevel,
		Recipients:       cfg.AgeRecipients,
		Repositories:     cfg.Repositories,
		Fs:               cfg.FS,
	}
	packer, err := blob.NewPacker(packerCfg)
	if err != nil {
		log.Error("Failed to create packer", "error", err)
		return nil
	}

	var progress *ProgressReporter
	if cfg.EnableProgress {
		progress = NewProgressReporter()
	}

	// Compile exclude patterns
	compiledExclude := compileExcludePatterns(cfg.Exclude)

	return &Scanner{
		fs:                 cfg.FS,
		chunker:            chunker.NewChunker(cfg.ChunkSize),
		packer:             packer,
		repos:              cfg.Repositories,
		storage:            cfg.Storage,
		maxBlobSize:        cfg.MaxBlobSize,
		compressionLevel:   cfg.CompressionLevel,
		ageRecipient:       strings.Join(cfg.AgeRecipients, ","),
		exclude:            cfg.Exclude,
		compiledExclude:    compiledExclude,
		progress:           progress,
		skipErrors:         cfg.SkipErrors,
		pendingChunkHashes: make(map[string]struct{}),
	}
}

// Scan scans a directory and populates the database
func (s *Scanner) Scan(ctx context.Context, path string, snapshotID string) (*ScanResult, error) {
	s.snapshotID = snapshotID
	s.currentSourcePath = path // Store source path for file records (used during restore)
	s.scanCtx = ctx
	result := &ScanResult{
		StartTime: time.Now().UTC(),
	}

	// Set blob handler for concurrent upload
	if s.storage != nil {
		log.Debug("Setting blob handler for storage uploads")
		s.packer.SetBlobHandler(s.handleBlobReady)
	} else {
		log.Debug("No storage configured, blobs will not be uploaded")
	}

	// Start progress reporting if enabled
	if s.progress != nil {
		s.progress.Start()
		defer s.progress.Stop()
	}

	// Phase 0: Load known files and chunks from database into memory for fast lookup
	fmt.Println("Loading known files from database...")
	knownFiles, err := s.loadKnownFiles(ctx, path)
	if err != nil {
		return nil, fmt.Errorf("loading known files: %w", err)
	}
	fmt.Printf("Loaded %s known files from database\n", formatNumber(len(knownFiles)))

	fmt.Println("Loading known chunks from database...")
	if err := s.loadKnownChunks(ctx); err != nil {
		return nil, fmt.Errorf("loading known chunks: %w", err)
	}
	fmt.Printf("Loaded %s known chunks from database\n", formatNumber(len(s.knownChunks)))

	// Phase 1: Scan directory, collect files to process, and track existing files
	// (builds existingFiles map during walk to avoid double traversal)
	log.Info("Phase 1/3: Scanning directory structure")
	existingFiles := make(map[string]struct{})
	scanResult, err := s.scanPhase(ctx, path, result, existingFiles, knownFiles)
	if err != nil {
		return nil, fmt.Errorf("scan phase failed: %w", err)
	}
	filesToProcess := scanResult.FilesToProcess

	// Phase 1b: Detect deleted files by comparing DB against scanned files
	if err := s.detectDeletedFilesFromMap(ctx, knownFiles, existingFiles, result); err != nil {
		return nil, fmt.Errorf("detecting deleted files: %w", err)
	}

	// Phase 1c: Associate unchanged files with this snapshot (no new records needed)
	if len(scanResult.UnchangedFileIDs) > 0 {
		fmt.Printf("Associating %s unchanged files with snapshot...\n", formatNumber(len(scanResult.UnchangedFileIDs)))
		if err := s.batchAddFilesToSnapshot(ctx, scanResult.UnchangedFileIDs); err != nil {
			return nil, fmt.Errorf("associating unchanged files: %w", err)
		}
	}

	// Calculate total size to process
	var totalSizeToProcess int64
	for _, file := range filesToProcess {
		totalSizeToProcess += file.FileInfo.Size()
	}

	// Update progress with total size and file count
	if s.progress != nil {
		s.progress.SetTotalSize(totalSizeToProcess)
		s.progress.GetStats().TotalFiles.Store(int64(len(filesToProcess)))
	}

	log.Info("Phase 1 complete",
		"total_files", len(filesToProcess),
		"total_size", humanize.Bytes(uint64(totalSizeToProcess)),
		"files_skipped", result.FilesSkipped,
		"bytes_skipped", humanize.Bytes(uint64(result.BytesSkipped)))

	// Print scan summary
	fmt.Printf("Scan complete: %s examined (%s), %s to process (%s)",
		formatNumber(result.FilesScanned),
		humanize.Bytes(uint64(totalSizeToProcess+result.BytesSkipped)),
		formatNumber(len(filesToProcess)),
		humanize.Bytes(uint64(totalSizeToProcess)))
	if result.FilesDeleted > 0 {
		fmt.Printf(", %s deleted (%s)",
			formatNumber(result.FilesDeleted),
			humanize.Bytes(uint64(result.BytesDeleted)))
	}
	fmt.Println()

	// Phase 2: Process files and create chunks
	if len(filesToProcess) > 0 {
		fmt.Printf("Processing %s files...\n", formatNumber(len(filesToProcess)))
		log.Info("Phase 2/3: Creating snapshot (chunking, compressing, encrypting, and uploading blobs)")
		if err := s.processPhase(ctx, filesToProcess, result); err != nil {
			return nil, fmt.Errorf("process phase failed: %w", err)
		}
	} else {
		fmt.Printf("No files need processing. Creating metadata-only snapshot.\n")
		log.Info("Phase 2/3: Skipping (no files need processing, metadata-only snapshot)")
	}

