Handle non homogeneous arrays

ultraplot/internals/inputs.py

@@ -150,13 +150,17 @@ def _to_numpy_array(data, strip_units=False):
         data = data.data  # support pint quantities that get unit-stripped later
     elif isinstance(data, (DataFrame, Series, Index)):
         data = data.values
+
     if Quantity is not ndarray and isinstance(data, Quantity):
         if strip_units:
             return np.atleast_1d(data.magnitude)
         else:
             return np.atleast_1d(data.magnitude) * data.units
-    else:
+    try:
         return np.atleast_1d(data)  # natively preserves masked arrays
+    except (TypeError, ValueError):
+        # handle  non-homogeneous data
+        return np.array(data, dtype=object)
 
 
 def _to_masked_array(data, *, copy=False):

ultraplot/tests/test_inputs.py

@@ -0,0 +1,18 @@
+import ultraplot as uplt, pytest, numpy as np
+
+
+@pytest.mark.parametrize(
+    "data, dtype",
+    [
+        ([1, 2, 3], int),
+        ([[1, 2], [1, 2, 3]], object),
+        (["hello", 1], np.dtype("<U21")),  # will convert 1 to string
+        ([["hello"], 1], object),  # non-homogeneous  # mixed types
+    ],
+)
+def test_to_numpy_array(data, dtype):
+    """
+    Test that to_numpy_array works with various data types.
+    """
+    arr = uplt.internals.inputs._to_numpy_array(data)
+    assert arr.dtype == dtype, f"Expected dtype {dtype}, got {arr.dtype}"

ultraplot/tests/test_plot.py

@@ -433,3 +433,17 @@ def test_color_parsing_for_none():
     for artist in ax[0].collections:
         assert artist.get_facecolor().shape[0] == 0
     uplt.close(fig)
+
+
+@pytest.mark.mpl_image_compare
+def test_inhomogeneous_violin(rng):
+    """
+    Test that inhomogeneous violin plots work correctly.
+    """
+    fig, ax = uplt.subplots()
+    data = [rng.normal(size=100), np.random.normal(size=200)]
+    violins = ax.violinplot(data, vert=True, labels=["A", "B"])
+    assert len(violins) == 2
+    for violin in violins:
+        assert violin.get_paths()  # Ensure paths are created
+    return fig