feat: add progress bar to restore operation

- Add progress bar for restore and verify operations using progressbar/v3
- Replace in-memory blob cache with disk-based LRU cache (closes #29)
- Add helper wrappers for stdin/stdout/stderr IO
- Move FetchAndDecryptBlob into restore.go, remove blob_fetch_stub.go
- Use v.Stdout/v.Stdin instead of os.Stdout for all user-facing output

Rebased onto main, squashed to resolve conflicts.

internal/vaultik/blobcache.go

@@ -0,0 +1,210 @@
+package vaultik
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+	"sync"
+)
+
+// defaultMaxBlobCacheBytes is the default maximum size of the disk blob cache (10 GB).
+const defaultMaxBlobCacheBytes = 10 << 30 // 10 GiB
+
+// blobDiskCacheEntry tracks a cached blob on disk.
+type blobDiskCacheEntry struct {
+	key  string
+	size int64
+	prev *blobDiskCacheEntry
+	next *blobDiskCacheEntry
+}
+
+// blobDiskCache is an LRU cache that stores blobs on disk instead of in memory.
+// Blobs are written to a temp directory keyed by their hash. When total size
+// exceeds maxBytes, the least-recently-used entries are evicted (deleted from disk).
+type blobDiskCache struct {
+	mu       sync.Mutex
+	dir      string
+	maxBytes int64
+	curBytes int64
+	items    map[string]*blobDiskCacheEntry
+	head     *blobDiskCacheEntry // most recent
+	tail     *blobDiskCacheEntry // least recent
+}
+
+// newBlobDiskCache creates a new disk-based blob cache with the given max size.
+func newBlobDiskCache(maxBytes int64) (*blobDiskCache, error) {
+	dir, err := os.MkdirTemp("", "vaultik-blobcache-*")
+	if err != nil {
+		return nil, fmt.Errorf("creating blob cache dir: %w", err)
+	}
+	return &blobDiskCache{
+		dir:      dir,
+		maxBytes: maxBytes,
+		items:    make(map[string]*blobDiskCacheEntry),
+	}, nil
+}
+
+func (c *blobDiskCache) path(key string) string {
+	return filepath.Join(c.dir, key)
+}
+
+func (c *blobDiskCache) unlink(e *blobDiskCacheEntry) {
+	if e.prev != nil {
+		e.prev.next = e.next
+	} else {
+		c.head = e.next
+	}
+	if e.next != nil {
+		e.next.prev = e.prev
+	} else {
+		c.tail = e.prev
+	}
+	e.prev = nil
+	e.next = nil
+}
+
+func (c *blobDiskCache) pushFront(e *blobDiskCacheEntry) {
+	e.prev = nil
+	e.next = c.head
+	if c.head != nil {
+		c.head.prev = e
+	}
+	c.head = e
+	if c.tail == nil {
+		c.tail = e
+	}
+}
+
+func (c *blobDiskCache) evictLRU() {
+	if c.tail == nil {
+		return
+	}
+	victim := c.tail
+	c.unlink(victim)
+	delete(c.items, victim.key)
+	c.curBytes -= victim.size
+	_ = os.Remove(c.path(victim.key))
+}
+
+// Put writes blob data to disk cache. Entries larger than maxBytes are silently skipped.
+func (c *blobDiskCache) Put(key string, data []byte) error {
+	entrySize := int64(len(data))
+
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if entrySize > c.maxBytes {
+		return nil
+	}
+
+	// Remove old entry if updating
+	if e, ok := c.items[key]; ok {
+		c.unlink(e)
+		c.curBytes -= e.size
+		_ = os.Remove(c.path(key))
+		delete(c.items, key)
+	}
+
+	if err := os.WriteFile(c.path(key), data, 0600); err != nil {
+		return fmt.Errorf("writing blob to cache: %w", err)
+	}
+
+	e := &blobDiskCacheEntry{key: key, size: entrySize}
+	c.pushFront(e)
+	c.items[key] = e
+	c.curBytes += entrySize
+
+	for c.curBytes > c.maxBytes && c.tail != nil {
+		c.evictLRU()
+	}
+
+	return nil
+}
+
+// Get reads a cached blob from disk. Returns data and true on hit.
+func (c *blobDiskCache) Get(key string) ([]byte, bool) {
+	c.mu.Lock()
+	e, ok := c.items[key]
+	if !ok {
+		c.mu.Unlock()
+		return nil, false
+	}
+	c.unlink(e)
+	c.pushFront(e)
+	c.mu.Unlock()
+
+	data, err := os.ReadFile(c.path(key))
+	if err != nil {
+		c.mu.Lock()
+		if e2, ok2 := c.items[key]; ok2 && e2 == e {
+			c.unlink(e)
+			delete(c.items, key)
+			c.curBytes -= e.size
+		}
+		c.mu.Unlock()
+		return nil, false
+	}
+	return data, true
+}
+
+// ReadAt reads a slice of a cached blob without loading the entire blob into memory.
+func (c *blobDiskCache) ReadAt(key string, offset, length int64) ([]byte, error) {
+	c.mu.Lock()
+	e, ok := c.items[key]
+	if !ok {
+		c.mu.Unlock()
+		return nil, fmt.Errorf("key %q not in cache", key)
+	}
+	if offset+length > e.size {
+		c.mu.Unlock()
+		return nil, fmt.Errorf("read beyond blob size: offset=%d length=%d size=%d", offset, length, e.size)
+	}
+	c.unlink(e)
+	c.pushFront(e)
+	c.mu.Unlock()
+
+	f, err := os.Open(c.path(key))
+	if err != nil {
+		return nil, err
+	}
+	defer func() { _ = f.Close() }()
+
+	buf := make([]byte, length)
+	if _, err := f.ReadAt(buf, offset); err != nil {
+		return nil, err
+	}
+	return buf, nil
+}
+
+// Has returns whether a key exists in the cache.
+func (c *blobDiskCache) Has(key string) bool {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	_, ok := c.items[key]
+	return ok
+}
+
+// Size returns current total cached bytes.
+func (c *blobDiskCache) Size() int64 {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	return c.curBytes
+}
+
+// Len returns number of cached entries.
+func (c *blobDiskCache) Len() int {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	return len(c.items)
+}
+
+// Close removes the cache directory and all cached blobs.
+func (c *blobDiskCache) Close() error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	c.items = nil
+	c.head = nil
+	c.tail = nil
+	c.curBytes = 0
+	return os.RemoveAll(c.dir)
+}

