refactor: use nplike.asarray

src/awkward/_backends.py

@@ -5,10 +5,10 @@
 import awkward_cpp
 
 import awkward as ak
-from awkward._kernels import CupyKernel, JaxKernel, NumpyKernel
+from awkward._kernels import CupyKernel, JaxKernel, NumpyKernel, TypeTracerKernel
 from awkward._nplikes import Cupy, Jax, Numpy, NumpyLike, NumpyMetadata, nplike_of
 from awkward._singleton import Singleton
-from awkward._typetracer import NoKernel, TypeTracer
+from awkward._typetracer import TypeTracer
 from awkward.typing import Callable, Final, Tuple, TypeAlias, TypeVar, Unpack
 
 np = NumpyMetadata.instance()
@@ -160,8 +160,8 @@ def index_nplike(self) -> TypeTracer:
     def __init__(self):
         self._typetracer = TypeTracer.instance()
 
-    def __getitem__(self, index: KernelKeyType) -> NoKernel:
-        return NoKernel(index)
+    def __getitem__(self, index: KernelKeyType) -> TypeTracerKernel:
+        return TypeTracerKernel(index)
 
 
 def _backend_for_nplike(nplike: ak._nplikes.NumpyLike) -> Backend:

src/awkward/_broadcasting.py

@@ -537,7 +537,7 @@ def continuation():
 
                 combos = backend.index_nplike.stack(tagslist, axis=-1)
 
-                all_combos = backend.index_nplike.array(
+                all_combos = backend.index_nplike.asarray(
                     list(itertools.product(*[range(x) for x in numtags])),
                     dtype=[(str(i), combos.dtype) for i in range(len(tagslist))],
                 )

src/awkward/_connect/jax/reducers.py

@@ -83,7 +83,7 @@ def apply(cls, array, parents, outlength):
 
         if array.dtype.kind == "m":
             return ak.contents.NumpyArray(
-                array.backend.nplike.asarray(result, array.dtype)
+                array.backend.nplike.asarray(result, dtype=array.dtype)
             )
         elif array.dtype.type in (np.complex128, np.complex64):
             return ak.contents.NumpyArray(result.view(array.dtype))
@@ -184,7 +184,9 @@ def apply(cls, array, parents, outlength):
 
         if array.dtype.type in (np.complex128, np.complex64):
             return ak.contents.NumpyArray(
-                array.backend.nplike.array(result.view(array.dtype), array.dtype),
+                array.backend.nplike.asarray(
+                    result.view(array.dtype), dtype=array.dtype
+                ),
                 backend=array.backend,
             )
         else:
@@ -230,7 +232,9 @@ def apply(cls, array, parents, outlength):
         result = jax.numpy.maximum(result, cls._max_initial(cls.initial, array.dtype))
         if array.dtype.type in (np.complex128, np.complex64):
             return ak.contents.NumpyArray(
-                array.backend.nplike.array(result.view(array.dtype), array.dtype),
+                array.backend.nplike.asarray(
+                    result.view(array.dtype), dtype=array.dtype
+                ),
                 backend=array.backend,
             )
         else:

src/awkward/_connect/numba/arrayview.py

@@ -5,6 +5,7 @@
 import numba
 import numba.core.typing
 import numba.core.typing.ctypes_utils
+import numpy
 
 import awkward as ak
 
@@ -865,7 +866,7 @@ def array_supported(dtype):
     ) or isinstance(dtype, (numba.types.NPDatetime, numba.types.NPTimedelta))
 
 
-@numba.extending.overload(ak._nplikes.numpy.array)
+@numba.extending.overload(numpy.array)
 def overload_np_array(array, dtype=None):
     if isinstance(array, ArrayViewType):
         ndim = array.type.ndim
@@ -934,7 +935,7 @@ def array_impl(array, dtype=None):
             )
 
 
-@numba.extending.type_callable(ak._nplikes.numpy.asarray)
+@numba.extending.type_callable(numpy.asarray)
 def type_asarray(context):
     def typer(arrayview):
         if (
@@ -948,7 +949,7 @@ def typer(arrayview):
     return typer
 
