Merge pull request #171 from djhoese/ci-fix-np2

Fix various numpy 2 and new xarray casting issues

.github/workflows/ci.yaml

@@ -59,7 +59,6 @@ jobs:
           git+https://github.com/dask/distributed \
           git+https://github.com/rasterio/rasterio \
           git+https://github.com/pydata/bottleneck \
-          git+https://github.com/zarr-developers/zarr \
           git+https://github.com/pydata/xarray;
           python -m pip install -e . --no-deps --no-build-isolation;
 

pyproject.toml

@@ -20,6 +20,8 @@ xfail_strict = true
 filterwarnings = [
     "error",
     "ignore::rasterio.errors.NotGeoreferencedWarning",
+    # remove when fixed by xarray
+    "ignore:__array_wrap__ must accept context and return_scalar arguments:DeprecationWarning:numpy",
     # dateutil needs a new release
     # https://github.com/dateutil/dateutil/issues/1314
     'ignore:datetime.datetime.utcfromtimestamp\(\) is deprecated and scheduled for removal:DeprecationWarning:dateutil',

trollimage/xrimage.py

@@ -673,6 +673,9 @@ def _scale_to_dtype(self, data, dtype, fill_value=None):
         """
         attrs = data.attrs.copy()
         if np.issubdtype(dtype, np.integer):
+            if np.issubdtype(data, np.bool_):
+                # convert boolean masks to floats so they can be scaled to the output integer dtype
+                data = data.astype(np.float64)
             if np.issubdtype(data, np.integer):
                 # preserve integer data type
                 data = data.clip(np.iinfo(dtype).min, np.iinfo(dtype).max)
@@ -1109,6 +1112,10 @@ def crude_stretch(self, min_stretch=None, max_stretch=None):
 
         attrs = self.data.attrs
         offset = -min_stretch * scale_factor
+        try:
+            offset = offset.astype(scale_factor.dtype)
+        except AttributeError:
+            offset = scale_factor.dtype.type(offset)
 
         self.data = np.multiply(self.data, scale_factor, dtype=scale_factor.dtype) + offset
         self.data.attrs = attrs
@@ -1225,7 +1232,7 @@ def stretch_logarithmic(self, factor=100., base="e", min_stretch=None, max_stret
         log_func = np.log if base == "e" else getattr(np, "log" + base)
         min_stretch, max_stretch = self._convert_log_minmax_stretch(min_stretch, max_stretch)
 
-        b__ = float(crange[1] - crange[0]) / log_func(factor)
+        b__ = float(crange[1] - crange[0]) / self.data.dtype.type(log_func(factor))
         c__ = float(crange[0])
 
         def _band_log(arr, min_input, max_input):