	// Get final stats from packer
	blobs := s.packer.GetFinishedBlobs()
	result.BlobsCreated += len(blobs)

	// Query database for actual blob count created during this snapshot
	// The database is authoritative, especially for concurrent blob uploads
	// We count uploads rather than all snapshot_blobs to get only NEW blobs
	if s.snapshotID != "" {
		uploadCount, err := s.repos.Uploads.GetCountBySnapshot(ctx, s.snapshotID)
		if err != nil {
			log.Warn("Failed to query upload count from database", "error", err)
		} else {
			result.BlobsCreated = int(uploadCount)
		}
	}

	result.EndTime = time.Now().UTC()
	return result, nil
}

// loadKnownFiles loads all known files from the database into a map for fast lookup
// This avoids per-file database queries during the scan phase
func (s *Scanner) loadKnownFiles(ctx context.Context, path string) (map[string]*database.File, error) {
	files, err := s.repos.Files.ListByPrefix(ctx, path)
	if err != nil {
		return nil, fmt.Errorf("listing files by prefix: %w", err)
	}

	result := make(map[string]*database.File, len(files))
	for _, f := range files {
		result[f.Path.String()] = f
	}

	return result, nil
}

// loadKnownChunks loads all known chunk hashes from the database into a map for fast lookup
// This avoids per-chunk database queries during file processing
func (s *Scanner) loadKnownChunks(ctx context.Context) error {
	chunks, err := s.repos.Chunks.List(ctx)
	if err != nil {
		return fmt.Errorf("listing chunks: %w", err)
	}

	s.knownChunksMu.Lock()
	s.knownChunks = make(map[string]struct{}, len(chunks))
	for _, c := range chunks {
		s.knownChunks[c.ChunkHash.String()] = struct{}{}
	}
	s.knownChunksMu.Unlock()

	return nil
}

// chunkExists checks if a chunk hash exists in the in-memory cache
func (s *Scanner) chunkExists(hash string) bool {
	s.knownChunksMu.RLock()
	_, exists := s.knownChunks[hash]
	s.knownChunksMu.RUnlock()
	return exists
}

// addKnownChunk adds a chunk hash to the in-memory cache
func (s *Scanner) addKnownChunk(hash string) {
	s.knownChunksMu.Lock()
	s.knownChunks[hash] = struct{}{}
	s.knownChunksMu.Unlock()
}

// addPendingChunkHash marks a chunk as pending (not yet committed to DB)
func (s *Scanner) addPendingChunkHash(hash string) {
	s.pendingChunkHashesMu.Lock()
	s.pendingChunkHashes[hash] = struct{}{}
	s.pendingChunkHashesMu.Unlock()
}

// removePendingChunkHashes removes committed chunk hashes from the pending set
func (s *Scanner) removePendingChunkHashes(hashes []string) {
	log.Debug("removePendingChunkHashes: starting", "count", len(hashes))
	start := time.Now()
	s.pendingChunkHashesMu.Lock()
	for _, hash := range hashes {
		delete(s.pendingChunkHashes, hash)
	}
	s.pendingChunkHashesMu.Unlock()
	log.Debug("removePendingChunkHashes: done", "count", len(hashes), "duration", time.Since(start))
}

// isChunkPending returns true if the chunk is still pending (not yet committed to DB)
func (s *Scanner) isChunkPending(hash string) bool {
	s.pendingChunkHashesMu.Lock()
	_, pending := s.pendingChunkHashes[hash]
	s.pendingChunkHashesMu.Unlock()
	return pending
}

// addPendingFile adds a file to the pending buffer
// Files are NOT auto-flushed here - they are flushed when their chunks are committed
// (in handleBlobReady after blob finalize)
func (s *Scanner) addPendingFile(_ context.Context, data pendingFileData) {
	s.pendingFilesMu.Lock()
	s.pendingFiles = append(s.pendingFiles, data)
	s.pendingFilesMu.Unlock()
}

// flushPendingFiles writes all pending files to the database in a single transaction
func (s *Scanner) flushPendingFiles(ctx context.Context) error {
	s.pendingFilesMu.Lock()
	files := s.pendingFiles
	s.pendingFiles = nil
	s.pendingFilesMu.Unlock()

	if len(files) == 0 {
		return nil
	}

	return s.repos.WithTx(ctx, func(txCtx context.Context, tx *sql.Tx) error {
		for _, data := range files {
			// Create or update the file record
			if err := s.repos.Files.Create(txCtx, tx, data.file); err != nil {
				return fmt.Errorf("creating file record: %w", err)
			}

			// Delete any existing file_chunks and chunk_files for this file
			if err := s.repos.FileChunks.DeleteByFileID(txCtx, tx, data.file.ID); err != nil {
				return fmt.Errorf("deleting old file chunks: %w", err)
			}
			if err := s.repos.ChunkFiles.DeleteByFileID(txCtx, tx, data.file.ID); err != nil {
				return fmt.Errorf("deleting old chunk files: %w", err)
			}

			// Create file-chunk mappings
			for i := range data.fileChunks {
				if err := s.repos.FileChunks.Create(txCtx, tx, &data.fileChunks[i]); err != nil {
					return fmt.Errorf("creating file chunk: %w", err)
				}
			}

			// Create chunk-file mappings
			for i := range data.chunkFiles {
				if err := s.repos.ChunkFiles.Create(txCtx, tx, &data.chunkFiles[i]); err != nil {
					return fmt.Errorf("creating chunk file: %w", err)
				}
			}

			// Add file to snapshot
			if err := s.repos.Snapshots.AddFileByID(txCtx, tx, s.snapshotID, data.file.ID); err != nil {
				return fmt.Errorf("adding file to snapshot: %w", err)
			}
		}
		return nil
	})
}

