feat(*): add check_vector

src/fbench/__init__.py

@@ -3,5 +3,6 @@
 __version__ = version("fbench")
 
 from .core import *
+from .validation import *
 
-del core
+del core, validation

src/fbench/core.py

@@ -1,5 +1,7 @@
 import numpy as np
 
+from fbench.validation import check_vector
+
 __all__ = (
     "ackley",
     "rastrigin",
@@ -44,7 +46,7 @@ def ackley(x):
     >>> round(ackley([1, 1]), 4)
     3.6254
     """
-    x = np.asarray(x)
+    x = check_vector(x, 1)
     return float(
         -20 * np.exp(-0.2 * np.sqrt((x**2).mean()))
         - np.exp((np.cos(2 * np.pi * x)).sum() / len(x))
@@ -87,7 +89,7 @@ def rastrigin(x):
     >>> round(rastrigin([4.5, 4.5]), 4)
     80.5
     """
-    x = np.asarray(x)
+    x = check_vector(x, 1)
     return float(10 * len(x) + (x**2 - 10 * np.cos(2 * np.pi * x)).sum())
 
 
@@ -130,7 +132,7 @@ def rosenbrock(x):
     >>> round(rosenbrock([3, 3]), 4)
     3604.0
     """
-    x = np.asarray(x)
+    x = check_vector(x, 2)
     return float((100 * (x[1:] - x[:-1] ** 2) ** 2 + (1 - x[:-1]) ** 2).sum())
 
 
@@ -164,4 +166,5 @@ def sphere(x):
     >>> sphere([1, 1])
     2.0
     """
-    return float((np.asarray(x) ** 2).sum())
+    x = check_vector(x, 1)
+    return float((x**2).sum())

src/fbench/exception.py

@@ -0,0 +1,10 @@
+class IncorrectNumberOfElements(ValueError):
+    """Raise if number of elements is incorrect."""
+
+    pass
+
+
+class NotAVectorError(TypeError):
+    """Raise if object is not vector-like."""
+
+    pass

src/fbench/validation.py

@@ -0,0 +1,43 @@
+import numpy as np
+
+__all__ = ("check_vector",)
+
+from fbench.exception import IncorrectNumberOfElements, NotAVectorError
+
+
+def check_vector(x, min_number_of_elements):
+    """Validate an n-dimensional vector.
+
+    Parameters
+    ----------
+    x : array_like
+        Input data with :math:`n` elements that can be converted to an array.
+    min_number_of_elements : int
+        Specify the minimum number of elements ``x`` must have.
+
+    Returns
+    -------
+    np.ndarray
+        An n-dimensional vector.
+
+    Raises
+    ------
+    NotAVectorError
+        If ``x`` is not vector-like.
+    IncorrectNumberOfElements
+        If ``x`` does not satisfy the ``min_number_of_elements`` condition.
+    """
+    x = np.asarray(x)
+
+    if len(x.shape) != 1:
+        raise NotAVectorError(
+            f"input must be vector-like object - it has shape={x.shape}"
+        )
+
+    if not len(x) >= min_number_of_elements:
+        raise IncorrectNumberOfElements(
+            f"number of elements must be at least {min_number_of_elements} "
+            f"- it has {x.shape[0]}"
+        )
+
+    return x

tests/test_core.py

@@ -39,7 +39,6 @@ def test_rastrigin(arg, expected):
 @pytest.mark.parametrize(
     "arg, expected",
     [
-        ([0], 0.0),
         ([0, 0], 1.0),
         ([1, 1], 0.0),
         ([1, 1, 1], 0.0),

tests/test_validation.py

@@ -0,0 +1,17 @@
+import numpy as np
+import numpy.testing as npt
+import pytest
+
+from fbench import exception, validation
+
+
+def test_check_vector():
+    x = [1, 2, 3]
+    actual = validation.check_vector(x, 1)
+    npt.assert_array_equal(actual, np.array(x))
+
+    with pytest.raises(exception.NotAVectorError):
+        validation.check_vector([[1, 2]], 1)
+
+    with pytest.raises(exception.IncorrectNumberOfElements):
+        validation.check_vector([1, 2], 3)