tests: add some property based tests (#540)

* tests: add some property based tests

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: henryiii

setup.py

@@ -49,12 +49,7 @@
 
 
 extras = {
-    "test": [
-        "pytest",
-        "pytest-benchmark",
-        "typing_extensions; python_version<'3.8'",
-        "cloudpickle",
-    ],
+    "test": ["pytest", "pytest-benchmark", "cloudpickle", "hypothesis >=6.0"],
     "docs": [
         "Sphinx~=3.0",
         "recommonmark>=0.5.0",

tests/test_accumulators.py

@@ -1,7 +1,11 @@
+import pytest
 from pytest import approx
 
 import boost_histogram as bh
 
+hypothesis = pytest.importorskip("hypothesis")
+st = pytest.importorskip("hypothesis.strategies")
+
 
 def test_weighted_sum():
     a = bh.accumulators.WeightedSum(1.5, 2.5)
@@ -89,15 +93,25 @@ def test_mean():
     assert repr(a) == "Mean(count=3, value=2, variance=1)"
 
 
-def test_sum_mean():
+float_st = st.floats(
+    allow_nan=False, allow_infinity=False, min_value=-1e30, max_value=1e30
+)
+simple_list_st = st.lists(float_st, min_size=1, max_size=10)
+
+
+@hypothesis.given(
+    simple_list_st,
+    simple_list_st,
+)
+def test_sum_mean(list1, list2):
     a = bh.accumulators.Mean()
-    a.fill([1, 2, 3])
+    a.fill(list1)
 
     b = bh.accumulators.Mean()
-    b.fill([5, 6])
+    b.fill(list2)
 
     c = bh.accumulators.Mean()
-    c.fill([1, 2, 3, 5, 6])
+    c.fill(list1 + list2)
 
     ab = a + b
     assert ab.value == approx(c.value)
@@ -110,15 +124,30 @@ def test_sum_mean():
     assert a.count == approx(c.count)
 
 
-def test_sum_weighed_mean():
+dual_lists_st = st.integers(min_value=1, max_value=10).flatmap(
+    lambda n: st.tuples(
+        st.lists(float_st, min_size=n, max_size=n),
+        st.lists(
+            st.floats(
+                allow_nan=False, allow_infinity=False, min_value=1e-30, max_value=1e30
+            ),
+            min_size=n,
+            max_size=n,
+        ),
+    )
+)
+
+
+@hypothesis.given(dual_lists_st, dual_lists_st)
+def test_sum_weighed_mean(pair1, pair2):
     a = bh.accumulators.WeightedMean()
-    a.fill([1, 2, 3], weight=[2, 5, 3])
+    a.fill(pair1[0], weight=pair1[1])
 
     b = bh.accumulators.WeightedMean()
-    b.fill([5, 6], weight=[12, 17])
+    b.fill(pair2[0], weight=pair2[1])
 
     c = bh.accumulators.WeightedMean()
-    c.fill([1, 2, 3, 5, 6], weight=[2, 5, 3, 12, 17])
+    c.fill(pair1[0] + pair2[0], weight=pair1[1] + pair2[1])
 
     ab = a + b
     assert ab.value == approx(c.value)

tests/test_pbt.py

@@ -0,0 +1,32 @@
+import numpy as np
+import pytest
+from pytest import approx
+
+import boost_histogram as bh
+
+hypothesis = pytest.importorskip("hypothesis")
+nst = pytest.importorskip("hypothesis.extra.numpy")
+
+
+@hypothesis.given(
+    nst.arrays(
+        float,
+        (4,),
+        elements=dict(min_value=1, max_value=100, exclude_min=True, allow_nan=False),
+    ),
+    nst.arrays(
+        float, (4,), elements=dict(min_value=-100, max_value=100, allow_nan=False)
+    ),
+    nst.arrays(
+        float,
+        (4,),
+        elements=dict(min_value=0, max_value=100, allow_nan=False, exclude_min=True),
+    ),
+)
+def test_variance_setting(cnt, val, var):
+    h = bh.Histogram(bh.axis.Regular(4, 0, 1), storage=bh.storage.Mean())
+    h[...] = np.stack([cnt, val, var * cnt], axis=-1)
+
+    assert h.counts() == approx(cnt)
+    assert h.values() == approx(val)
+    assert h.variances() == approx(var)  # , abs=1e-3, rel=1e-3)