// flushAllPending flushes all pending files to the database
func (s *Scanner) flushAllPending(ctx context.Context) error {
	return s.flushPendingFiles(ctx)
}

// flushCompletedPendingFiles flushes only files whose chunks are all committed to DB
// Files with pending chunks are kept in the queue for later flushing
func (s *Scanner) flushCompletedPendingFiles(ctx context.Context) error {
	flushStart := time.Now()
	log.Debug("flushCompletedPendingFiles: starting")

	log.Debug("flushCompletedPendingFiles: acquiring pendingFilesMu lock")
	s.pendingFilesMu.Lock()
	log.Debug("flushCompletedPendingFiles: acquired lock", "pending_files", len(s.pendingFiles))

	// Separate files into complete (can flush) and incomplete (keep pending)
	var canFlush []pendingFileData
	var stillPending []pendingFileData

	log.Debug("flushCompletedPendingFiles: checking which files can flush")
	checkStart := time.Now()
	for _, data := range s.pendingFiles {
		allChunksCommitted := true
		for _, fc := range data.fileChunks {
			if s.isChunkPending(fc.ChunkHash.String()) {
				allChunksCommitted = false
				break
			}
		}
		if allChunksCommitted {
			canFlush = append(canFlush, data)
		} else {
			stillPending = append(stillPending, data)
		}
	}
	log.Debug("flushCompletedPendingFiles: check done", "duration", time.Since(checkStart), "can_flush", len(canFlush), "still_pending", len(stillPending))

	s.pendingFiles = stillPending
	s.pendingFilesMu.Unlock()
	log.Debug("flushCompletedPendingFiles: released lock")

	if len(canFlush) == 0 {
		log.Debug("flushCompletedPendingFiles: nothing to flush")
		return nil
	}

	log.Debug("Flushing completed files after blob finalize",
		"files_to_flush", len(canFlush),
		"files_still_pending", len(stillPending))

	// Collect all data for batch operations
	log.Debug("flushCompletedPendingFiles: collecting data for batch ops")
	collectStart := time.Now()
	var allFileChunks []database.FileChunk
	var allChunkFiles []database.ChunkFile
	var allFileIDs []types.FileID
	var allFiles []*database.File

	for _, data := range canFlush {
		allFileChunks = append(allFileChunks, data.fileChunks...)
		allChunkFiles = append(allChunkFiles, data.chunkFiles...)
		allFileIDs = append(allFileIDs, data.file.ID)
		allFiles = append(allFiles, data.file)
	}
	log.Debug("flushCompletedPendingFiles: collected data",
		"duration", time.Since(collectStart),
		"file_chunks", len(allFileChunks),
		"chunk_files", len(allChunkFiles),
		"files", len(allFiles))

	// Flush the complete files using batch operations
	log.Debug("flushCompletedPendingFiles: starting transaction")
	txStart := time.Now()
	err := s.repos.WithTx(ctx, func(txCtx context.Context, tx *sql.Tx) error {
		log.Debug("flushCompletedPendingFiles: inside transaction")

		// Batch delete old file_chunks and chunk_files
		log.Debug("flushCompletedPendingFiles: deleting old file_chunks")
		opStart := time.Now()
		if err := s.repos.FileChunks.DeleteByFileIDs(txCtx, tx, allFileIDs); err != nil {
			return fmt.Errorf("batch deleting old file chunks: %w", err)
		}
		log.Debug("flushCompletedPendingFiles: deleted file_chunks", "duration", time.Since(opStart))

		log.Debug("flushCompletedPendingFiles: deleting old chunk_files")
		opStart = time.Now()
		if err := s.repos.ChunkFiles.DeleteByFileIDs(txCtx, tx, allFileIDs); err != nil {
			return fmt.Errorf("batch deleting old chunk files: %w", err)
		}
		log.Debug("flushCompletedPendingFiles: deleted chunk_files", "duration", time.Since(opStart))

		// Batch create/update file records
		log.Debug("flushCompletedPendingFiles: creating files")
		opStart = time.Now()
		if err := s.repos.Files.CreateBatch(txCtx, tx, allFiles); err != nil {
			return fmt.Errorf("batch creating file records: %w", err)
		}
		log.Debug("flushCompletedPendingFiles: created files", "duration", time.Since(opStart))

		// Batch insert file_chunks
		log.Debug("flushCompletedPendingFiles: inserting file_chunks")
		opStart = time.Now()
		if err := s.repos.FileChunks.CreateBatch(txCtx, tx, allFileChunks); err != nil {
			return fmt.Errorf("batch creating file chunks: %w", err)
		}
		log.Debug("flushCompletedPendingFiles: inserted file_chunks", "duration", time.Since(opStart))

		// Batch insert chunk_files
		log.Debug("flushCompletedPendingFiles: inserting chunk_files")
		opStart = time.Now()
		if err := s.repos.ChunkFiles.CreateBatch(txCtx, tx, allChunkFiles); err != nil {
			return fmt.Errorf("batch creating chunk files: %w", err)
		}
		log.Debug("flushCompletedPendingFiles: inserted chunk_files", "duration", time.Since(opStart))

		// Batch add files to snapshot
		log.Debug("flushCompletedPendingFiles: adding files to snapshot")
		opStart = time.Now()
		if err := s.repos.Snapshots.AddFilesByIDBatch(txCtx, tx, s.snapshotID, allFileIDs); err != nil {
			return fmt.Errorf("batch adding files to snapshot: %w", err)
		}
		log.Debug("flushCompletedPendingFiles: added files to snapshot", "duration", time.Since(opStart))

		log.Debug("flushCompletedPendingFiles: transaction complete")
		return nil
	})
	log.Debug("flushCompletedPendingFiles: transaction done", "duration", time.Since(txStart))
	log.Debug("flushCompletedPendingFiles: total duration", "duration", time.Since(flushStart))
	return err
}

