Simplify draw_values_uniform and add tests

src/fmu/tools/sensitivities/design_distributions.py

@@ -234,23 +234,24 @@ def draw_values_uniform(dist_parameters, numreals, normalscoresamples=None):
     Returns:
         list of values
     """
-    distribution = None
-    status, msg = _check_dist_params_uniform(dist_parameters)
-    if status:
-        low = float(dist_parameters[0])
-        high = float(dist_parameters[1])
-        uscale = high - low
-        if normalscoresamples is not None:
-            values = scipy.stats.uniform.ppf(
-                scipy.stats.norm.cdf(normalscoresamples), loc=low, scale=uscale
-            )
+    if numreals < 0:
+        raise ValueError("numreal must be a positive integer")
 
-        else:
-            distribution = scipy.stats.uniform(loc=low, scale=uscale)
-            values = distribution.rvs(size=numreals)
-    else:
+    status, msg = _check_dist_params_uniform(dist_parameters)
+    if not status:
         raise ValueError(msg)
-    return values
+
+    low = float(dist_parameters[0])
+    high = float(dist_parameters[1])
+    uscale = high - low
+
+    if normalscoresamples is not None:
+        return scipy.stats.uniform.ppf(
+            scipy.stats.norm.cdf(normalscoresamples), loc=low, scale=uscale
+        )
+
+    distribution = scipy.stats.uniform(loc=low, scale=uscale)
+    return distribution.rvs(size=numreals)
 
 
 def draw_values_triangular(dist_parameters, numreals, normalscoresamples=None):

tests/sensitivities/test_design_distributions.py

@@ -183,15 +183,49 @@ def test_draw_values_lognormal():
     assert all(value > 0 for value in values)
 
 
+def test_draw_uniform_with_correlation():
+    rng = np.random.default_rng()
+    correlation = 0.6
+    n_samples = 1000
+
+    cov_matrix = [[1.0, correlation], [correlation, 1.0]]
+    normal_scores = rng.multivariate_normal([0, 0], cov_matrix, size=n_samples)
+
+    values1 = dists.draw_values_uniform([0, 100], n_samples, normal_scores[:, 0])
+    values2 = dists.draw_values_uniform([10, 50], n_samples, normal_scores[:, 1])
+
+    actual_correlation = np.corrcoef(values1, values2)[0, 1]
+    assert (actual_correlation - correlation) < 0.1
+
+
 def test_draw_values_uniform():
-    """Test drawing uniform values"""
-    assert not dists.draw_values_uniform([10, 100], 0).size
+    values = dists.draw_values_uniform([10, 100], 0)
+    assert len(values) == 0
 
     values = dists.draw_values_uniform([10, 100], 20)
     assert len(values) == 20
     assert all(isinstance(value, numbers.Number) for value in values)
     assert all(10 <= value <= 100 for value in values)
 
+    with pytest.raises(
+        ValueError,
+        match="Uniform distribution must have 2 parameters, but had 3 parameters.",
+    ):
+        values = dists.draw_values_uniform([10, 50, 100], 10)
+
+    with pytest.raises(
+        ValueError, match="Uniform distribution must have dist_param2 >= dist_param1"
+    ):
+        values = dists.draw_values_uniform([50, 10], 10)
+
+    with pytest.raises(
+        ValueError, match="Parameters for uniform distribution must be numbers."
+    ):
+        values = dists.draw_values_uniform(["a", 10], 10)
+
+    with pytest.raises(ValueError, match="numreal must be a positive integer"):
+        values = dists.draw_values_uniform([10, 50], -10)
+
 
 def test_draw_values_triangular():
     """Test drawing triangular values"""