dataframely 3.1.0__tar.gz → 3.1.1__tar.gz

{dataframely-3.1.0 → dataframely-3.1.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dataframely
-Version: 3.1.0
+Version: 3.1.1
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11

{dataframely-3.1.0 → dataframely-3.1.1}/dataframely/columns/categorical.py

@@ -132,7 +132,10 @@ class Categorical(Column):
         return str
 
     def _sample_unchecked(self, generator: Generator, n: int) -> pl.Series:
-        # We simply sample low-cardinality strings here
+        # Two-letter strings allow 702 categories, exceeding UInt8's capacity.
+        max_length = 1 if self._categories.physical() == pl.UInt8 else 2
         return generator.sample_string(
-            n, regex=r"[a-z]{1,2}", null_probability=self._null_probability
+            n,
+            regex=rf"[a-z]{{1,{max_length}}}",
+            null_probability=self._null_probability,
         ).cast(self.dtype)

{dataframely-3.1.0 → dataframely-3.1.1}/pyproject.toml

@@ -26,7 +26,7 @@ description = "A declarative, polars-native data frame validation library"
 name = "dataframely"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "3.1.0"
+version = "3.1.1"
 
 [project.optional-dependencies]
 pydantic = ["pydantic>=2"]

{dataframely-3.1.0 → dataframely-3.1.1}/tests/columns/test_sample.py

@@ -175,6 +175,31 @@ def test_sample_enum(generator: Generator) -> None:
     assert set(samples) == {"a", "b", "c"}
 
 
+@pytest.mark.parametrize("physical", [pl.UInt8, pl.UInt16, pl.UInt32])
+@pytest.mark.parametrize("explicit_categories", [False, True])
+@pytest.mark.parametrize("nullable", [False, True])
+@pytest.mark.parametrize("n", [0, 10_000])
+def test_sample_categorical(
+    physical: type[pl.DataType],
+    explicit_categories: bool,
+    nullable: bool,
+    n: int,
+    generator: Generator,
+) -> None:
+    categories = (
+        pl.Categories("sample", physical=physical) if explicit_categories else physical
+    )
+    schema = create_schema("test", {"a": dy.Categorical(categories, nullable=nullable)})
+    column = schema.columns()["a"]
+    samples = sample_and_validate(column, generator, n=n)
+    assert len(samples) == n
+    assert samples.dtype == column.dtype
+    assert samples.to_physical().dtype == physical
+    assert samples.drop_nulls().n_unique() <= (256 if physical == pl.UInt8 else 702)
+    if n:
+        assert samples.is_null().any() == nullable
+
+
 def test_sample_list(generator: Generator) -> None:
     column = dy.List(
         dy.String(regex="[abc]"), nullable=True, min_length=5, max_length=10