// ScanPhaseResult contains the results of the scan phase
type ScanPhaseResult struct {
	FilesToProcess   []*FileToProcess
	UnchangedFileIDs []types.FileID // IDs of unchanged files to associate with snapshot
}

// scanPhase performs the initial directory scan to identify files to process
// It uses the pre-loaded knownFiles map for fast change detection without DB queries
// It also populates existingFiles map for deletion detection
// Returns files needing processing and IDs of unchanged files for snapshot association
func (s *Scanner) scanPhase(ctx context.Context, path string, result *ScanResult, existingFiles map[string]struct{}, knownFiles map[string]*database.File) (*ScanPhaseResult, error) {
	// Use known file count as estimate for progress (accurate for subsequent backups)
	estimatedTotal := int64(len(knownFiles))

	var filesToProcess []*FileToProcess
	var unchangedFileIDs []types.FileID // Just IDs - no new records needed
	var mu sync.Mutex

	// Set up periodic status output
	startTime := time.Now()
	lastStatusTime := time.Now()
	statusInterval := 15 * time.Second
	var filesScanned int64

	log.Debug("Starting directory walk", "path", path)
	err := afero.Walk(s.fs, path, func(filePath string, info os.FileInfo, err error) error {
		if err != nil {
			if s.skipErrors {
				log.Error("ERROR: Failed to access file (skipping due to --skip-errors)", "path", filePath, "error", err)
				fmt.Printf("ERROR: Failed to access %s: %v (skipping)\n", filePath, err)
				return nil // Continue scanning
			}
			log.Debug("Error accessing filesystem entry", "path", filePath, "error", err)
			return err
		}

		// Check context cancellation
		select {
		case <-ctx.Done():
			return ctx.Err()
		default:
		}

		// Check exclude patterns - for directories, skip the entire subtree
		if s.shouldExclude(filePath, path) {
			if info.IsDir() {
				return filepath.SkipDir
			}
			return nil
		}

		// Skip non-regular files for processing (but still count them)
		if !info.Mode().IsRegular() {
			return nil
		}

		// Track this file as existing (for deletion detection)
		existingFiles[filePath] = struct{}{}

		// Check file against in-memory map (no DB query!)
		file, needsProcessing := s.checkFileInMemory(filePath, info, knownFiles)

		mu.Lock()
		if needsProcessing {
			// New or changed file - will create record after processing
			filesToProcess = append(filesToProcess, &FileToProcess{
				Path:     filePath,
				FileInfo: info,
				File:     file,
			})
		} else if !file.ID.IsZero() {
			// Unchanged file with existing ID - just need snapshot association
			unchangedFileIDs = append(unchangedFileIDs, file.ID)
		}
		filesScanned++
		changedCount := len(filesToProcess)
		mu.Unlock()

		// Update result stats
		if needsProcessing {
			result.BytesScanned += info.Size()
		} else {
			result.FilesSkipped++
			result.BytesSkipped += info.Size()
		}
		result.FilesScanned++

		// Output periodic status
		if time.Since(lastStatusTime) >= statusInterval {
			elapsed := time.Since(startTime)
			rate := float64(filesScanned) / elapsed.Seconds()

			// Build status line - use estimate if available (not first backup)
			if estimatedTotal > 0 {
				// Show actual scanned vs estimate (may exceed estimate if files were added)
				pct := float64(filesScanned) / float64(estimatedTotal) * 100
				if pct > 100 {
					pct = 100 // Cap at 100% for display
				}
				remaining := estimatedTotal - filesScanned
				if remaining < 0 {
					remaining = 0
				}
				var eta time.Duration
				if rate > 0 && remaining > 0 {
					eta = time.Duration(float64(remaining)/rate) * time.Second
				}
				fmt.Printf("Scan: %s files (~%.0f%%), %s changed/new, %.0f files/sec, %s elapsed",
					formatNumber(int(filesScanned)),
					pct,
					formatNumber(changedCount),
					rate,
					elapsed.Round(time.Second))
				if eta > 0 {
					fmt.Printf(", ETA %s", eta.Round(time.Second))
				}
				fmt.Println()
			} else {
				// First backup - no estimate available
				fmt.Printf("Scan: %s files, %s changed/new, %.0f files/sec, %s elapsed\n",
					formatNumber(int(filesScanned)),
					formatNumber(changedCount),
					rate,
					elapsed.Round(time.Second))
			}
			lastStatusTime = time.Now()
		}

		return nil
	})

	if err != nil {
		return nil, err
	}

	return &ScanPhaseResult{
		FilesToProcess:   filesToProcess,
		UnchangedFileIDs: unchangedFileIDs,
	}, nil
}

// checkFileInMemory checks if a file needs processing using the in-memory map
// No database access is performed - this is purely CPU/memory work
func (s *Scanner) checkFileInMemory(path string, info os.FileInfo, knownFiles map[string]*database.File) (*database.File, bool) {
	// Get file stats
	stat, ok := info.Sys().(interface {
		Uid() uint32
		Gid() uint32
	})

	var uid, gid uint32
	if ok {
		uid = stat.Uid()
		gid = stat.Gid()
	}

	// Check against in-memory map first to get existing ID if available
	existingFile, exists := knownFiles[path]

