xbarray 0.0.1a1__tar.gz → 0.0.1a2__tar.gz

{xbarray-0.0.1a1/xbarray.egg-info → xbarray-0.0.1a2}/PKG-INFO

@@ -1,6 +1,7 @@
 Metadata-Version: 2.4
 Name: xbarray
-Version: 0.0.1a1
+Version: 0.0.1a2
+Summary: Cross-backend Python array library based on the Array API Standard.
 Requires-Python: >=3.10
 License-File: LICENSE
 Requires-Dist: typing_extensions>=4.5

{xbarray-0.0.1a1 → xbarray-0.0.1a2}/README.md

@@ -1,17 +1,23 @@
-# xarray
-Cross-backend Python N-dimensional array library based on Array API.
+# xbarray
+Cross-backend Python array library based on the Array API Standard.
 
 This allows you to write array transformations that can run on different backends like NumPy, PyTorch, and Jax.
 
+## Installation
+
+```bash
+pip install xbarray
+```
+
 ## Usage:
 
 Abstract typing:
 
 ```python
-from xarray import ComputeBackend, BArrayType, BDeviceType, BDtypeType, BRNGType
+from xbarray import ComputeBackend, BArrayType, BDeviceType, BDtypeType, BRNGType
 from typing import Generic
 
-class ABC(Generic[BArrayType, BDeviceType, BDtypeType, BRNGType]):
+class Behavior(Generic[BArrayType, BDeviceType, BDtypeType, BRNGType]):
     def __init__(self, backend : ComputeBackend[BArrayType, BDeviceType, BDtypeType, BRNGType]) -> None:
         self.backend = backend
 
@@ -22,8 +28,8 @@ class ABC(Generic[BArrayType, BDeviceType, BDtypeType, BRNGType]):
 Concrete usage:
 
 ```python
-from xarray import pytorch as pytorch_backend
+from xbarray.pytorch import PyTorchComputeBackend
 
