feat(edit-dataset): add concatenate_videos opt-out to merge (#3663)

* feat(edit-dataset): add `concatenate_videos` opt-out to merge

When merging datasets, source mp4s are concatenated into shards capped at
`video_files_size_in_mb` (default 200 MB). This is great for dataloader
throughput but destroys per-episode (or per-source) video boundaries,
which is undesirable when you want to inspect, ship, or reuse the
individual mp4s.

Add a `concatenate_videos: bool = True` knob plumbed through
`MergeConfig` → `merge_datasets` → `aggregate_datasets` → `aggregate_videos`.
When False, each source mp4 is copied 1:1 to its own destination mp4 with
no re-muxing, so the merge preserves source video boundaries.

Usage:

    lerobot-edit-dataset \
        --new_repo_id user/merged \
        --operation.type=merge \
        --operation.repo_ids "['user/a', 'user/b']" \
        --operation.concatenate_videos=false

Defaults are unchanged; the dataloader path is unaffected because the
`episodes.parquet` `from_timestamp`/`to_timestamp` index keeps working
regardless of whether each mp4 holds one or many episodes.

* feat(edit-dataset): extend concatenate opt-out to data files

Following review, add a concatenate_data flag mirroring concatenate_videos,
threaded through MergeConfig, merge_datasets, aggregate_datasets, aggregate_data
and append_or_create_parquet_file. Metadata index files still always concatenate.

Also trim the verbose docstrings and comments since the names are
self-explanatory, and extend the existing merge test to cover data files.

tests/datasets/test_aggregate.py

@@ -289,6 +289,52 @@ def test_aggregate_datasets(tmp_path, lerobot_dataset_factory):
     assert_dataset_iteration_works(aggr_ds)
 
 
+def test_aggregate_datasets_without_concatenation(tmp_path, lerobot_dataset_factory):
+    """With concatenation disabled, each source file is kept as its own destination file."""
+    ds_0 = lerobot_dataset_factory(
+        root=tmp_path / "no_stitch_0",
+        repo_id=f"{DUMMY_REPO_ID}_no_stitch_0",
+        total_episodes=3,
+        total_frames=60,
+    )
+    ds_1 = lerobot_dataset_factory(
+        root=tmp_path / "no_stitch_1",
+        repo_id=f"{DUMMY_REPO_ID}_no_stitch_1",
+        total_episodes=4,
+        total_frames=80,
+    )
+
+    aggr_root = tmp_path / "no_stitch_aggr"
+    aggregate_datasets(
+        repo_ids=[ds_0.repo_id, ds_1.repo_id],
+        roots=[ds_0.root, ds_1.root],
+        aggr_repo_id=f"{DUMMY_REPO_ID}_no_stitch_aggr",
+        aggr_root=aggr_root,
+        concatenate_videos=False,
+        concatenate_data=False,
+    )
+
+    with (
+        patch("lerobot.datasets.dataset_metadata.get_safe_version") as mock_get_safe_version,
+        patch("lerobot.datasets.dataset_metadata.snapshot_download") as mock_snapshot_download,
+    ):
+        mock_get_safe_version.return_value = "v3.0"
+        mock_snapshot_download.return_value = str(aggr_root)
+        aggr_ds = LeRobotDataset(f"{DUMMY_REPO_ID}_no_stitch_aggr", root=aggr_root)
+
+    assert_episode_and_frame_counts(
+        aggr_ds, ds_0.num_episodes + ds_1.num_episodes, ds_0.num_frames + ds_1.num_frames
+    )
+    assert_dataset_iteration_works(aggr_ds)
+    assert_video_timestamps_within_bounds(aggr_ds)
+
+    # Two single-file sources stay as two files each, instead of being packed together.
+    assert len(list((aggr_root / "data").rglob("*.parquet"))) == 2
+    assert aggr_ds.meta.video_keys, "Test fixture should produce at least one video feature"
+    for key in aggr_ds.meta.video_keys:
+        assert len(list((aggr_root / "videos" / key).rglob("*.mp4"))) == 2
+
+
 @pytest.mark.parametrize("mutation", ["mismatched_value", "missing_key"])
 def test_aggregate_incomplete_video_encoder_info_warns_and_nuls_encoders(
     tmp_path, lerobot_dataset_factory, caplog, mutation

tests/scripts/test_edit_dataset_parsing.py

@@ -66,6 +66,20 @@ class TestOperationTypeParsing:
         with pytest.raises(ValueError, match="--new_repo_id is required for merge"):
             _validate_config(cfg)
 
+    @pytest.mark.parametrize("flag", ["concatenate_videos", "concatenate_data"])
+    def test_merge_concatenate_flag_defaults_true(self, flag):
+        cfg = parse_cfg(["--new_repo_id", "test/merged", "--operation.type", "merge"])
+        assert isinstance(cfg.operation, MergeConfig)
+        assert getattr(cfg.operation, flag) is True
+
+    @pytest.mark.parametrize("flag", ["concatenate_videos", "concatenate_data"])
+    def test_merge_concatenate_flag_can_be_disabled(self, flag):
+        cfg = parse_cfg(
+            ["--new_repo_id", "test/merged", "--operation.type", "merge", f"--operation.{flag}", "false"]
+        )
+        assert isinstance(cfg.operation, MergeConfig)
+        assert getattr(cfg.operation, flag) is False
+
     def test_non_merge_requires_repo_id(self):
         cfg = parse_cfg(["--operation.type", "delete_episodes"])
         with pytest.raises(ValueError, match="--repo_id is required for delete_episodes"):