	// Create file record with ID set upfront
	// For new files, generate UUID immediately so it's available for chunk associations
	// For existing files, reuse the existing ID
	var fileID types.FileID
	if exists {
		fileID = existingFile.ID
	} else {
		fileID = types.NewFileID()
	}

	file := &database.File{
		ID:         fileID,
		Path:       types.FilePath(path),
		SourcePath: types.SourcePath(s.currentSourcePath), // Store source directory for restore path stripping
		MTime:      info.ModTime(),
		CTime:      info.ModTime(), // afero doesn't provide ctime
		Size:       info.Size(),
		Mode:       uint32(info.Mode()),
		UID:        uid,
		GID:        gid,
	}

	// New file - needs processing
	if !exists {
		return file, true
	}

	// Check if file has changed
	if existingFile.Size != file.Size ||
		existingFile.MTime.Unix() != file.MTime.Unix() ||
		existingFile.Mode != file.Mode ||
		existingFile.UID != file.UID ||
		existingFile.GID != file.GID {
		return file, true
	}

	// File unchanged
	return file, false
}

// batchAddFilesToSnapshot adds existing file IDs to the snapshot association table
// This is used for unchanged files that already have records in the database
func (s *Scanner) batchAddFilesToSnapshot(ctx context.Context, fileIDs []types.FileID) error {
	const batchSize = 1000

	startTime := time.Now()
	lastStatusTime := time.Now()
	statusInterval := 5 * time.Second

	for i := 0; i < len(fileIDs); i += batchSize {
		// Check context cancellation
		select {
		case <-ctx.Done():
			return ctx.Err()
		default:
		}

		end := i + batchSize
		if end > len(fileIDs) {
			end = len(fileIDs)
		}
		batch := fileIDs[i:end]

		err := s.repos.WithTx(ctx, func(ctx context.Context, tx *sql.Tx) error {
			for _, fileID := range batch {
				if err := s.repos.Snapshots.AddFileByID(ctx, tx, s.snapshotID, fileID); err != nil {
					return fmt.Errorf("adding file to snapshot: %w", err)
				}
			}
			return nil
		})
		if err != nil {
			return err
		}

		// Periodic status
		if time.Since(lastStatusTime) >= statusInterval {
			elapsed := time.Since(startTime)
			rate := float64(end) / elapsed.Seconds()
			pct := float64(end) / float64(len(fileIDs)) * 100
			fmt.Printf("Associating files: %s/%s (%.1f%%), %.0f files/sec\n",
				formatNumber(end), formatNumber(len(fileIDs)), pct, rate)
			lastStatusTime = time.Now()
		}
	}

	elapsed := time.Since(startTime)
	rate := float64(len(fileIDs)) / elapsed.Seconds()
	fmt.Printf("Associated %s unchanged files in %s (%.0f files/sec)\n",
		formatNumber(len(fileIDs)), elapsed.Round(time.Second), rate)

	return nil
}

// processPhase processes the files that need backing up
func (s *Scanner) processPhase(ctx context.Context, filesToProcess []*FileToProcess, result *ScanResult) error {
	// Calculate total bytes to process
	var totalBytes int64
	for _, f := range filesToProcess {
		totalBytes += f.FileInfo.Size()
	}

	// Set up periodic status output
	lastStatusTime := time.Now()
	statusInterval := 15 * time.Second
	startTime := time.Now()
	filesProcessed := 0
	var bytesProcessed int64
	totalFiles := len(filesToProcess)

	// Process each file
	for _, fileToProcess := range filesToProcess {
		// Update progress
		if s.progress != nil {
			s.progress.GetStats().CurrentFile.Store(fileToProcess.Path)
		}

		// Process file in streaming fashion
		if err := s.processFileStreaming(ctx, fileToProcess, result); err != nil {
			// Handle files that were deleted between scan and process phases
			if errors.Is(err, os.ErrNotExist) {
				log.Warn("File was deleted during backup, skipping", "path", fileToProcess.Path)
				result.FilesSkipped++
				continue
			}
			// Skip file read errors if --skip-errors is enabled
			if s.skipErrors {
				log.Error("ERROR: Failed to process file (skipping due to --skip-errors)", "path", fileToProcess.Path, "error", err)
				fmt.Printf("ERROR: Failed to process %s: %v (skipping)\n", fileToProcess.Path, err)
				result.FilesSkipped++
				continue
			}
			return fmt.Errorf("processing file %s: %w", fileToProcess.Path, err)
		}

		// Update files processed counter
		if s.progress != nil {
			s.progress.GetStats().FilesProcessed.Add(1)
		}

		filesProcessed++
		bytesProcessed += fileToProcess.FileInfo.Size()

		// Output periodic status
		if time.Since(lastStatusTime) >= statusInterval {
			elapsed := time.Since(startTime)
			pct := float64(bytesProcessed) / float64(totalBytes) * 100
			byteRate := float64(bytesProcessed) / elapsed.Seconds()
			fileRate := float64(filesProcessed) / elapsed.Seconds()

			// Calculate ETA based on bytes (more accurate than files)
			remainingBytes := totalBytes - bytesProcessed
			var eta time.Duration
			if byteRate > 0 {
				eta = time.Duration(float64(remainingBytes)/byteRate) * time.Second
			}

			// Format: Progress [5.7k/610k] 6.7 GB/44 GB (15.4%), 106MB/sec, 500 files/sec, running for 1m30s, ETA: 5m49s
			fmt.Printf("Progress [%s/%s] %s/%s (%.1f%%), %s/sec, %.0f files/sec, running for %s",
				formatCompact(filesProcessed),
				formatCompact(totalFiles),
				humanize.Bytes(uint64(bytesProcessed)),
				humanize.Bytes(uint64(totalBytes)),
				pct,
				humanize.Bytes(uint64(byteRate)),
				fileRate,
				elapsed.Round(time.Second))
			if eta > 0 {
				fmt.Printf(", ETA: %s", eta.Round(time.Second))
			}
			fmt.Println()
			lastStatusTime = time.Now()
		}
	}