 
-@numba.extending.lower_builtin(ak._nplikes.numpy.asarray, ArrayViewType)
+@numba.extending.lower_builtin(numpy.asarray, ArrayViewType)
 def lower_asarray(context, builder, sig, args):
     rettype, (viewtype,) = sig.return_type, sig.args
     (viewval,) = args

src/awkward/_connect/numba/builder.py

@@ -3,13 +3,11 @@
 import numba
 import numba.core.typing
 import numba.core.typing.ctypes_utils
+import numpy
 from awkward_cpp import libawkward
 
 import awkward as ak
 
-numpy = ak._nplikes.Numpy.instance()
-
-
 dynamic_addrs = {}
 
 

src/awkward/_connect/numba/layout.py

@@ -6,9 +6,6 @@
 
 import awkward as ak
 
-np = ak._nplikes.NumpyMetadata.instance()
-numpy = ak._nplikes.Numpy.instance()
-
 
 @numba.extending.typeof_impl.register(ak.contents.Content)
 @numba.extending.typeof_impl.register(ak.index.Index)

src/awkward/_lookup.py

@@ -19,7 +19,7 @@ def arrayptr(x):
         self.nplike = layout.backend.nplike
         self.generator = generator
         self.positions = positions
-        self.arrayptrs = self.nplike.array(
+        self.arrayptrs = self.nplike.asarray(
             [arrayptr(x) for x in positions], dtype=np.intp
         )
 

src/awkward/_nplikes.py

@@ -81,13 +81,26 @@ class NumpyLike(Singleton):
 
     ############################ array creation
 
-    def array(self, *args, **kwargs):
-        # data[, dtype=[, copy=]]
-        return self._module.array(*args, **kwargs)
-
-    def asarray(self, *args, **kwargs):
-        # array[, dtype=][, order=]
-        return self._module.asarray(*args, **kwargs)
+    def asarray(
+        self,
+        obj,
+        *,
+        dtype: numpy.dtype | None = None,
+        copy: bool | None = None,
+    ):
+        if copy:
+            return self._module.array(obj, dtype=dtype, copy=True)
+        elif copy is None:
+            return self._module.asarray(obj, dtype=dtype)
+        else:
+            if getattr(obj, "dtype", dtype) != dtype:
+                raise ak._errors.wrap_error(
+                    ValueError(
+                        "asarray was called with copy=False for an array of a different dtype"
+                    )
+                )
+            else:
+                return self._module.asarray(obj, dtype=dtype)
 
     def ascontiguousarray(self, *args, **kwargs):
         # array[, dtype=]

src/awkward/_reducers.py

@@ -334,7 +334,7 @@ def apply(self, array, parents, outlength):
 
         if array.dtype.kind == "m":
             return ak.contents.NumpyArray(
-                array.backend.nplike.asarray(result, array.dtype)
+                array.backend.nplike.asarray(result, dtype=array.dtype)
             )
         elif array.dtype.type in (np.complex128, np.complex64):
             return ak.contents.NumpyArray(result.view(array.dtype))
@@ -348,7 +348,7 @@ def identity_for(self, dtype: np.dtype | None):
         if dtype in {np.timedelta64, np.datetime64}:
             return np.timedelta64(0)
         else:
-            return numpy.array(0, dtype=dtype)[()]
+            return numpy.asarray(0, dtype=dtype)[()]
 
 
 class Prod(Reducer):
@@ -435,7 +435,7 @@ def identity_for(self, dtype: np.dtype | None):
         if dtype in {np.timedelta64, np.datetime64}:
             return np.timedelta64(0)
         else:
-            return numpy.array(1, dtype=dtype)[()]
+            return numpy.asarray(1, dtype=dtype)[()]
 
 
 class Any(Reducer):
@@ -633,11 +633,13 @@ def apply(self, array, parents, outlength):
             )
         if array.dtype.type in (np.complex128, np.complex64):
             return ak.contents.NumpyArray(
-                array.backend.nplike.array(result.view(array.dtype), array.dtype)
+                array.backend.nplike.asarray(
+                    result.view(array.dtype), dtype=array.dtype
+                )
             )
         else:
             return ak.contents.NumpyArray(
-                array.backend.nplike.array(result, array.dtype)
+                array.backend.nplike.asarray(result, dtype=array.dtype)
             )
 
