moveaxis extension

ivy/array/extensions.py

@@ -1,6 +1,6 @@
 # global
 import abc
-from typing import Optional, Union, Tuple
+from typing import Optional, Union, Tuple, Sequence
 
 # local
 import ivy
@@ -230,3 +230,46 @@ def max_pool2d(
             data_format=data_format,
             out=out,
         )
+
+    def moveaxis(
+        self: ivy.Array,
+        source: Union[int, Sequence[int]],
+        destination: Union[int, Sequence[int]],
+        /,
+        *,
+        out: Optional[ivy.Array] = None,
+    ) -> ivy.Array:
+        """ivy.Array instance method variant of ivy.moveaxis. This method simply
+        wraps the function, and so the docstring for ivy.unstack also applies to
+        this method with minimal changes.
+
+        Parameters
+        ----------
+        a
+            The array whose axes should be reordered.
+        source
+            Original positions of the axes to move. These must be unique.
+        destination
+            Destination positions for each of the original axes.
+            These must also be unique.
+        out
+            optional output array, for writing the result to.
+
+        Returns
+        -------
+        ret
+            Array with moved axes. This array is a view of the input array.
+
+        Examples
+        --------
+        >>> x = ivy.zeros((3, 4, 5))
+        >>> x.moveaxis(0, -1).shape
+        (4, 5, 3)
+        >>> x.moveaxis(-1, 0).shape
+        (5, 3, 4)
+        """
+        return ivy.moveaxis(
+            self._data, 
+            source,
+            destination,
+            out=out)

ivy/container/extensions.py

@@ -1,5 +1,5 @@
 # global
-from typing import Optional, Union, List, Dict, Tuple
+from typing import Optional, Union, List, Dict, Tuple, Sequence
 
 # local
 import ivy
@@ -493,3 +493,109 @@ def max_pool2d(
             map_sequences=map_sequences,
             out=out,
         )
+
+    @staticmethod
+    def static_moveaxis(
+        a: Union[ivy.Array, ivy.NativeArray, ivy.Container],
+        source: Union[int, Sequence[int]],
+        destination: Union[int, Sequence[int]],
+        /,
+        *,
+        key_chains: Optional[Union[List[str], Dict[str, str]]] = None,
+        to_apply: bool = True,
+        prune_unapplied: bool = False,
+        map_sequences: bool = False,
+        out: Optional[ivy.Container] = None,
+    ) -> ivy.Container:
+        """
+        ivy.Container static method variant of ivy.moveaxis. This method simply wraps 
+        the function, and so the docstring for ivy.moveaxis also applies to this method
+        with minimal changes.
+
+        Parameters
+        ----------
+        a
+            The container with the arrays whose axes should be reordered.
+        source
+            Original positions of the axes to move. These must be unique.
+        destination
+            Destination positions for each of the original axes.
+            These must also be unique.
+        out
+            optional output container, for writing the result to.
+
+        Returns
+        -------
+        ret
+            Container including arrays with moved axes.
+
+        Examples
+        --------
+        With one :class:`ivy.Container` input:
+
+        >>> x = ivy.Container(a=ivy.zeros((3, 4, 5)), b=ivy.zeros((2,7,6)))
+        >>> ivy.Container.static_moveaxis(x, 0, -1).shape
+
+        {
+            a: (4, 5, 3)
+            b: (7, 6, 2)
+        }
+        """
+        return ContainerBase.multi_map_in_static_method(
+            "moveaxis",
+            a,
+            source,
+            destination,
+            key_chains=key_chains,
+            to_apply=to_apply,
+            prune_unapplied=prune_unapplied,
+            map_sequences=map_sequences,
+            out=out,
+        )
+
+    def moveaxis(
+        self: ivy.Container,
+        source: Union[int, Sequence[int]],
+        destination: Union[int, Sequence[int]],
+        /,
+        *,
+        out: Optional[ivy.Container] = None,
+    ) -> ivy.Container:
+        """ivy.Container instance method variant of ivy.moveaxis. This method simply
+        wraps the function, and so the docstring for ivy.flatten also applies to this
+        method with minimal changes.
+
+        Parameters
+        ----------
+        self
+            The container with the arrays whose axes should be reordered.
+        source
+            Original positions of the axes to move. These must be unique.
+        destination
+            Destination positions for each of the original axes.
+            These must also be unique.
+        out
+            optional output container, for writing the result to.
+
+        Returns
+        -------
+        ret
+            Container including arrays with moved axes.
+
+        Examples
+        --------
+        With one :class:`ivy.Container` input:
+
+        >>> x = ivy.Container(a=ivy.zeros((3, 4, 5)), b=ivy.zeros((2,7,6)))
+        >>> x.moveaxis(0, -1).shape
+
+        {
+            a: (4, 5, 3)
+            b: (7, 6, 2)
+        }
+        """
+        return self.static_moveaxis(
+            self,
+            source,
+            destination,
+            out=out)

ivy/functional/backends/jax/extensions.py

@@ -1,5 +1,5 @@
 import logging
-from typing import Optional, Union, Tuple
+from typing import Optional, Union, Tuple, Sequence
 import ivy
 from ivy.functional.ivy.extensions import (
     _verify_coo_components,
@@ -156,3 +156,14 @@ def max_pool2d(
         return jnp.transpose(res, (0, 3, 1, 2))
 
     return res
+
+
+def moveaxis(
+    a: JaxArray,
+    source: Union[int, Sequence[int]],
+    destination: Union[int, Sequence[int]],
+    /,
+    *,
+    out: Optional[JaxArray] = None,
+) -> JaxArray:
+    return jnp.moveaxis(a, source, destination)