	// Final packer flush first - this commits remaining chunks to DB
	// and handleBlobReady will flush files whose chunks are now committed
	s.packerMu.Lock()
	if err := s.packer.Flush(); err != nil {
		s.packerMu.Unlock()
		return fmt.Errorf("flushing packer: %w", err)
	}
	s.packerMu.Unlock()

	// Flush any remaining pending files (e.g., files with only pre-existing chunks
	// that didn't trigger a blob finalize)
	if err := s.flushAllPending(ctx); err != nil {
		return fmt.Errorf("flushing remaining pending files: %w", err)
	}

	// If no storage configured, store any remaining blobs locally
	if s.storage == nil {
		blobs := s.packer.GetFinishedBlobs()
		for _, b := range blobs {
			// Blob metadata is already stored incrementally during packing
			// Just add the blob to the snapshot
			blobID, err := types.ParseBlobID(b.ID)
			if err != nil {
				return fmt.Errorf("parsing blob ID: %w", err)
			}
			err = s.repos.WithTx(ctx, func(ctx context.Context, tx *sql.Tx) error {
				return s.repos.Snapshots.AddBlob(ctx, tx, s.snapshotID, blobID, types.BlobHash(b.Hash))
			})
			if err != nil {
				return fmt.Errorf("storing blob metadata: %w", err)
			}
		}
		result.BlobsCreated += len(blobs)
	}

	return nil
}

// handleBlobReady is called by the packer when a blob is finalized
func (s *Scanner) handleBlobReady(blobWithReader *blob.BlobWithReader) error {
	startTime := time.Now().UTC()
	finishedBlob := blobWithReader.FinishedBlob

	// Report upload start
	if s.progress != nil {
		s.progress.ReportUploadStart(finishedBlob.Hash, finishedBlob.Compressed)
	}

	// Upload to storage first (without holding any locks)
	// Use scan context for cancellation support
	ctx := s.scanCtx
	if ctx == nil {
		ctx = context.Background()
	}

	// Track bytes uploaded for accurate speed calculation
	lastProgressTime := time.Now()
	lastProgressBytes := int64(0)

	progressCallback := func(uploaded int64) error {
		// Calculate instantaneous speed
		now := time.Now()
		elapsed := now.Sub(lastProgressTime).Seconds()
		if elapsed > 0.5 { // Update speed every 0.5 seconds
			bytesSinceLastUpdate := uploaded - lastProgressBytes
			speed := float64(bytesSinceLastUpdate) / elapsed

			if s.progress != nil {
				s.progress.ReportUploadProgress(finishedBlob.Hash, uploaded, finishedBlob.Compressed, speed)
			}

			lastProgressTime = now
			lastProgressBytes = uploaded
		}

		// Check for cancellation
		select {
		case <-ctx.Done():
			return ctx.Err()
		default:
			return nil
		}
	}

	// Create sharded path: blobs/ca/fe/cafebabe...
	blobPath := fmt.Sprintf("blobs/%s/%s/%s", finishedBlob.Hash[:2], finishedBlob.Hash[2:4], finishedBlob.Hash)
	if err := s.storage.PutWithProgress(ctx, blobPath, blobWithReader.Reader, finishedBlob.Compressed, progressCallback); err != nil {
		return fmt.Errorf("uploading blob %s to storage: %w", finishedBlob.Hash, err)
	}

	uploadDuration := time.Since(startTime)

	// Calculate upload speed
	uploadSpeedBps := float64(finishedBlob.Compressed) / uploadDuration.Seconds()

	// Print blob stored message
	fmt.Printf("Blob stored: %s (%s, %s/sec, %s)\n",
		finishedBlob.Hash[:12]+"...",
		humanize.Bytes(uint64(finishedBlob.Compressed)),
		humanize.Bytes(uint64(uploadSpeedBps)),
		uploadDuration.Round(time.Millisecond))

	// Log upload stats
	uploadSpeedBits := uploadSpeedBps * 8 // bits per second
	log.Info("Successfully uploaded blob to storage",
		"path", blobPath,
		"size", humanize.Bytes(uint64(finishedBlob.Compressed)),
		"duration", uploadDuration,
		"speed", humanize.SI(uploadSpeedBits, "bps"))

	// Report upload complete
	if s.progress != nil {
		s.progress.ReportUploadComplete(finishedBlob.Hash, finishedBlob.Compressed, uploadDuration)
	}

	// Update progress
	if s.progress != nil {
		stats := s.progress.GetStats()
		stats.BlobsUploaded.Add(1)
		stats.BytesUploaded.Add(finishedBlob.Compressed)
		stats.BlobsCreated.Add(1)
	}

	// Store metadata in database (after upload is complete)
	dbCtx := s.scanCtx
	if dbCtx == nil {
		dbCtx = context.Background()
	}

	// Parse blob ID for typed operations
	finishedBlobID, err := types.ParseBlobID(finishedBlob.ID)
	if err != nil {
		return fmt.Errorf("parsing finished blob ID: %w", err)
	}

	err = s.repos.WithTx(dbCtx, func(ctx context.Context, tx *sql.Tx) error {
		// Update blob upload timestamp
		if err := s.repos.Blobs.UpdateUploaded(ctx, tx, finishedBlob.ID); err != nil {
			return fmt.Errorf("updating blob upload timestamp: %w", err)
		}

