Keep recordings on Ctrl+C, drop dataset repair, make Stop immediate

CLAUDE.md

@@ -35,13 +35,12 @@ Run the Python tests with `pytest` (config in [pyproject.toml](pyproject.toml);
 
 [server.py](lelab/server.py) is a thin FastAPI router. Each feature lives in its own module that owns its global state and exposes `handle_*` functions plus a Pydantic request model:
 
-- [record.py](lelab/record.py) — dataset recording (wraps `lerobot.record.record`); patches `lerobot.common.utils.control_utils` keyboard listener so frontend buttons replace arrow-key controls.
+- [record.py](lelab/record.py) — dataset recording (wraps `lerobot.record.record`); patches `lerobot.common.utils.control_utils` keyboard listener so frontend buttons replace arrow-key controls. The worker is a daemon thread, so the server's shutdown hook calls `stop_recording_and_wait()` to let an active session finalize on Ctrl+C.
 - [teleoperate.py](lelab/teleoperate.py) — leader→follower teleoperation (wraps `lerobot.teleoperate`).
 - [calibrate.py](lelab/calibrate.py) — step-by-step web calibration with a `CalibrationManager` singleton and `_step_complete` threading.Event.
 - [train.py](lelab/train.py) — wraps the LeRobot training CLI as a subprocess (psutil for lifecycle, queue for log streaming).
 - [datasets.py](lelab/datasets.py) — dataset listing (local cache + Hub, merged under a `source` field) plus the episode-browsing handlers behind `/dataset-episodes`, `/dataset-episode`, `/dataset-frame`, `/dataset-thumbnails`, `/dataset-motion` and `/dataset-video`.
-- [episode_media.py](lelab/episode_media.py) — reads the LeRobot v3.0 on-disk layout (`meta/episodes/*.parquet` + `videos/`) and decodes frames with PyAV. Deliberately does **not** go through `LeRobotDataset`, so browsing a dataset never builds a robot config or imports torch. Several episodes share one mp4, so locating an episode resolves both the file and the `from_/to_timestamp` window inside it — which is why `/dataset-video` is addressed by chunk/file rather than by episode.
-- [dataset_repair.py](lelab/dataset_repair.py) — rebuilds `meta/episodes/` for a recording interrupted before `LeRobotDataset.finalize()` ran. Without an episode index LeRobot reads a local dataset as "not downloaded yet" and fetches it from the Hub, which 404s for a dataset that was never pushed. Call `repair_local_dataset(repo_id)` before opening a recorded dataset with `LeRobotDataset`.
+- [episode_media.py](lelab/episode_media.py) — reads the LeRobot v3.0 on-disk layout (`meta/episodes/*.parquet` + `videos/`) and decodes frames with PyAV. Deliberately does **not** go through `LeRobotDataset`, so browsing a dataset never builds a robot config or imports torch. Several episodes share one mp4, so locating an episode resolves both the file and the `from_/to_timestamp` window inside it — which is why `/dataset-video` is addressed by chunk/file rather than by episode. Call `ensure_episode_index(repo_id)` before opening a local dataset with `LeRobotDataset`: LeRobot writes the index only in `finalize()`, and without it reads the dataset as "not downloaded yet" and fetches it from the Hub, which 404s for a dataset that was never pushed.
 - [utils/config.py](lelab/utils/config.py) — shared paths and persistence: calibration JSON, saved ports, saved config selections. **Import shared constants from here, do not hardcode paths in feature modules.**
 
 ### State model

lelab/datasets.py

@@ -27,7 +27,6 @@
 from lerobot.datasets.dataset_tools import merge_datasets
 
 from . import episode_media
-from .dataset_repair import DatasetRepairError, repair_local_dataset
 from .utils.hf_auth import cached_whoami, shared_hf_api
 
 logger = logging.getLogger(__name__)
@@ -203,13 +202,10 @@ def handle_merge_local_datasets(source_repo_ids: list[str], output_name: str) ->
     if output_dir.exists():
         raise ValueError(f"A local dataset already exists at {output_repo_id}")
 
-    # An interrupted local recording can have data and info.json but no episode
-    # index. LeRobot then treats it as absent and attempts a Hub lookup, so make
-    # each source readable before constructing a dataset from its local path.
     try:
         for repo_id in source_repo_ids:
-            repair_local_dataset(repo_id)
-    except DatasetRepairError as exc:
+            episode_media.ensure_episode_index(repo_id)
+    except episode_media.UnreadableDatasetError as exc:
         raise ValueError(str(exc)) from exc
 
     source_dirs = [episode_media.resolve_dataset_dir(repo_id) for repo_id in source_repo_ids]