 
@@ -734,9 +736,11 @@ def apply(self, array, parents, outlength):
             )
         if array.dtype.type in (np.complex128, np.complex64):
             return ak.contents.NumpyArray(
-                array.backend.nplike.array(result.view(array.dtype), array.dtype)
+                array.backend.nplike.asarray(
+                    result.view(array.dtype), dtype=array.dtype
+                )
             )
         else:
             return ak.contents.NumpyArray(
-                array.backend.nplike.array(result, array.dtype)
+                array.backend.nplike.asarray(result, dtype=array.dtype)
             )

src/awkward/_typetracer.py

@@ -1,4 +1,5 @@
 # BSD 3-Clause License; see https://github.com/scikit-hep/awkward-1.0/blob/main/LICENSE
+from __future__ import annotations
 
 import numbers
 
@@ -12,32 +13,6 @@
 np = _nplikes.NumpyMetadata.instance()
 
 
-class NoError:
-    def __init__(self):
-        self.str = None
-        self.filename = None
-        self.pass_through = False
-        self.attempt = ak._util.kSliceNone
-        self.id = ak._util.kSliceNone
-
-
-class NoKernel:
-    def __init__(self, index):
-        self._name_and_types = index
-
-    def __call__(self, *args):
-        for x in args:
-            try_touch_data(x)
-        return NoError()
-
-    def __repr__(self):
-        return "<{} {}{}>".format(
-            type(self).__name__,
-            self._name_and_types[0],
-            "".join(", " + str(numpy.dtype(x)) for x in self._name_and_types[1:]),
-        )
-
-
 class UnknownLengthType:
     def __repr__(self):
         return "UnknownLength"
@@ -712,9 +687,6 @@ def to_rectilinear(self, array, *args, **kwargs):
         try_touch_shape(array)
         raise ak._errors.wrap_error(NotImplementedError)
 
-    def __getitem__(self, name_and_types):
-        return NoKernel(name_and_types)
-
     @property
     def ma(self):
         raise ak._errors.wrap_error(NotImplementedError)
@@ -747,15 +719,28 @@ def raw(self, array, nplike):
 
     ############################ array creation
 
-    def array(self, data, dtype=None, **kwargs):
-        # data[, dtype=[, copy=]]
-        try_touch_data(data)
-        return TypeTracerArray.from_array(data, dtype=dtype)
-
-    def asarray(self, array, dtype=None, **kwargs):
-        # array[, dtype=][, order=]
-        try_touch_data(array)
-        return TypeTracerArray.from_array(array, dtype=dtype)
+    def asarray(
+        self,
+        obj,
+        *,
+        dtype: numpy.dtype | None = None,
+        copy: bool | None = None,
+    ):
+        try_touch_data(obj)
+        result = TypeTracerArray.from_array(obj, dtype=dtype)
+        # If we want a copy, by the dtypes don't match
+        if (
+            not (copy is None or copy)
+            and dtype is not None
+            and getattr(obj, "dtype", dtype) != dtype
+        ):
+            raise ak._errors.wrap_error(
+                ValueError(
+                    "asarray was called with copy=False for an array of a different dtype"
+                )
+            )
+        else:
+            return result
 
     def ascontiguousarray(self, array, dtype=None, **kwargs):
         # array[, dtype=]

src/awkward/contents/indexedoptionarray.py

@@ -1416,7 +1416,7 @@ def _pad_none(self, target, axis, depth, clip):
             )
 
     def _to_arrow(self, pyarrow, mask_node, validbytes, length, options):
-        index = numpy.array(self._index, copy=True)
+        index = numpy.asarray(self._index, copy=True)
         this_validbytes = self.mask_as_bool(valid_when=True)
         index[~this_validbytes] = 0
 