		// Add the blob to the snapshot
		if err := s.repos.Snapshots.AddBlob(ctx, tx, s.snapshotID, finishedBlobID, types.BlobHash(finishedBlob.Hash)); err != nil {
			return fmt.Errorf("adding blob to snapshot: %w", err)
		}

		// Record upload metrics
		upload := &database.Upload{
			BlobHash:   finishedBlob.Hash,
			SnapshotID: s.snapshotID,
			UploadedAt: startTime,
			Size:       finishedBlob.Compressed,
			DurationMs: uploadDuration.Milliseconds(),
		}
		if err := s.repos.Uploads.Create(ctx, tx, upload); err != nil {
			return fmt.Errorf("recording upload metrics: %w", err)
		}

		return nil
	})

	// Cleanup temp file if needed
	if blobWithReader.TempFile != nil {
		tempName := blobWithReader.TempFile.Name()
		if err := blobWithReader.TempFile.Close(); err != nil {
			log.Fatal("Failed to close temp file", "file", tempName, "error", err)
		}
		if err := s.fs.Remove(tempName); err != nil {
			log.Fatal("Failed to remove temp file", "file", tempName, "error", err)
		}
	}

	if err != nil {
		return err
	}

	// Chunks from this blob are now committed to DB - remove from pending set
	log.Debug("handleBlobReady: removing pending chunk hashes")
	s.removePendingChunkHashes(blobWithReader.InsertedChunkHashes)
	log.Debug("handleBlobReady: removed pending chunk hashes")

	// Flush files whose chunks are now all committed
	// This maintains database consistency after each blob
	log.Debug("handleBlobReady: calling flushCompletedPendingFiles")
	if err := s.flushCompletedPendingFiles(dbCtx); err != nil {
		return fmt.Errorf("flushing completed files: %w", err)
	}
	log.Debug("handleBlobReady: flushCompletedPendingFiles returned")

	log.Debug("handleBlobReady: complete")
	return nil
}

// processFileStreaming processes a file by streaming chunks directly to the packer
func (s *Scanner) processFileStreaming(ctx context.Context, fileToProcess *FileToProcess, result *ScanResult) error {
	// Open the file
	file, err := s.fs.Open(fileToProcess.Path)
	if err != nil {
		return fmt.Errorf("opening file: %w", err)
	}
	defer func() { _ = file.Close() }()

	// We'll collect file chunks for database storage
	// but process them for packing as we go
	type chunkInfo struct {
		fileChunk database.FileChunk
		offset    int64
		size      int64
	}
	var chunks []chunkInfo
	chunkIndex := 0

	// Process chunks in streaming fashion and get full file hash
	fileHash, err := s.chunker.ChunkReaderStreaming(file, func(chunk chunker.Chunk) error {
		// Check for cancellation
		select {
		case <-ctx.Done():
			return ctx.Err()
		default:
		}

		log.Debug("Processing content-defined chunk from file",
			"file", fileToProcess.Path,
			"chunk_index", chunkIndex,
			"hash", chunk.Hash,
			"size", chunk.Size)

		// Check if chunk already exists (fast in-memory lookup)
		chunkExists := s.chunkExists(chunk.Hash)

		// Queue new chunks for batch insert when blob finalizes
		// This dramatically reduces transaction overhead
		if !chunkExists {
			s.packer.AddPendingChunk(chunk.Hash, chunk.Size)
			// Add to in-memory cache immediately for fast duplicate detection
			s.addKnownChunk(chunk.Hash)
			// Track as pending until blob finalizes and commits to DB
			s.addPendingChunkHash(chunk.Hash)
		}

		// Track file chunk association for later storage
		chunks = append(chunks, chunkInfo{
			fileChunk: database.FileChunk{
				FileID:    fileToProcess.File.ID,
				Idx:       chunkIndex,
				ChunkHash: types.ChunkHash(chunk.Hash),
			},
			offset: chunk.Offset,
			size:   chunk.Size,
		})

		// Update stats
		if chunkExists {
			result.FilesSkipped++ // Track as skipped for now
			result.BytesSkipped += chunk.Size
			if s.progress != nil {
				s.progress.GetStats().BytesSkipped.Add(chunk.Size)
			}
		} else {
			result.ChunksCreated++
			result.BytesScanned += chunk.Size
			if s.progress != nil {
				s.progress.GetStats().ChunksCreated.Add(1)
				s.progress.GetStats().BytesProcessed.Add(chunk.Size)
				s.progress.UpdateChunkingActivity()
			}
		}

		// Add chunk to packer immediately (streaming)
		// This happens outside the database transaction
		if !chunkExists {
			s.packerMu.Lock()
			err := s.packer.AddChunk(&blob.ChunkRef{
				Hash: chunk.Hash,
				Data: chunk.Data,
			})
			if err == blob.ErrBlobSizeLimitExceeded {
				// Finalize current blob and retry
				if err := s.packer.FinalizeBlob(); err != nil {
					s.packerMu.Unlock()
					return fmt.Errorf("finalizing blob: %w", err)
				}
				// Retry adding the chunk
				if err := s.packer.AddChunk(&blob.ChunkRef{
					Hash: chunk.Hash,
					Data: chunk.Data,
				}); err != nil {
					s.packerMu.Unlock()
					return fmt.Errorf("adding chunk after finalize: %w", err)
				}
			} else if err != nil {
				s.packerMu.Unlock()
				return fmt.Errorf("adding chunk to packer: %w", err)
			}
			s.packerMu.Unlock()
		}