internal/vaultik/blobcache_test.go

@@ -0,0 +1,189 @@
+package vaultik
+
+import (
+	"bytes"
+	"crypto/rand"
+	"fmt"
+	"testing"
+)
+
+func TestBlobDiskCache_BasicGetPut(t *testing.T) {
+	cache, err := newBlobDiskCache(1 << 20)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer func() { _ = cache.Close() }()
+
+	data := []byte("hello world")
+	if err := cache.Put("key1", data); err != nil {
+		t.Fatal(err)
+	}
+
+	got, ok := cache.Get("key1")
+	if !ok {
+		t.Fatal("expected cache hit")
+	}
+	if !bytes.Equal(got, data) {
+		t.Fatalf("got %q, want %q", got, data)
+	}
+
+	_, ok = cache.Get("nonexistent")
+	if ok {
+		t.Fatal("expected cache miss")
+	}
+}
+
+func TestBlobDiskCache_EvictionUnderPressure(t *testing.T) {
+	maxBytes := int64(1000)
+	cache, err := newBlobDiskCache(maxBytes)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer func() { _ = cache.Close() }()
+
+	for i := 0; i < 5; i++ {
+		data := make([]byte, 300)
+		if err := cache.Put(fmt.Sprintf("key%d", i), data); err != nil {
+			t.Fatal(err)
+		}
+	}
+
+	if cache.Size() > maxBytes {
+		t.Fatalf("cache size %d exceeds max %d", cache.Size(), maxBytes)
+	}
+
+	if !cache.Has("key4") {
+		t.Fatal("expected key4 to be cached")
+	}
+	if cache.Has("key0") {
+		t.Fatal("expected key0 to be evicted")
+	}
+}
+
+func TestBlobDiskCache_OversizedEntryRejected(t *testing.T) {
+	cache, err := newBlobDiskCache(100)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer func() { _ = cache.Close() }()
+
+	data := make([]byte, 200)
+	if err := cache.Put("big", data); err != nil {
+		t.Fatal(err)
+	}
+
+	if cache.Has("big") {
+		t.Fatal("oversized entry should not be cached")
+	}
+}
+
+func TestBlobDiskCache_UpdateInPlace(t *testing.T) {
+	cache, err := newBlobDiskCache(1 << 20)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer func() { _ = cache.Close() }()
+
+	if err := cache.Put("key1", []byte("v1")); err != nil {
+		t.Fatal(err)
+	}
+	if err := cache.Put("key1", []byte("version2")); err != nil {
+		t.Fatal(err)
+	}
+
+	got, ok := cache.Get("key1")
+	if !ok {
+		t.Fatal("expected hit")
+	}
+	if string(got) != "version2" {
+		t.Fatalf("got %q, want %q", got, "version2")
+	}
+	if cache.Len() != 1 {
+		t.Fatalf("expected 1 entry, got %d", cache.Len())
+	}
+	if cache.Size() != int64(len("version2")) {
+		t.Fatalf("expected size %d, got %d", len("version2"), cache.Size())
+	}
+}
+
+func TestBlobDiskCache_ReadAt(t *testing.T) {
+	cache, err := newBlobDiskCache(1 << 20)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer func() { _ = cache.Close() }()
+
+	data := make([]byte, 1024)
+	if _, err := rand.Read(data); err != nil {
+		t.Fatal(err)
+	}
+	if err := cache.Put("blob1", data); err != nil {
+		t.Fatal(err)
+	}
+
+	chunk, err := cache.ReadAt("blob1", 100, 200)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !bytes.Equal(chunk, data[100:300]) {
+		t.Fatal("ReadAt returned wrong data")
+	}
+
+	_, err = cache.ReadAt("blob1", 900, 200)
+	if err == nil {
+		t.Fatal("expected error for out-of-bounds read")
+	}
+
+	_, err = cache.ReadAt("missing", 0, 10)
+	if err == nil {