@@ -1468,9 +1468,9 @@ def _to_backend_array(self, allow_missing, backend):
                 elif issubclass(content.dtype.type, np.integer):
                     data[mask0] = np.iinfo(content.dtype).max
                 elif issubclass(content.dtype.type, (np.datetime64, np.timedelta64)):
-                    data[mask0] = nplike.array([np.iinfo(np.int64).max], content.dtype)[
-                        0
-                    ]
+                    data[mask0] = nplike.asarray(
+                        [np.iinfo(np.int64).max], dtype=content.dtype
+                    )[0]
                 else:
                     raise ak._errors.wrap_error(
                         AssertionError(f"unrecognized dtype: {content.dtype}")

src/awkward/contents/listoffsetarray.py

@@ -245,7 +245,7 @@ def _getitem_range(self, where):
         offsets = self._offsets[start : stop + 1]
         if offsets.length == 0:
             offsets = Index(
-                self._backend.index_nplike.array([0], dtype=self._offsets.dtype),
+                self._backend.index_nplike.asarray([0], dtype=self._offsets.dtype),
                 nplike=self._backend.index_nplike,
             )
         return ListOffsetArray(offsets, self._content, parameters=self._parameters)
@@ -1862,8 +1862,8 @@ def _to_arrow(self, pyarrow, mask_node, validbytes, length, options):
             nonzeros = npoffsets[1:] != npoffsets[:-1]
             maskedbytes = validbytes == 0
             if numpy.any(maskedbytes & nonzeros):  # null and count > 0
-                new_starts = numpy.array(npoffsets[:-1], copy=True)
-                new_stops = numpy.array(npoffsets[1:], copy=True)
+                new_starts = numpy.asarray(npoffsets[:-1], copy=True)
+                new_stops = numpy.asarray(npoffsets[1:], copy=True)
                 new_starts[maskedbytes] = 0
                 new_stops[maskedbytes] = 0
                 next = ak.contents.ListArray(
@@ -1952,7 +1952,7 @@ def _to_arrow(self, pyarrow, mask_node, validbytes, length, options):
     def _to_backend_array(self, allow_missing, backend):
         array_param = self.parameter("__array__")
         if array_param in {"bytestring", "string"}:
-            return backend.nplike.array(self.to_list())
+            return backend.nplike.asarray(self.to_list())
 
         return self.to_RegularArray()._to_backend_array(allow_missing, backend)
 

src/awkward/contents/numpyarray.py

@@ -191,8 +191,8 @@ def to_RegularArray(self):
     def maybe_to_NumpyArray(self) -> Self:
         return self
 
-    def __array__(self, *args, **kwargs):
-        return self._backend.nplike.asarray(self._data, *args, **kwargs)
+    def __array__(self, dtype=None):
+        return self._backend.nplike.asarray(self._data, dtype=dtype)
 
     def __iter__(self):
         return iter(self._data)
@@ -685,7 +685,7 @@ def _unique(self, negaxis, starts, parents, outlength):
             )
 
             return ak.contents.NumpyArray(
-                self._backend.nplike.asarray(out[: nextlength[0]], self.dtype),
+                self._backend.nplike.asarray(out[: nextlength[0]], dtype=self.dtype),
                 parameters=None,
                 backend=self._backend,
             )
@@ -997,7 +997,7 @@ def _sort_next(self, negaxis, starts, parents, outlength, ascending, stable):
                 )
             )
             return ak.contents.NumpyArray(
-                self._backend.nplike.asarray(out, self.dtype),
+                self._backend.nplike.asarray(out, dtype=self.dtype),
                 parameters=None,
                 backend=self._backend,
             )

src/awkward/contents/regulararray.py

@@ -1141,7 +1141,7 @@ def _pad_none(self, target, axis, depth, clip):
     def _to_backend_array(self, allow_missing, backend):
         array_param = self.parameter("__array__")
         if array_param in {"bytestring", "string"}:
-            return backend.nplike.array(self.to_list())
+            return backend.nplike.asarray(self.to_list())
 
         out = self._content._to_backend_array(allow_missing, backend)
         shape = (self._length, self._size) + out.shape[1:]