		// Clear chunk data from memory immediately after use
		chunk.Data = nil

		chunkIndex++
		return nil
	})

	if err != nil {
		return fmt.Errorf("chunking file: %w", err)
	}

	log.Debug("Completed snapshotting file",
		"path", fileToProcess.Path,
		"file_hash", fileHash,
		"chunks", len(chunks))

	// Build file data for batch insertion
	// Update chunk associations with the file ID
	fileChunks := make([]database.FileChunk, len(chunks))
	chunkFiles := make([]database.ChunkFile, len(chunks))
	for i, ci := range chunks {
		fileChunks[i] = database.FileChunk{
			FileID:    fileToProcess.File.ID,
			Idx:       ci.fileChunk.Idx,
			ChunkHash: ci.fileChunk.ChunkHash,
		}
		chunkFiles[i] = database.ChunkFile{
			ChunkHash:  ci.fileChunk.ChunkHash,
			FileID:     fileToProcess.File.ID,
			FileOffset: ci.offset,
			Length:     ci.size,
		}
	}

	// Queue file for batch insertion
	// Files will be flushed when their chunks are committed (after blob finalize)
	s.addPendingFile(ctx, pendingFileData{
		file:       fileToProcess.File,
		fileChunks: fileChunks,
		chunkFiles: chunkFiles,
	})
	return nil
}

// GetProgress returns the progress reporter for this scanner
func (s *Scanner) GetProgress() *ProgressReporter {
	return s.progress
}

// detectDeletedFilesFromMap finds files that existed in previous snapshots but no longer exist
// Uses pre-loaded maps to avoid any filesystem or database access
func (s *Scanner) detectDeletedFilesFromMap(ctx context.Context, knownFiles map[string]*database.File, existingFiles map[string]struct{}, result *ScanResult) error {
	if len(knownFiles) == 0 {
		return nil
	}

	// Check each known file against the enumerated set (no filesystem access needed)
	for path, file := range knownFiles {
		// Check context cancellation periodically
		select {
		case <-ctx.Done():
			return ctx.Err()
		default:
		}

		// Check if the file exists in our enumerated set
		if _, exists := existingFiles[path]; !exists {
			// File has been deleted
			result.FilesDeleted++
			result.BytesDeleted += file.Size
			log.Debug("Detected deleted file", "path", path, "size", file.Size)
		}
	}

	if result.FilesDeleted > 0 {
		fmt.Printf("Found %s deleted files\n", formatNumber(result.FilesDeleted))
	}

	return nil
}

// compileExcludePatterns compiles the exclude patterns into glob matchers
func compileExcludePatterns(patterns []string) []compiledPattern {
	var compiled []compiledPattern
	for _, p := range patterns {
		if p == "" {
			continue
		}

		// Check if pattern is anchored (starts with /)
		anchored := strings.HasPrefix(p, "/")
		pattern := p
		if anchored {
			pattern = p[1:] // Remove leading /
		}

		// Remove trailing slash if present (directory indicator)
		pattern = strings.TrimSuffix(pattern, "/")

		// Compile the glob pattern
		// For patterns without path separators, we need to match them as components
		// e.g., ".git" should match ".git" anywhere in the path
		g, err := glob.Compile(pattern, '/')
		if err != nil {
			log.Warn("Invalid exclude pattern, skipping", "pattern", p, "error", err)
			continue
		}

		compiled = append(compiled, compiledPattern{
			pattern:  g,
			anchored: anchored,
			original: p,
		})
	}
	return compiled
}

// shouldExclude checks if a path should be excluded based on exclude patterns
// filePath is the full path to the file
// rootPath is the root of the backup source directory
func (s *Scanner) shouldExclude(filePath, rootPath string) bool {
	if len(s.compiledExclude) == 0 {
		return false
	}

	// Get the relative path from root
	relPath, err := filepath.Rel(rootPath, filePath)
	if err != nil {
		return false
	}

	// Never exclude the root directory itself
	if relPath == "." {
		return false
	}

	// Normalize path separators
	relPath = filepath.ToSlash(relPath)

	// Check each pattern
	for _, cp := range s.compiledExclude {
		if cp.anchored {
			// Anchored pattern: must match from the root
			// Match the relative path directly
			if cp.pattern.Match(relPath) {
				return true
			}
			// Also check if any prefix of the path matches (for directory patterns)
			parts := strings.Split(relPath, "/")
			for i := 1; i <= len(parts); i++ {
				prefix := strings.Join(parts[:i], "/")
				if cp.pattern.Match(prefix) {
					return true
				}
			}
		} else {
			// Unanchored pattern: can match anywhere in path
			// Check the full relative path
			if cp.pattern.Match(relPath) {
				return true
			}
			// Check each path component and subpath
			parts := strings.Split(relPath, "/")
			for i := range parts {
				// Match individual component (e.g., ".git" matches ".git" directory)
				if cp.pattern.Match(parts[i]) {
					return true
				}
				// Match subpath from this component onwards
				subpath := strings.Join(parts[i:], "/")
				if cp.pattern.Match(subpath) {
					return true
				}
			}
		}
	}

	return false
}

// formatNumber formats a number with comma separators
func formatNumber(n int) string {
	if n < 1000 {
		return fmt.Sprintf("%d", n)
	}
	return humanize.Comma(int64(n))
}

// formatCompact formats a number compactly with k/M suffixes (e.g., 5.7k, 1.2M)
func formatCompact(n int) string {
	if n < 1000 {
		return fmt.Sprintf("%d", n)
	}
	if n < 10000 {
		return fmt.Sprintf("%.1fk", float64(n)/1000)
	}
	if n < 1000000 {
		return fmt.Sprintf("%.0fk", float64(n)/1000)
	}
	return fmt.Sprintf("%.1fM", float64(n)/1000000)
}