add scanner metadata cache reuse and control flags

src/vlm/scanner.py

@@ -30,7 +30,9 @@ def _normalize_to_utc(timestamp: datetime) -> datetime:
 def scan_library(
     root: Path,
     config: Config,
-    progress_callback: Optional[Callable[[int, int], None]] = None
+    progress_callback: Optional[Callable[[int, int], None]] = None,
+    include_video_metadata: bool = True,
+    metadata_cache: Optional[dict[str, VideoFile]] = None
 ) -> list[VideoFile]:
     """Recursively scan library for video files.
     
@@ -40,6 +42,9 @@ def scan_library(
     Args:
         root: Root directory to scan
         config: Configuration object with video extensions and settings
+        progress_callback: Optional callback for progress updates
+        include_video_metadata: Whether to run ffprobe metadata extraction
+        metadata_cache: Optional cached inventory keyed by file path
         
     Returns:
         List of VideoFile objects representing discovered files
@@ -68,7 +73,13 @@ def scan_library(
         progress_callback(0, total_paths)
 
     for index, file_path in enumerate(discovered_paths, start=1):
-        video_file = _create_video_file(file_path, root, config.categories)
+        video_file = _create_video_file(
+            file_path,
+            root,
+            config.categories,
+            include_video_metadata=include_video_metadata,
+            metadata_cache=metadata_cache
+        )
         if video_file is None:
             if progress_callback is not None:
                 progress_callback(index, total_paths)
@@ -215,7 +226,9 @@ def _is_video_file(file_path: Path, video_extensions: list[str]) -> bool:
 def _create_video_file(
     file_path: Path,
     library_root: Path,
-    categories_config: dict[str, list[str]]
+    categories_config: dict[str, list[str]],
+    include_video_metadata: bool = True,
+    metadata_cache: Optional[dict[str, VideoFile]] = None
 ) -> Optional[VideoFile]:
     """Create VideoFile object from file path.
 
@@ -226,6 +239,8 @@ def _create_video_file(
         file_path: Path to video file
         library_root: Root of the library (for categorization)
         categories_config: Mapping of category names to directory name lists
+        include_video_metadata: Whether to run ffprobe metadata extraction
+        metadata_cache: Optional cached inventory keyed by file path
 
     Returns:
         VideoFile object or None if file cannot be accessed
@@ -238,9 +253,28 @@ def _create_video_file(
         
         # Categorize based on directory structure
         category = categorize_file(file_path, library_root, categories_config)
-        
+
-        # Extract video metadata using ffprobe (optional, non-blocking)
+        # Reuse cached metadata when file identity is unchanged.
-        video_metadata = extract_metadata(file_path)
+        video_metadata: dict = {}
+        if include_video_metadata:
+            cache_key = str(file_path)
+            cached_file = metadata_cache.get(cache_key) if metadata_cache else None
+            file_mtime_seconds = int(modified_timestamp.timestamp())
+
+            if (
+                cached_file is not None
+                and cached_file.size_bytes == size_bytes
+                and int(_normalize_to_utc(cached_file.modified_timestamp).timestamp()) == file_mtime_seconds
+            ):
+                video_metadata = {
+                    'resolution': cached_file.resolution,
+                    'codec': cached_file.codec,
+                    'duration_seconds': cached_file.duration_seconds,
+                    'bitrate_kbps': cached_file.bitrate_kbps
+                }
+            else:
+                # Extract video metadata using ffprobe (optional, non-blocking)
+                video_metadata = extract_metadata(file_path)
         
         # Create VideoFile object with optional metadata
         return VideoFile(
@@ -260,6 +294,42 @@ def _create_video_file(
         return None
 
 
+def load_inventory_csv(input_path: Path) -> list[VideoFile]:
+    """Load inventory CSV into VideoFile objects."""
+    from datetime import timezone
+
+    video_files: list[VideoFile] = []
+    with open(input_path, 'r', encoding='utf-8') as csvfile:
+        lines = [line for line in csvfile if not line.startswith('#')]
+        reader = csv.DictReader(lines)
+
+        for row in reader:
+            modified_timestamp = datetime.fromisoformat(row['modified_timestamp'])
+            if modified_timestamp.tzinfo is None:
+                modified_timestamp = modified_timestamp.replace(tzinfo=timezone.utc)
+
+            resolution = row.get('resolution') or None
+            codec = row.get('codec') or None
+            duration_seconds = float(row['duration_seconds']) if row.get('duration_seconds') else None
+            bitrate_kbps = int(row['bitrate_kbps']) if row.get('bitrate_kbps') else None
+
+            video_files.append(
+                VideoFile(
+                    path=Path(row['path']),
+                    filename=row['filename'],
+                    size_bytes=int(row['size_bytes']),
+                    modified_timestamp=modified_timestamp,
+                    category=row['category'],
+                    resolution=resolution,
+                    codec=codec,
+                    duration_seconds=duration_seconds,
+                    bitrate_kbps=bitrate_kbps
+                )
+            )
+
+    return video_files
+
+
 def categorize_file(
     file_path: Path,
     library_root: Path,

