[BREAKING] FIX Reject empty dataset filter criteria

pyrit/datasets/seed_datasets/seed_metadata.py

@@ -184,7 +184,7 @@ class SeedDatasetFilter:
             SeedDatasetMetadata(size={"large"}, modalities={"image"}),
         ])
 
-    Passing both flat kwargs and criteria raises ValueError.
+    Passing both flat kwargs and criteria, or an empty criteria list, raises ValueError.
 
     Special tags:
     - "all": Returns every dataset, ignores all other fields. This tag will
@@ -193,7 +193,7 @@ class SeedDatasetFilter:
       strict_match=True, loses its shortcut and is treated as a normal tag.
 
     Args:
-        criteria: Explicit list of SeedDatasetMetadata to OR-match against.
+        criteria: Non-empty list of SeedDatasetMetadata to OR-match against.
         strict_match: If True, within-axis matching uses AND (all filter values
             must be present) instead of OR (any overlap suffices).
         **kwargs: Flat metadata fields (tags, size, modalities, etc.) for simple use.
@@ -221,12 +221,12 @@ def __init__(
             ])
 
         Args:
-            criteria: Explicit list of SeedDatasetMetadata to OR-match against.
+            criteria: Non-empty list of SeedDatasetMetadata to OR-match against.
             strict_match: If True, within-axis matching uses AND instead of OR.
             **kwargs: Flat metadata fields passed to SeedDatasetMetadata.
 
         Raises:
-            ValueError: If both criteria and flat kwargs are provided.
+            ValueError: If both criteria and flat kwargs are provided, or criteria is empty.
         """
         if criteria is not None and kwargs:
             raise ValueError("Cannot pass both 'criteria' and flat metadata kwargs. Use one or the other.")
@@ -248,6 +248,11 @@ def _normalize_criterion(c: SeedDatasetMetadata) -> SeedDatasetMetadata:
 
         self.criteria = [_normalize_criterion(c) for c in self.criteria]
 
+        if not self.criteria:
+            raise ValueError(
+                "'criteria' must contain at least one criterion. Omit 'criteria' for an unconstrained filter."
+            )
+
         self.strict_match = strict_match
         self._validate()
 

tests/unit/datasets/test_seed_dataset_metadata.py

@@ -371,3 +371,18 @@ def test_tags_values(self):
     def test_harm_categories_values(self):
         f = SeedDatasetFilter(harm_categories={"violence", "cybercrime"})
         assert "violence" in f.criteria[0].harm_categories
+
+
+@pytest.mark.parametrize("strict_match", [False, True])
+def test_empty_criteria_raises(strict_match: bool) -> None:
+    with pytest.raises(ValueError, match="at least one criterion"):
+        SeedDatasetFilter(criteria=[], strict_match=strict_match)
+
+
+def test_empty_criteria_with_kwargs_preserves_conflicting_arguments_error() -> None:
+    with pytest.raises(ValueError, match="Cannot pass both"):
+        SeedDatasetFilter(criteria=[], tags={"safety"})
+
+
+def test_unconstrained_criterion_preserves_default_filter() -> None:
+    assert SeedDatasetFilter(criteria=[SeedDatasetMetadata()]).criteria == SeedDatasetFilter().criteria