fix miscellaneous numpy=2.0 errors (#8117)

* replace `np.unicode_` with `np.str_`

* replace `np.NaN` with `np.nan`

* replace more instances of `np.unicode_`

note that with more modern versions of `numpy` the `.astype(np.str_)`
don't actually change the dtype, so maybe we can remove those.

* more instances of renamed / removed dtypes

* more dtype replacements

* use `str.encode(encoding)` instead of `bytes(str, encoding)`

* explicitly import `RankWarning`

* left-over `np.RankWarning`

* use `float` instead of the removed `np.float_`

* ignore missing stubs for `numpy.exceptions`

---------

Co-authored-by: Kai Mühlbauer <kmuehlbauer@wradlib.org>
Co-authored-by: Mathias Hauser <mathause@users.noreply.github.com>
Co-authored-by: Maximilian Roos <5635139+max-sixty@users.noreply.github.com>
Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>

pyproject.toml

@@ -64,6 +64,7 @@ module = [
   "sparse.*",
   "toolz.*",
   "zarr.*",
+  "numpy.exceptions.*",  # remove once support for `numpy<2.0` has been dropped
 ]
 
 [[tool.mypy.overrides]]

xarray/backends/netCDF4_.py

@@ -207,7 +207,7 @@ def _ensure_fill_value_valid(data, attributes):
     # work around for netCDF4/scipy issue where _FillValue has the wrong type:
     # https://github.com/Unidata/netcdf4-python/issues/271
     if data.dtype.kind == "S" and "_FillValue" in attributes:
-        attributes["_FillValue"] = np.string_(attributes["_FillValue"])
+        attributes["_FillValue"] = np.bytes_(attributes["_FillValue"])
 
 
 def _force_native_endianness(var):

xarray/coding/strings.py

@@ -100,7 +100,7 @@ def ensure_fixed_length_bytes(var):
     dims, data, attrs, encoding = unpack_for_encoding(var)
     if check_vlen_dtype(data.dtype) == bytes:
         # TODO: figure out how to handle this with dask
-        data = np.asarray(data, dtype=np.string_)
+        data = np.asarray(data, dtype=np.bytes_)
     return Variable(dims, data, attrs, encoding)
 
 
@@ -151,7 +151,7 @@ def bytes_to_char(arr):
 def _numpy_bytes_to_char(arr):
     """Like netCDF4.stringtochar, but faster and more flexible."""
     # ensure the array is contiguous
-    arr = np.array(arr, copy=False, order="C", dtype=np.string_)
+    arr = np.array(arr, copy=False, order="C", dtype=np.bytes_)
     return arr.reshape(arr.shape + (1,)).view("S1")
 
 
@@ -168,7 +168,7 @@ def char_to_bytes(arr):
 
     if not size:
         # can't make an S0 dtype
-        return np.zeros(arr.shape[:-1], dtype=np.string_)
+        return np.zeros(arr.shape[:-1], dtype=np.bytes_)
 
     if is_chunked_array(arr):
         chunkmanager = get_chunked_array_type(arr)

xarray/coding/times.py

@@ -467,7 +467,7 @@ def convert_times(times, date_type, raise_on_invalid: bool = True) -> np.ndarray
     Useful to convert between calendars in numpy and cftime or between cftime calendars.
 
     If raise_on_valid is True (default), invalid dates trigger a ValueError.
-    Otherwise, the invalid element is replaced by np.NaN for cftime types and np.NaT for np.datetime64.
+    Otherwise, the invalid element is replaced by np.nan for cftime types and np.NaT for np.datetime64.
     """
     if date_type in (pd.Timestamp, np.datetime64) and not is_np_datetime_like(
         times.dtype
@@ -489,7 +489,7 @@ def convert_times(times, date_type, raise_on_invalid: bool = True) -> np.ndarray
                     f"{date_type(2000, 1, 1).calendar} calendar.  Reason: {e}."
                 )
             else:
-                dt = np.NaN
+                dt = np.nan
 
         new[i] = dt
     return new

xarray/core/accessor_str.py

@@ -471,7 +471,7 @@ def cat(self, *others, sep: str | bytes | Any = "") -> T_DataArray:
         ... )
         >>> values_2 = np.array(3.4)
         >>> values_3 = ""
-        >>> values_4 = np.array("test", dtype=np.unicode_)
+        >>> values_4 = np.array("test", dtype=np.str_)
 
