really be more tolerant

Author: FynnBe

tests/test_proc_ops.py

@@ -30,7 +30,7 @@ def test_scale_linear(tid: MemberId):
     op(sample)
 
     expected = xr.DataArray(np.array([[[1, 4, 48], [4, 10, 57]]]), dims=("x", "y", "c"))
-    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 def test_scale_linear_no_channel(tid: MemberId):
@@ -42,7 +42,7 @@ def test_scale_linear_no_channel(tid: MemberId):
     op(sample)
 
     expected = xr.DataArray(np.array([[1, 3, 5], [7, 9, 11]]), dims=("x", "y"))
-    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 T = TypeVar("T")
@@ -75,7 +75,7 @@ def test_zero_mean_unit_variance(tid: MemberId):
         ),
         dims=("x", "y"),
     )
-    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 def test_zero_mean_unit_variance_fixed(tid: MemberId):
@@ -102,7 +102,7 @@ def test_zero_mean_unit_variance_fixed(tid: MemberId):
     )
     sample = Sample(members={tid: Tensor.from_xarray(data)}, stat={}, id=None)
     op(sample)
-    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 def test_zero_mean_unit_across_axes(tid: MemberId):
@@ -123,7 +123,7 @@ def test_zero_mean_unit_across_axes(tid: MemberId):
         [(data[i : i + 1] - data[i].mean()) / data[i].std() for i in range(2)], dim="c"
     )
     op(sample)
-    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 def test_zero_mean_unit_variance_fixed2(tid: MemberId):
@@ -139,7 +139,7 @@ def test_zero_mean_unit_variance_fixed2(tid: MemberId):
     sample = Sample(members={tid: Tensor.from_xarray(data)}, stat={}, id=None)
     expected = xr.DataArray((np_data - mean) / (std + eps), dims=("x", "y"))
     op(sample)
-    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 def test_binarize(tid: MemberId):
@@ -223,7 +223,7 @@ def test_combination_of_op_steps_with_dims_specified(tid: MemberId):
     )
 
     op(sample)
-    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 @pytest.mark.parametrize(
@@ -255,7 +255,7 @@ def test_scale_mean_variance(tid: MemberId, axes: Optional[Tuple[AxisId, ...]]):
     )
     sample.stat = compute_measures(op.required_measures, [sample])
     op(sample)
-    xr.testing.assert_allclose(ref_data, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(ref_data, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 @pytest.mark.parametrize(
@@ -291,13 +291,13 @@ def test_scale_mean_variance_per_channel(tid: MemberId, axes_str: Optional[str])
     if axes is not None and AxisId("c") not in axes:
         # mean,std per channel should match exactly
         xr.testing.assert_allclose(
-            ref_data, sample.members[tid].data, rtol=1e-6, atol=1e-7
+            ref_data, sample.members[tid].data, rtol=1e-5, atol=1e-7
         )
     else:
         # mean,std across channels should not match
         with pytest.raises(AssertionError):
             xr.testing.assert_allclose(
-                ref_data, sample.members[tid].data, rtol=1e-6, atol=1e-7
+                ref_data, sample.members[tid].data, rtol=1e-5, atol=1e-7
             )
 
 
@@ -317,7 +317,7 @@ def test_scale_range(tid: MemberId):
 
     op(sample)
     # NOTE xarray.testing.assert_allclose compares irrelavant properties here and fails although the result is correct
-    np.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    np.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 def test_scale_range_axes(tid: MemberId):
@@ -352,7 +352,7 @@ def test_scale_range_axes(tid: MemberId):
 
     op(sample)
     # NOTE xarray.testing.assert_allclose compares irrelavant properties here and fails although the result is correct
-    np.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    np.testing.assert_allclose(expected, sample.members[tid].data, rtol=1e-5, atol=1e-7)
 
 
 def test_sigmoid(tid: MemberId):
@@ -367,4 +367,4 @@ def test_sigmoid(tid: MemberId):
     sigmoid(sample)
 
     exp = xr.DataArray(1.0 / (1 + np.exp(-np_data)), dims=axes)
-    xr.testing.assert_allclose(exp, sample.members[tid].data, rtol=1e-6, atol=1e-7)
+    xr.testing.assert_allclose(exp, sample.members[tid].data, rtol=1e-5, atol=1e-7)