Add joint dataset coverage tests

Author: francois-drielsma

test/test_io/test_dataset.py

@@ -1585,6 +1585,57 @@ def __getitem__(self, idx):
         )
 
 
+def test_joint_dataset_rejects_overlay_metadata_mismatch():
+    """JointDataset should reject matching keys with conflicting metadata."""
+
+    class DummyDataset:
+        data_types = {"index": "scalar"}
+
+        def __init__(self, overlay_methods):
+            self.overlay_methods = overlay_methods
+
+        def __len__(self):
+            return 1
+
+        def __getitem__(self, idx):
+            return {"index": idx}
+
+    with pytest.raises(ValueError, match="overlay mismatch"):
+        joint_dataset_module.JointDataset(
+            primary=DummyDataset({"index": "cat"}),
+            secondary=DummyDataset({"index": "first"}),
+        )
+
+
+def test_joint_dataset_rejects_empty_sources():
+    """JointDataset should require both primary and secondary samples."""
+
+    class DummyDataset:
+        data_types = {"index": "scalar"}
+        overlay_methods = {"index": "cat"}
+
+        def __init__(self, size):
+            self.size = size
+
+        def __len__(self):
+            return self.size
+
+        def __getitem__(self, idx):
+            return {"index": idx}
+
+    with pytest.raises(ValueError, match="primary dataset must expose"):
+        joint_dataset_module.JointDataset(
+            primary=DummyDataset(0),
+            secondary=DummyDataset(1),
+        )
+
+    with pytest.raises(ValueError, match="secondary dataset must expose"):
+        joint_dataset_module.JointDataset(
+            primary=DummyDataset(1),
+            secondary=DummyDataset(0),
+        )
+
+
 def test_joint_dataset_accepts_explicit_pair_index():
     """Pair indexes should let an external sampler own secondary pairing."""
 
@@ -1611,6 +1662,34 @@ def __getitem__(self, idx):
     assert dataset[(1, None)]["index"] == 1
 
 
+def test_joint_dataset_rejects_bad_pair_indexes():
+    """JointDataset should validate tuple shape and secondary bounds."""
+
+    class DummyDataset:
+        data_types = {"index": "scalar"}
+        overlay_methods = {"index": "cat"}
+
+        def __init__(self, indexes):
+            self.indexes = indexes
+
+        def __len__(self):
+            return len(self.indexes)
+
+        def __getitem__(self, idx):
+            return {"index": self.indexes[idx]}
+
+    dataset = joint_dataset_module.JointDataset(
+        primary=DummyDataset([0, 1]),
+        secondary=DummyDataset([10]),
+    )
+
+    with pytest.raises(ValueError, match="length 2"):
+        dataset[(0, 1, 2)]
+
+    with pytest.raises(ValueError, match="outside of bounds"):
+        dataset[(0, 4)]
+
+
 def test_joint_dataset_merges_shared_dataset_config(monkeypatch):
     """Shared dataset config should let source blocks override only paths."""
 
@@ -1675,6 +1754,48 @@ def build_dataset(source, dtype):
     ]
 
 
+def test_joint_dataset_resolve_source_config_rejects_string_base():
+    """A string base cannot be merged into mapping source overrides."""
+    with pytest.raises(ValueError, match="shared `base`"):
+        joint_dataset_module.JointDataset.resolve_source_config(
+            "hdf5", {"file_keys": "input.h5"}
+        )
+
+
+def test_joint_dataset_build_dataset_uses_factory(monkeypatch):
+    """Mapping-based source configs should instantiate through dataset_factory."""
+
+    class DummyDataset:
+        data_types = {"index": "scalar"}
+        overlay_methods = {"index": "cat"}
+
+        def __len__(self):
+            return 1
+
+        def __getitem__(self, idx):
+            return {"index": idx}
+
+    calls = []
+
+    def fake_dataset_factory(source, entry_list=None, dtype=None):
+        calls.append((source, entry_list, dtype))
+        return DummyDataset()
+
+    monkeypatch.setattr("spine.io.factories.dataset_factory", fake_dataset_factory)
+
+    dataset = joint_dataset_module.JointDataset(
+        primary={"name": "hdf5", "file_keys": "a.h5"},
+        secondary={"name": "hdf5", "file_keys": "b.h5"},
+        dtype="float32",
+    )
+
+    assert len(dataset) == 1
+    assert calls == [
+        ({"name": "hdf5", "file_keys": "a.h5"}, None, "float32"),
+        ({"name": "hdf5", "file_keys": "b.h5"}, None, "float32"),
+    ]
+
+
 def test_larcv_dataset_uses_augmenter_and_length(monkeypatch):
     """The LArCV dataset should initialize and apply the configured augmenter."""
     seen = {}

test/test_io/test_factories.py

@@ -279,6 +279,42 @@ def __init__(self, dataset, **kwargs):
     assert captured["pin_memory"] is True
 
 
+def test_loader_factory_requires_sampler_for_joint_dataset(monkeypatch):
+    """Joint datasets should fail early when no joint sampler is configured."""
+
+    class DummyDataset:
+        joint = True
+        data_types = {"value": "scalar"}
+        overlay_methods = {"value": "cat"}
+
+    class DummyDataLoader:
+        def __init__(self, dataset, **kwargs):
+            self.dataset = dataset
+            self.kwargs = kwargs
+
+    fake_torch = types.ModuleType("torch")
+    fake_utils = types.ModuleType("torch.utils")
+    fake_data = types.ModuleType("torch.utils.data")
+    fake_data.DataLoader = DummyDataLoader
+    fake_utils.data = fake_data
+    fake_torch.utils = fake_utils
+
+    monkeypatch.setattr(factories_module, "TORCH_AVAILABLE", True)
+    monkeypatch.setattr(
+        factories_module, "dataset_factory", lambda *args, **kwargs: DummyDataset()
+    )
+    monkeypatch.setitem(sys.modules, "torch", fake_torch)
+    monkeypatch.setitem(sys.modules, "torch.utils", fake_utils)
+    monkeypatch.setitem(sys.modules, "torch.utils.data", fake_data)
+
+    with pytest.raises(ValueError, match="explicit joint sampler"):
+        factories_module.loader_factory(
+            dataset={"name": "joint"},
+            dtype="float32",
+            minibatch_size=2,
+        )
+
+
 def test_loader_factory_distributed_requires_explicit_rank(monkeypatch):
     """Distributed loader setup should reject an unspecified process rank."""