tests/test_cli_scan.py

@@ -0,0 +1,89 @@
+"""Tests for CLI scan command options."""
+
+from pathlib import Path
+from unittest.mock import patch
+
+from click.testing import CliRunner
+
+from vlm.cli import main
+
+
+def test_scan_no_metadata_passes_flag_to_scanner(tmp_path):
+    """Test `scan --no-metadata` disables ffprobe metadata extraction."""
+    library_root = tmp_path / "library"
+    library_root.mkdir(parents=True)
+
+    config_file = tmp_path / "config.yaml"
+    config_file.write_text(
+        f"""
+library_root: {library_root}
+video_extensions:
+  - .mp4
+categories:
+  movie: [movie, movies]
+  series: [series, tv, shows]
+  anime: [anime]
+""".strip()
+    )
+
+    runner = CliRunner()
+    with patch("vlm.scanner.scan_library", return_value=[]) as mock_scan, patch(
+        "vlm.scanner.save_inventory_csv"
+    ) as mock_save:
+        result = runner.invoke(
+            main,
+            ["--config", str(config_file), "scan", "--no-metadata", "--output", str(tmp_path / "inv.csv")]
+        )
+
+    assert result.exit_code == 0
+    assert mock_scan.call_count == 1
+    _, kwargs = mock_scan.call_args
+    assert kwargs["include_video_metadata"] is False
+    assert "progress_callback" in kwargs
+    assert mock_save.call_count == 1
+
+
+def test_scan_force_refresh_metadata_disables_cache(tmp_path):
+    """Test `scan --force-refresh-metadata` skips cache loading."""
+    library_root = tmp_path / "library"
+    library_root.mkdir(parents=True)
+
+    config_file = tmp_path / "config.yaml"
+    config_file.write_text(
+        f"""
+library_root: {library_root}
+video_extensions:
+  - .mp4
+categories:
+  movie: [movie, movies]
+  series: [series, tv, shows]
+  anime: [anime]
+""".strip()
+    )
+
+    output_file = tmp_path / "inventory.csv"
+    output_file.write_text(
+        "# Generated: 2026-02-09T00:00:00\n"
+        f"# Library Root: {library_root}\n"
+        "path,filename,size_bytes,modified_timestamp,category,resolution,codec,duration_seconds,bitrate_kbps\n"
+    )
+
+    runner = CliRunner()
+    with patch("vlm.scanner.load_inventory_csv") as mock_load_cache, patch(
+        "vlm.scanner.scan_library", return_value=[]
+    ) as mock_scan, patch("vlm.scanner.save_inventory_csv") as mock_save:
+        result = runner.invoke(
+            main,
+            [
+                "--config", str(config_file), "scan",
+                "--output", str(output_file),
+                "--force-refresh-metadata"
+            ]
+        )
+
+    assert result.exit_code == 0
+    mock_load_cache.assert_not_called()
+    _, kwargs = mock_scan.call_args
+    assert kwargs["include_video_metadata"] is True
+    assert kwargs["metadata_cache"] is None
+    assert mock_save.call_count == 1

tests/test_scanner.py

@@ -3,6 +3,7 @@
 import os
 import tempfile
 import time
+import csv
 from datetime import datetime, timezone
 from pathlib import Path
 from unittest.mock import patch, MagicMock
@@ -17,6 +18,8 @@ from vlm.scanner import (
     categorize_file,
     scan_library,
     extract_metadata,
+    save_inventory_csv,
+    load_inventory_csv,
 )
 
 
@@ -698,6 +701,63 @@ class TestExtractMetadata:
             assert vf.duration_seconds is None
             assert vf.bitrate_kbps is None
 
+    def test_scan_library_with_metadata_disabled(self, tmp_path):
+        """Test that metadata extraction can be disabled for faster scans."""
+        movie_dir = tmp_path / "movie"
+        movie_dir.mkdir()
+
+        video_file = movie_dir / "test.mp4"
+        video_file.touch()
+
+        config = Config(library_root=tmp_path)
+
+        with patch('subprocess.run') as mock_run:
+            result = scan_library(tmp_path, config, include_video_metadata=False)
+
+        assert len(result) == 1
+        assert result[0].filename == "test.mp4"
+        assert result[0].resolution is None
+        assert result[0].codec is None
+        assert result[0].duration_seconds is None
+        assert result[0].bitrate_kbps is None
+        mock_run.assert_not_called()
+
+    def test_scan_library_reuses_cached_metadata_when_unchanged(self, tmp_path):
+        """Test unchanged files reuse metadata from previous inventory cache."""
+        movie_dir = tmp_path / "movie"
+        movie_dir.mkdir()
+        video_file = movie_dir / "test.mp4"
+        video_file.write_text("test")
+
+        original = VideoFile(
+            path=video_file,
+            filename=video_file.name,
+            size_bytes=video_file.stat().st_size,
+            modified_timestamp=datetime.fromtimestamp(video_file.stat().st_mtime, tz=timezone.utc),
+            category="movie",
+            resolution="1920x1080",
+            codec="h264",
+            duration_seconds=120.5,
+            bitrate_kbps=5000,
+        )
+
+        cache_file = tmp_path / "cached_inventory.csv"
+        save_inventory_csv([original], cache_file, tmp_path)
+        cached_entries = load_inventory_csv(cache_file)
+        metadata_cache = {str(vf.path): vf for vf in cached_entries}
+
+        config = Config(library_root=tmp_path)
+        with patch("subprocess.run") as mock_run:
+            result = scan_library(tmp_path, config, metadata_cache=metadata_cache)
+
+        assert len(result) == 1
+        scanned = result[0]
+        assert scanned.resolution == "1920x1080"
+        assert scanned.codec == "h264"
+        assert scanned.duration_seconds == 120.5
+        assert scanned.bitrate_kbps == 5000
+        mock_run.assert_not_called()
+
 
 
 class TestInventoryReports: