Improve progress reporting and video format coverage

tests/test_cli.py

@@ -1,3 +1,5 @@
+import time
+
 from typer.testing import CliRunner
 
 from tgvideo_check.cli import app
@@ -167,6 +169,52 @@ def test_scan_command_reports_discovery_and_stage_progress(tmp_path, monkeypatch
     assert "Stage2 progress: 1/1 (100%) passed_stage2=1" in result.stdout
 
 
+def test_scan_command_reports_stage2_heartbeat_while_waiting(tmp_path, monkeypatch) -> None:
+    root = tmp_path / "videos"
+    root.mkdir()
+    for index in range(60):
+        (root / f"risky-{index:03d}.mp4").write_bytes(b"risky")
+
+    def fake_run_ffprobe(path, **kwargs):
+        return ProbeExecutionResult(
+            status="completed",
+            command=("ffprobe", str(path)),
+            exit_code=0,
+            stdout_text='{"format":{"format_name":"mov,mp4,m4a,3gp,3g2,mj2","duration":"20.0"},"streams":[{"codec_type":"video"}]}',
+            stderr_text="",
+            elapsed_ms=1,
+        )
+
+    def fake_run_ffmpeg(path, **kwargs):
+        time.sleep(0.02)
+        return FfmpegExecutionResult(
+            status="completed",
+            command=("ffmpeg", str(path)),
+            exit_code=0,
+            stdout_text="frame=10",
+            stderr_text="",
+            elapsed_ms=20,
+            reason_code=None,
+        )
+
+    monkeypatch.setattr("tgvideo_check.cli.PROGRESS_HEARTBEAT_SECONDS", 0.01)
+    monkeypatch.setattr("tgvideo_check.cli.run_ffprobe", fake_run_ffprobe)
+    monkeypatch.setattr("tgvideo_check.cli.run_ffmpeg_verification", fake_run_ffmpeg)
+
+    result = runner.invoke(
+        app,
+        [
+            "scan",
+            str(root),
+            "--db",
+            str(tmp_path / "scan_state.db"),
+        ],
+    )
+
+    assert result.exit_code == 0
+    assert "Stage2 heartbeat:" in result.stdout
+
+
 def test_scan_resume_preserves_reason_code_in_exported_report(tmp_path, monkeypatch) -> None:
     root = tmp_path / "videos"
     root.mkdir()

tests/test_profiles.py

@@ -1,4 +1,4 @@
-from tgvideo_check.config import ScanConfig
+from tgvideo_check.config import DEFAULT_VIDEO_EXTENSIONS, ScanConfig
 from tgvideo_check.profiles import get_profile, resolve_config
 
 
@@ -12,6 +12,46 @@ def test_tg_hdd_profile_uses_conservative_defaults() -> None:
     assert profile.high_risk_extensions == ("mp4", "mov", "m4v", "mkv", "webm")
 
 
+def test_default_extensions_cover_broad_common_video_formats() -> None:
+    assert DEFAULT_VIDEO_EXTENSIONS == (
+        "mp4",
+        "mov",
+        "m4v",
+        "mkv",
+        "webm",
+        "avi",
+        "flv",
+        "f4v",
+        "ts",
+        "m2ts",
+        "mts",
+        "mpg",
+        "mpeg",
+        "mpe",
+        "mpv",
+        "m2v",
+        "wmv",
+        "asf",
+        "3gp",
+        "3g2",
+        "ogv",
+        "ogm",
+        "vob",
+        "qt",
+        "rm",
+        "rmvb",
+        "mxf",
+        "divx",
+    )
+    assert ScanConfig(roots=()).extensions == DEFAULT_VIDEO_EXTENSIONS
+
+
+def test_deep_verify_profile_treats_all_supported_formats_as_high_risk() -> None:
+    profile = get_profile("deep-verify")
+
+    assert profile.high_risk_extensions == DEFAULT_VIDEO_EXTENSIONS
+
+
 def test_resolve_config_applies_profile_overrides() -> None:
     config = resolve_config(
         ScanConfig(
@@ -26,4 +66,3 @@ def test_resolve_config_applies_profile_overrides() -> None:
     assert config.stage1_workers == 3
     assert config.stage2_workers == 1
     assert config.extensions == ("mp4", "mkv")
-

tests/test_scheduler_hdd.py

@@ -169,3 +169,35 @@ def test_iter_stage_batch_results_yields_items_as_they_finish() -> None:
     results = list(iter_stage_batch_results(batch, worker))
 
     assert [(item.file_id, output) for item, output in results] == [(2, 20), (1, 10)]
+
+
+def test_iter_stage_batch_results_can_emit_heartbeat_while_waiting() -> None:
+    batch = build_stage_plan(
+        stage1_files=[
+            make_stored_file(file_id=1, discovery_order=0, stage1_status="pending_stage1"),
+        ],
+        stage2_files=[],
+        stage1_workers=1,
+        stage2_workers=1,
+        queue_limit=64,
+    )[0]
+    heartbeats: list[tuple[int, ...]] = []
+
+    def worker(item: StoredFile) -> int:
+        time.sleep(0.03)
+        return item.file_id
+
+    results = list(
+        iter_stage_batch_results(
+            batch,
+            worker,
+            heartbeat_seconds=0.01,
+            on_heartbeat=lambda items: heartbeats.append(
+                tuple(entry.file_id for entry in items)
+            ),
+        )
+    )
+
+    assert results == [(batch.files[0], 1)]
+    assert heartbeats
+    assert heartbeats[0] == (1,)