+		t.Fatal("expected error for missing key")
+	}
+}
+
+func TestBlobDiskCache_Close(t *testing.T) {
+	cache, err := newBlobDiskCache(1 << 20)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	if err := cache.Put("key1", []byte("data")); err != nil {
+		t.Fatal(err)
+	}
+	if err := cache.Close(); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestBlobDiskCache_LRUOrder(t *testing.T) {
+	cache, err := newBlobDiskCache(200)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer func() { _ = cache.Close() }()
+
+	d := make([]byte, 100)
+	if err := cache.Put("a", d); err != nil {
+		t.Fatal(err)
+	}
+	if err := cache.Put("b", d); err != nil {
+		t.Fatal(err)
+	}
+
+	// Access "a" to make it most recently used
+	cache.Get("a")
+
+	// Adding "c" should evict "b" (LRU), not "a"
+	if err := cache.Put("c", d); err != nil {
+		t.Fatal(err)
+	}
+
+	if !cache.Has("a") {
+		t.Fatal("expected 'a' to survive")
+	}
+	if !cache.Has("c") {
+		t.Fatal("expected 'c' to be present")
+	}
+	if cache.Has("b") {
+		t.Fatal("expected 'b' to be evicted")
+	}
+}

internal/vaultik/restore.go

@@ -109,28 +109,83 @@ func (v *Vaultik) Restore(opts *RestoreOptions) error {
 
 	// Step 5: Restore files
 	result := &RestoreResult{}
-	blobCache := make(map[string][]byte) // Cache downloaded and decrypted blobs
+	blobCache, err := newBlobDiskCache(defaultMaxBlobCacheBytes)
+	if err != nil {
+		return fmt.Errorf("creating blob cache: %w", err)
+	}
+	defer func() { _ = blobCache.Close() }()
 
-	for i, file := range files {
+	// Calculate total bytes for progress bar
+	var totalBytes int64
+	for _, file := range files {
+		totalBytes += file.Size
+	}
+
+	_, _ = fmt.Fprintf(v.Stdout, "Restoring %d files (%s)...\n",
+		len(files),
+		humanize.Bytes(uint64(totalBytes)),
+	)
+
+	// Create progress bar if stderr is a terminal
+	isTTY := isTerminal(v.Stderr)
+	var bar *progressbar.ProgressBar
+	if isTTY {
+		bar = progressbar.NewOptions64(
+			totalBytes,
+			progressbar.OptionSetDescription("Restoring"),
+			progressbar.OptionSetWriter(v.Stderr),
+			progressbar.OptionShowBytes(true),
+			progressbar.OptionShowCount(),
+			progressbar.OptionSetWidth(40),
+			progressbar.OptionThrottle(100*time.Millisecond),
+			progressbar.OptionOnCompletion(func() {
+				v.printlnStderr()
+			}),
+			progressbar.OptionSetRenderBlankState(true),
+		)
+	}
+
+	filesProcessed := 0
+	for _, file := range files {
 		if v.ctx.Err() != nil {
 			return v.ctx.Err()
 		}
 
 		if err := v.restoreFile(v.ctx, repos, file, opts.TargetDir, identity, chunkToBlobMap, blobCache, result); err != nil {
 			log.Error("Failed to restore file", "path", file.Path, "error", err)
-			// Continue with other files
+			filesProcessed++
+			// Update progress bar even on failure
+			if bar != nil {
+				_ = bar.Add64(file.Size)
+			}
+			// Periodic structured log for non-terminal contexts (headless/CI)
+			if !isTTY && filesProcessed%100 == 0 {
+				log.Info("Restore progress",
+					"files", fmt.Sprintf("%d/%d", filesProcessed, len(files)),
+					"bytes_restored", humanize.Bytes(uint64(result.BytesRestored)),
+				)
+			}
 			continue
 		}
 