         Determine the separator to use
 

xarray/core/dataarray.py

@@ -5772,8 +5772,8 @@ def idxmin(
         >>> array = xr.DataArray(
         ...     [
         ...         [2.0, 1.0, 2.0, 0.0, -2.0],
-        ...         [-4.0, np.NaN, 2.0, np.NaN, -2.0],
-        ...         [np.NaN, np.NaN, 1.0, np.NaN, np.NaN],
+        ...         [-4.0, np.nan, 2.0, np.nan, -2.0],
+        ...         [np.nan, np.nan, 1.0, np.nan, np.nan],
         ...     ],
         ...     dims=["y", "x"],
         ...     coords={"y": [-1, 0, 1], "x": np.arange(5.0) ** 2},
@@ -5868,8 +5868,8 @@ def idxmax(
         >>> array = xr.DataArray(
         ...     [
         ...         [2.0, 1.0, 2.0, 0.0, -2.0],
-        ...         [-4.0, np.NaN, 2.0, np.NaN, -2.0],
-        ...         [np.NaN, np.NaN, 1.0, np.NaN, np.NaN],
+        ...         [-4.0, np.nan, 2.0, np.nan, -2.0],
+        ...         [np.nan, np.nan, 1.0, np.nan, np.nan],
         ...     ],
         ...     dims=["y", "x"],
         ...     coords={"y": [-1, 0, 1], "x": np.arange(5.0) ** 2},

xarray/core/dataset.py

@@ -24,6 +24,13 @@
 from typing import IO, TYPE_CHECKING, Any, Callable, Generic, Literal, cast, overload
 
 import numpy as np
+
+# remove once numpy 2.0 is the oldest supported version
+try:
+    from numpy.exceptions import RankWarning
+except ImportError:
+    from numpy import RankWarning
+
 import pandas as pd
 
 from xarray.coding.calendar_ops import convert_calendar, interp_calendar
@@ -8785,9 +8792,9 @@ def polyfit(
 
             with warnings.catch_warnings():
                 if full:  # Copy np.polyfit behavior
-                    warnings.simplefilter("ignore", np.RankWarning)
+                    warnings.simplefilter("ignore", RankWarning)
                 else:  # Raise only once per variable
-                    warnings.simplefilter("once", np.RankWarning)
+                    warnings.simplefilter("once", RankWarning)
 
                 coeffs, residuals = duck_array_ops.least_squares(
                     lhs, rhs.data, rcond=rcond, skipna=skipna_da
@@ -9077,8 +9084,8 @@ def idxmin(
         >>> array2 = xr.DataArray(
         ...     [
         ...         [2.0, 1.0, 2.0, 0.0, -2.0],
-        ...         [-4.0, np.NaN, 2.0, np.NaN, -2.0],
-        ...         [np.NaN, np.NaN, 1.0, np.NaN, np.NaN],
+        ...         [-4.0, np.nan, 2.0, np.nan, -2.0],
+        ...         [np.nan, np.nan, 1.0, np.nan, np.nan],
         ...     ],
         ...     dims=["y", "x"],
         ...     coords={"y": [-1, 0, 1], "x": ["a", "b", "c", "d", "e"]},
@@ -9174,8 +9181,8 @@ def idxmax(
         >>> array2 = xr.DataArray(
         ...     [
         ...         [2.0, 1.0, 2.0, 0.0, -2.0],
-        ...         [-4.0, np.NaN, 2.0, np.NaN, -2.0],
-        ...         [np.NaN, np.NaN, 1.0, np.NaN, np.NaN],
+        ...         [-4.0, np.nan, 2.0, np.nan, -2.0],
+        ...         [np.nan, np.nan, 1.0, np.nan, np.nan],
         ...     ],
         ...     dims=["y", "x"],
         ...     coords={"y": [-1, 0, 1], "x": ["a", "b", "c", "d", "e"]},

xarray/core/dtypes.py

@@ -40,7 +40,7 @@ def __eq__(self, other):
 PROMOTE_TO_OBJECT: tuple[tuple[type[np.generic], type[np.generic]], ...] = (
     (np.number, np.character),  # numpy promotes to character
     (np.bool_, np.character),  # numpy promotes to character
-    (np.bytes_, np.unicode_),  # numpy promotes to unicode
+    (np.bytes_, np.str_),  # numpy promotes to unicode
 )
 
 

xarray/core/missing.py

@@ -730,7 +730,7 @@ def interp_func(var, x, new_x, method: InterpOptions, kwargs):
         # scipy.interpolate.interp1d always forces to float.
         # Use the same check for blockwise as well:
         if not issubclass(var.dtype.type, np.inexact):
-            dtype = np.float_
+            dtype = float
         else:
             dtype = var.dtype
 

xarray/core/nputils.py

@@ -6,6 +6,12 @@
 import pandas as pd
 from numpy.core.multiarray import normalize_axis_index  # type: ignore[attr-defined]
 
+# remove once numpy 2.0 is the oldest supported version
+try:
+    from numpy.exceptions import RankWarning
+except ImportError:
+    from numpy import RankWarning
+
 from xarray.core.options import OPTIONS
 from xarray.core.pycompat import is_duck_array
 
@@ -194,7 +200,7 @@ def _nanpolyfit_1d(arr, x, rcond=None):
 
 def warn_on_deficient_rank(rank, order):
     if rank != order:
-        warnings.warn("Polyfit may be poorly conditioned", np.RankWarning, stacklevel=2)
+        warnings.warn("Polyfit may be poorly conditioned", RankWarning, stacklevel=2)
 
 
 def least_squares(lhs, rhs, rcond=None, skipna=False):