fix: rasterize on spread (#6933)

Author: hoxbro

holoviews/operation/datashader.py

@@ -824,7 +824,7 @@ def _process(self, element, key=None):
         if df is element.data:
             df = df.copy()
 
-        pos, neg = element.vdims[1:3] if len(element.vdims) > 2 else element.vdims[1:2] * 2
+        neg, pos = element.vdims[1:3] if len(element.vdims) > 2 else element.vdims[1:2] * 2
         yvals = df[y.name]
         df[y.name] = yvals + df[pos.name]
         df["_lower"] = yvals - df[neg.name]

holoviews/tests/operation/test_datashader.py

@@ -751,8 +751,8 @@ def test_spread_aggregate_asymmetric_count(self):
         )
         agg = rasterize(spread, width=4, height=4, dynamic=False)
         xs = [0.25, 0.75, 1.25, 1.75]
-        ys = [0.6125, 1.4375, 2.2625, 3.0875]
-        arr = np.array([[0, 0, 0, 0], [1, 0, 0, 0], [0, 1, 1, 0], [0, 0, 1, 1]])
+        ys = [1.025, 1.875, 2.725, 3.575]
+        arr = np.array([[0, 0, 0, 0], [1, 1, 1, 0], [0, 1, 1, 0], [0, 0, 0, 1]])
         expected = hv.Image((xs, ys, arr), vdims=hv.Dimension("Count", nodata=0))
         assert_element_equal(agg, expected)