-		// Progress logging
+		filesProcessed++
-		if (i+1)%100 == 0 || i+1 == len(files) {
+		// Update progress bar
+		if bar != nil {
+			_ = bar.Add64(file.Size)
+		}
+		// Periodic structured log for non-terminal contexts (headless/CI)
+		if !isTTY && (filesProcessed%100 == 0 || filesProcessed == len(files)) {
 			log.Info("Restore progress",
-				"files", fmt.Sprintf("%d/%d", i+1, len(files)),
+				"files", fmt.Sprintf("%d/%d", filesProcessed, len(files)),
-				"bytes", humanize.Bytes(uint64(result.BytesRestored)),
+				"bytes_restored", humanize.Bytes(uint64(result.BytesRestored)),
 			)
 		}
 	}
 
+	if bar != nil {
+		_ = bar.Finish()
+	}
+
 	result.Duration = time.Since(startTime)
 
 	log.Info("Restore complete",
@@ -299,7 +354,7 @@ func (v *Vaultik) restoreFile(
 	targetDir string,
 	identity age.Identity,
 	chunkToBlobMap map[string]*database.BlobChunk,
-	blobCache map[string][]byte,
+	blobCache *blobDiskCache,
 	result *RestoreResult,
 ) error {
 	// Calculate target path - use full original path under target directory
@@ -383,7 +438,7 @@ func (v *Vaultik) restoreRegularFile(
 	targetPath string,
 	identity age.Identity,
 	chunkToBlobMap map[string]*database.BlobChunk,
-	blobCache map[string][]byte,
+	blobCache *blobDiskCache,
 	result *RestoreResult,
 ) error {
 	// Get file chunks in order
@@ -417,13 +472,15 @@ func (v *Vaultik) restoreRegularFile(
 
 		// Download and decrypt blob if not cached
 		blobHashStr := blob.Hash.String()
-		blobData, ok := blobCache[blobHashStr]
+		blobData, ok := blobCache.Get(blobHashStr)
 		if !ok {
 			blobData, err = v.downloadBlob(ctx, blobHashStr, blob.CompressedSize, identity)
 			if err != nil {
 				return fmt.Errorf("downloading blob %s: %w", blobHashStr[:16], err)
 			}
-			blobCache[blobHashStr] = blobData
+			if putErr := blobCache.Put(blobHashStr, blobData); putErr != nil {
+				log.Debug("Failed to cache blob on disk", "hash", blobHashStr[:16], "error", putErr)
+			}
 			result.BlobsDownloaded++
 			result.BytesDownloaded += blob.CompressedSize
 		}
@@ -473,6 +530,53 @@ func (v *Vaultik) restoreRegularFile(
 	return nil
 }
 
+// BlobFetchResult holds the result of fetching and decrypting a blob.
+type BlobFetchResult struct {
+	Data           []byte
+	CompressedSize int64
+}
+
+// FetchAndDecryptBlob downloads a blob from storage, decrypts and decompresses it.
+func (v *Vaultik) FetchAndDecryptBlob(ctx context.Context, blobHash string, expectedSize int64, identity age.Identity) (*BlobFetchResult, error) {
+	// Construct blob path with sharding
+	blobPath := fmt.Sprintf("blobs/%s/%s/%s", blobHash[:2], blobHash[2:4], blobHash)
+
+	reader, err := v.Storage.Get(ctx, blobPath)
+	if err != nil {
+		return nil, fmt.Errorf("downloading blob: %w", err)
+	}
+	defer func() { _ = reader.Close() }()
+
+	// Read encrypted data
+	encryptedData, err := io.ReadAll(reader)
+	if err != nil {
+		return nil, fmt.Errorf("reading blob data: %w", err)
+	}
+
+	// Decrypt and decompress
+	blobReader, err := blobgen.NewReader(bytes.NewReader(encryptedData), identity)
+	if err != nil {
+		return nil, fmt.Errorf("creating decryption reader: %w", err)
+	}
+	defer func() { _ = blobReader.Close() }()
+
+	data, err := io.ReadAll(blobReader)
+	if err != nil {
+		return nil, fmt.Errorf("decrypting blob: %w", err)
+	}
+
+	log.Debug("Downloaded and decrypted blob",
+		"hash", blobHash[:16],
+		"encrypted_size", humanize.Bytes(uint64(len(encryptedData))),
+		"decrypted_size", humanize.Bytes(uint64(len(data))),
+	)
+
+	return &BlobFetchResult{
+		Data:           data,
+		CompressedSize: int64(len(encryptedData)),
+	}, nil
+}
+
 // downloadBlob downloads and decrypts a blob
 func (v *Vaultik) downloadBlob(ctx context.Context, blobHash string, expectedSize int64, identity age.Identity) ([]byte, error) {
 	result, err := v.FetchAndDecryptBlob(ctx, blobHash, expectedSize, identity)
@@ -518,7 +622,7 @@ func (v *Vaultik) verifyRestoredFiles(
 