-abc_pytorch_instance = ABC(pytorch_backend)
-abc_pytorch_array = abc_pytorch_instance.create_array()
+behavior_pytorch_instance = Behavior(PyTorchComputeBackend)
+behavior_pytorch_array = behavior_pytorch_instance.create_array()
 ```

{xbarray-0.0.1a1 → xbarray-0.0.1a2}/pyproject.toml

@@ -1,6 +1,7 @@
 [project]
 name = "xbarray"
-version = "0.0.1a1"
+description = "Cross-backend Python array library based on the Array API Standard."
+version = "0.0.1a2"
 requires-python = ">= 3.10"
 dependencies = [
     "typing_extensions>=4.5",

xbarray-0.0.1a2/xbarray/__init__.py

@@ -0,0 +1 @@
+from .base import * # Import Abstract Typings

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/jax/__init__.py

@@ -9,16 +9,22 @@ else:
     import jax.experimental.array_api as compat_module
     from jax.experimental.array_api import *
 
+from array_api_compat.common._helpers import *
+
 # Import and bind all functions from array_api_extra before exposing them
 import array_api_extra
 from functools import partial
 for api_name in dir(array_api_extra):
     if api_name.startswith('_'):
         continue
-    globals()[api_name] = partial(
-        getattr(array_api_extra, api_name),
-        xp=compat_module
-    )
+
+    if api_name == 'at':
+        globals()[api_name] = getattr(array_api_extra, api_name)
+    else:
+        globals()[api_name] = partial(
+            getattr(array_api_extra, api_name),
+            xp=compat_module
+        )
 
 from ._typing import *
 from ._extra import *

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/jax/_extra.py

@@ -77,7 +77,7 @@ def dtype_is_boolean(
 ) -> bool:
     return dtype == np.bool_ or dtype == bool
 
-from xbarray.common.implementations import *
+from .._common.implementations import *
 if hasattr(jax.numpy, "__array_api_version__"):
     compat_module = jax.numpy
 else:

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/jax/random.py

@@ -26,9 +26,9 @@ def random_number_generator(
 
 def random_discrete_uniform(
     shape : Union[int, Tuple[int, ...]], 
+    /,
     from_num : int, 
     to_num : int, 
-    /,
     *,
     rng : RNG_TYPE, 
     dtype : Optional[DTYPE_TYPE] = None, 
@@ -70,7 +70,6 @@ def random_exponential(
         data = jax.device_put(data, device)
     return new_rng, data
 
-@classmethod
 def random_normal(
     shape: Union[int, Tuple[int, ...]],
     /,

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/numpy/__init__.py

@@ -9,10 +9,14 @@ from functools import partial
 for api_name in dir(array_api_extra):
     if api_name.startswith('_'):
         continue
-    globals()[api_name] = partial(
-        getattr(array_api_extra, api_name),
-        xp=compat_module
-    )
+
+    if api_name == 'at':
+        globals()[api_name] = getattr(array_api_extra, api_name)
+    else:
+        globals()[api_name] = partial(
+            getattr(array_api_extra, api_name),
+            xp=compat_module
+        )
 
 from ._typing import *
 from ._extra import *

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/numpy/_extra.py

@@ -68,7 +68,7 @@ def dtype_is_boolean(
 ) -> bool:
     return dtype == np.bool_ or dtype == bool
 
-from xbarray.common.implementations import get_abbreviate_array_function, get_map_fn_over_arrays_function
+from .._common.implementations import get_abbreviate_array_function, get_map_fn_over_arrays_function
 from array_api_compat import numpy as compat_module
 abbreviate_array = get_abbreviate_array_function(
     backend=compat_module,

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/numpy/random.py

@@ -21,9 +21,9 @@ def random_number_generator(
 
 def random_discrete_uniform(
     shape : Union[int, Tuple[int, ...]], 
+    /,
     from_num : int, 
     to_num : int, 
-    /,
     *,
     rng : RNG_TYPE, 
     dtype : Optional[DTYPE_TYPE] = None, 
@@ -62,7 +62,6 @@ def random_exponential(
         t = t.astype(dtype)
     return rng, t
 
-@classmethod
 def random_normal(
     shape: Union[int, Tuple[int, ...]],
     /,

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/pytorch/__init__.py

@@ -10,10 +10,14 @@ from functools import partial
 for api_name in dir(array_api_extra):
     if api_name.startswith('_'):
         continue
-    globals()[api_name] = partial(
-        getattr(array_api_extra, api_name),
-        xp=compat_module
-    )
+
+    if api_name == 'at':
+        globals()[api_name] = getattr(array_api_extra, api_name)
+    else:
+        globals()[api_name] = partial(
+            getattr(array_api_extra, api_name),
+            xp=compat_module
+        )
 
 from ._typing import *
 from ._extra import *

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/pytorch/_extra.py

@@ -95,7 +95,7 @@ def dtype_is_boolean(
 ) -> bool:
     return dtype == torch.bool
 
-from xbarray.common.implementations import get_abbreviate_array_function, get_map_fn_over_arrays_function
+from .._common.implementations import get_abbreviate_array_function, get_map_fn_over_arrays_function
 from array_api_compat import torch as compat_module
 abbreviate_array = get_abbreviate_array_function(
     compat_module, 

{xbarray-0.0.1a1/xbarray → xbarray-0.0.1a2/xbarray/_src/implementations}/pytorch/random.py

@@ -27,9 +27,9 @@ def random_number_generator(
 
 def random_discrete_uniform(
     shape : Union[int, Tuple[int, ...]], 
+    /,
     from_num : int, 
     to_num : int, 
-    /,
     *,
     rng : RNG_TYPE, 
     dtype : Optional[DTYPE_TYPE] = None, 
@@ -64,7 +64,6 @@ def random_exponential(