 	// Create progress bar if output is a terminal
 	var bar *progressbar.ProgressBar
-	if isTerminal() {
+	if isTerminal(v.Stderr) {
 		bar = progressbar.NewOptions64(
 			totalBytes,
 			progressbar.OptionSetDescription("Verifying"),
@@ -626,7 +730,11 @@ func (v *Vaultik) verifyFile(
 	return bytesVerified, nil
 }
 
-// isTerminal returns true if stdout is a terminal
+// isTerminal returns true if the given writer is connected to a terminal.
-func isTerminal() bool {
+// Returns false if the writer does not expose a file descriptor (e.g. in tests).
-	return term.IsTerminal(int(os.Stdout.Fd()))
+func isTerminal(w io.Writer) bool {
+	if f, ok := w.(*os.File); ok {
+		return term.IsTerminal(int(f.Fd()))
+	}
+	return false
 }

internal/vaultik/snapshot.go

@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
+	"regexp"
 	"sort"
 	"strings"
 	"text/tabwriter"
@@ -1126,12 +1127,20 @@ func (v *Vaultik) PruneDatabase() (*PruneResult, error) {
 	return result, nil
 }
 
-// getTableCount returns the count of rows in a table
+// validTableNameRe matches table names containing only lowercase alphanumeric characters and underscores.
+var validTableNameRe = regexp.MustCompile(`^[a-z0-9_]+$`)
+
+// getTableCount returns the count of rows in a table.
+// The tableName is sanitized to only allow [a-z0-9_] characters to prevent SQL injection.
 func (v *Vaultik) getTableCount(tableName string) (int64, error) {
 	if v.DB == nil {
 		return 0, nil
 	}
 
+	if !validTableNameRe.MatchString(tableName) {
+		return 0, fmt.Errorf("invalid table name: %q", tableName)
+	}
+
 	var count int64
 	query := fmt.Sprintf("SELECT COUNT(*) FROM %s", tableName)
 	err := v.DB.Conn().QueryRowContext(v.ctx, query).Scan(&count)

internal/vaultik/vaultik.go

@@ -139,15 +139,33 @@ func (v *Vaultik) printlnStdout(args ...any) {
 	_, _ = fmt.Fprintln(v.Stdout, args...)
 }
 
+// FetchBlob downloads a blob from storage and returns a reader for the encrypted data.
+func (v *Vaultik) FetchBlob(ctx context.Context, blobHash string, expectedSize int64) (io.ReadCloser, int64, error) {
+	blobPath := fmt.Sprintf("blobs/%s/%s/%s", blobHash[:2], blobHash[2:4], blobHash)
+
+	reader, err := v.Storage.Get(ctx, blobPath)
+	if err != nil {
+		return nil, 0, fmt.Errorf("downloading blob: %w", err)
+	}
+
+	return reader, expectedSize, nil
+}
+
 // printfStderr writes formatted output to stderr.
 func (v *Vaultik) printfStderr(format string, args ...any) {
 	_, _ = fmt.Fprintf(v.Stderr, format, args...)
 }
 
+// printlnStderr writes a line to stderr.
+func (v *Vaultik) printlnStderr(args ...any) {
+	_, _ = fmt.Fprintln(v.Stderr, args...)
+}
+
 // scanStdin reads a line of input from stdin.
 func (v *Vaultik) scanStdin(a ...any) (int, error) {
 	return fmt.Fscanln(v.Stdin, a...)
 }
+
 // TestVaultik wraps a Vaultik with captured stdout/stderr for testing
 type TestVaultik struct {
 	*Vaultik