     t = t.exponential_(lambd, generator=rng)
     return rng, t
 
-@classmethod
 def random_normal(
     shape: Union[int, Tuple[int, ...]],
     /,

xbarray-0.0.1a2/xbarray/_src/serialization/__init__.py

@@ -0,0 +1 @@
+from .serialization_map import *

xbarray-0.0.1a2/xbarray/_src/serialization/serialization_map.py

@@ -0,0 +1,31 @@
+from typing import Type, Dict, Any, Optional, List, Callable
+from types import ModuleType
+import importlib
+
+__all__ = [
+    "implementation_module_to_name",
+    "name_to_implementation_module",
+]
+
+def implementation_module_to_name(module : ModuleType) -> str:
+    """
+    Convert a backend module to its simplified name.
+    """
+    full_name = module.__name__
+    if not full_name.startswith("xbarray.implementations."):
+        raise ValueError(f"Module {full_name} is not a valid xbarray backend module.")
+
+    submodule_name = full_name[len("xbarray.implementations"):]
+    if '.' in submodule_name:
+        raise ValueError(f"Module {full_name} is not a valid xbarray backend module.")
+    return submodule_name
+
+def name_to_implementation_module(name: str) -> Type[ModuleType]:
+    """
+    Convert a simplified backend name to its module.
+    """
+    try:
+        return importlib.import_module(f"xbarray.implementations.{name}")
+    except ImportError as e:
+        raise ImportError(f"Could not import backend module '{name}'.") from e
+    

{xbarray-0.0.1a1 → xbarray-0.0.1a2}/xbarray/base/base.py

@@ -40,9 +40,9 @@ class RNGBackend(Protocol[BArrayType, BDeviceType, BDtypeType, BRNGType]):
     def random_discrete_uniform(
         self, 
         shape : Union[int, Tuple[int, ...]], 
-        from_num : int, 
-        to_num : int, 
         /,
+        from_num : int, 
+        to_num : int,
         *,
         rng : BRNGType, 
         dtype : Optional[BDtypeType] = None, 

xbarray-0.0.1a2/xbarray/cls_impl/cls_base.py

@@ -0,0 +1,10 @@
+from types import ModuleType
+from typing import Type
+from xbarray._src.serialization import implementation_module_to_name, name_to_implementation_module
+
+class ComputeBackendImplCls(Type):
+    def __str__(self):
+        return self.simplified_name
+    
+    def __repr__(self):
+        return self.simplified_name

xbarray-0.0.1a2/xbarray/jax.py

@@ -0,0 +1,13 @@
+from xbarray.cls_impl.cls_base import ComputeBackendImplCls
+from ._src.implementations import jax as jax_impl
+
+class JaxComputeBackend(metaclass=ComputeBackendImplCls):
+    pass
+
+for name in dir(jax_impl):
+    if not name.startswith('_'):
+        setattr(JaxComputeBackend, name, getattr(jax_impl, name))
+
+__all__ = [
+    'JaxComputeBackend',
+]

xbarray-0.0.1a2/xbarray/numpy.py

@@ -0,0 +1,13 @@
+from xbarray.cls_impl.cls_base import ComputeBackendImplCls
+from ._src.implementations import numpy as numpy_impl
+
+class NumpyComputeBackend(metaclass=ComputeBackendImplCls):
+    pass
+
+for name in dir(numpy_impl):
+    if not name.startswith('_'):
+        setattr(NumpyComputeBackend, name, getattr(numpy_impl, name))
+
+__all__ = [
+    'NumpyComputeBackend',
+]

xbarray-0.0.1a2/xbarray/pytorch.py

@@ -0,0 +1,13 @@
+from xbarray.cls_impl.cls_base import ComputeBackendImplCls
+from ._src.implementations import pytorch as pytorch_impl
+
+class PytorchComputeBackend(metaclass=ComputeBackendImplCls):
+    pass
+
+for name in dir(pytorch_impl):
+    if not name.startswith('_'):
+        setattr(PytorchComputeBackend, name, getattr(pytorch_impl, name))
+
+__all__ = [
+    'PytorchComputeBackend',
+]

{xbarray-0.0.1a1 → xbarray-0.0.1a2/xbarray.egg-info}/PKG-INFO

@@ -1,6 +1,7 @@
 Metadata-Version: 2.4
 Name: xbarray
-Version: 0.0.1a1
+Version: 0.0.1a2
+Summary: Cross-backend Python array library based on the Array API Standard.
 Requires-Python: >=3.10
 License-File: LICENSE
 Requires-Dist: typing_extensions>=4.5

{xbarray-0.0.1a1 → xbarray-0.0.1a2}/xbarray.egg-info/SOURCES.txt

@@ -16,23 +16,29 @@ array_api_typing/typing_extra/__init__.py
 array_api_typing/typing_extra/_api_typing.py
 array_api_typing/typing_extra/_at.py
 xbarray/__init__.py
+xbarray/jax.py
+xbarray/numpy.py
+xbarray/pytorch.py
 xbarray.egg-info/PKG-INFO
 xbarray.egg-info/SOURCES.txt
 xbarray.egg-info/dependency_links.txt
 xbarray.egg-info/requires.txt
 xbarray.egg-info/top_level.txt
+xbarray/_src/implementations/_common/implementations.py
+xbarray/_src/implementations/jax/__init__.py
+xbarray/_src/implementations/jax/_extra.py
+xbarray/_src/implementations/jax/_typing.py
+xbarray/_src/implementations/jax/random.py
+xbarray/_src/implementations/numpy/__init__.py
+xbarray/_src/implementations/numpy/_extra.py
+xbarray/_src/implementations/numpy/_typing.py
+xbarray/_src/implementations/numpy/random.py
+xbarray/_src/implementations/pytorch/__init__.py
+xbarray/_src/implementations/pytorch/_extra.py
+xbarray/_src/implementations/pytorch/_typing.py
+xbarray/_src/implementations/pytorch/random.py
+xbarray/_src/serialization/__init__.py
+xbarray/_src/serialization/serialization_map.py
 xbarray/base/__init__.py
 xbarray/base/base.py
-xbarray/common/implementations.py
-xbarray/jax/__init__.py
-xbarray/jax/_extra.py
-xbarray/jax/_typing.py
-xbarray/jax/random.py
-xbarray/numpy/__init__.py
-xbarray/numpy/_extra.py
-xbarray/numpy/_typing.py
-xbarray/numpy/random.py
-xbarray/pytorch/__init__.py
-xbarray/pytorch/_extra.py
-xbarray/pytorch/_typing.py
-xbarray/pytorch/random.py
+xbarray/cls_impl/cls_base.py

xbarray-0.0.1a1/xbarray/__init__.py

@@ -1 +0,0 @@
-from .base import *