ivy/functional/backends/numpy/extensions.py

@@ -1,4 +1,4 @@
-from typing import Optional, Union, Tuple
+from typing import Optional, Union, Tuple, Sequence
 import logging
 import ivy
 import numpy as np
@@ -176,3 +176,17 @@ def max_pool2d(
     if data_format == "NCHW":
         return np.transpose(res, (0, 3, 1, 2))
     return res
+
+
+def moveaxis(
+    a: np.ndarray,
+    source: Union[int, Sequence[int]],
+    destination: Union[int, Sequence[int]],
+    /,
+    *,
+    out: Optional[np.ndarray] = None,
+) -> np.ndarray:
+    return np.moveaxis(a, source, destination)
+
+
+moveaxis.support_native_out = False

ivy/functional/backends/tensorflow/extensions.py

@@ -1,4 +1,4 @@
-from typing import Union, Optional, Tuple
+from typing import Union, Optional, Tuple, Sequence
 import ivy
 from ivy.functional.ivy.extensions import (
     _verify_coo_components,
@@ -138,3 +138,14 @@ def max_pool2d(
     if data_format == "NCHW":
         return tf.transpose(res, (0, 3, 1, 2))
     return res
+
+
+def moveaxis(
+    a: Union[tf.Tensor, tf.Variable],
+    source: Union[int, Sequence[int]],
+    destination: Union[int, Sequence[int]],
+    /,
+    *,
+    out: Optional[Union[tf.Tensor, tf.Variable]] = None,
+) -> Union[tf.Tensor, tf.Variable]:
+    return tf.experimental.numpy.moveaxis(a, source, destination)

ivy/functional/backends/torch/extensions.py

@@ -1,4 +1,4 @@
-from typing import Optional, Union, Tuple
+from typing import Optional, Union, Tuple, Sequence
 import ivy
 from ivy.functional.ivy.extensions import (
     _verify_coo_components,
@@ -183,3 +183,17 @@ def max_pool2d(
 
 
 max_pool2d.unsupported_dtypes = ("bfloat16", "float16")
+
+
+def moveaxis(
+    a: torch.Tensor,
+    source: Union[int, Sequence[int]],
+    destination: Union[int, Sequence[int]],
+    /,
+    *,
+    out: Optional[torch.Tensor] = None,
+) -> torch.Tensor:
+    return torch.moveaxis(a, source, destination)
+
+
+moveaxis.support_native_out = False

ivy/functional/ivy/extensions.py

@@ -1,4 +1,4 @@
-from typing import Optional, Union, Tuple
+from typing import Optional, Union, Tuple, Sequence
 import ivy
 from ivy.func_wrapper import (
     handle_out_argument,
@@ -783,3 +783,48 @@ def max_pool2d(
         """
 
     return ivy.current_backend(x).max_pool2d(x, kernel, strides, padding, out=out)
+
+
+@to_native_arrays_and_back
+@handle_out_argument
+@handle_nestable
+def moveaxis(
+    a: Union[ivy.Array, ivy.NativeArray],
+    source: Union[int, Sequence[int]],
+    destination: Union[int, Sequence[int]],
+    /,
+    *,
+    out: Optional[Union[ivy.Array, ivy.NativeArray]] = None,
+) -> Union[ivy.Array, ivy.NativeArray]:
+    """Move axes of an array to new positions..
+
+    Parameters
+    ----------
+    a
+        The array whose axes should be reordered.
+    source
+        Original positions of the axes to move. These must be unique.
+    destination
+        Destination positions for each of the original axes.
+        These must also be unique.
+    out
+        optional output array, for writing the result to.
+
+    Returns
+    -------
+    ret
+        Array with moved axes. This array is a view of the input array.
+
+    Examples
+    --------
+    With :class:`ivy.Array` input:
+
+    >>> x = ivy.zeros((3, 4, 5))
+    >>> ivy.moveaxis(x, 0, -1).shape
+    (4, 5, 3)
+    >>> ivy.moveaxis(x, -1, 0).shape
+    (5, 3, 4)
+    """
+    return ivy.current_backend().moveaxis(
+        a, source, destination, out=out
+    )

ivy_tests/test_ivy/test_functional/test_extensions.py

@@ -350,3 +350,81 @@ def test_max_pool2d(
         strides=stride,
         padding=pad
     )
+
+
+# moveaxis    
+@handle_cmd_line_args
+@given(
+    dtype_and_a=helpers.dtype_and_values(
+        available_dtypes=helpers.get_dtypes("float"),
+        min_value=-100,
+        max_value=100,
+        shape=st.shared(
+            helpers.get_shape(
+                min_num_dims=1,
+                max_num_dims=3,
+                min_dim_size=1,
+                max_dim_size=3,        
+            ),
+            key="a_s_d"
+        ),
+    ),
+    source=helpers.get_axis(
+        allow_none=False,
+        unique=True,
+        shape=st.shared(
+            helpers.get_shape(
+                min_num_dims=1,
+                max_num_dims=3,
+                min_dim_size=1,
+                max_dim_size=3,        
+            ),
+            key="a_s_d"
+        ),
+        min_size=1,
+        force_int=True,
+    ),
+    destination=helpers.get_axis(
+        allow_none=False,
+        unique=True,
+        shape=st.shared(
+            helpers.get_shape(
+                min_num_dims=1,
+                max_num_dims=3,
+                min_dim_size=1,
+                max_dim_size=3,        
+            ),
+            key="a_s_d"
+        ),
+        min_size=1,
+        force_int=True,
+    ),
+    num_positional_args=helpers.num_positional_args(fn_name="moveaxis"),
+)
+def test_moveaxis(
+    dtype_and_a,
+    source,
+    destination,
+    as_variable,
+    with_out,
+    num_positional_args,
+    native_array,
+    container,
+    instance_method,
+    fw,
+):
+    input_dtype, a = dtype_and_a
+    helpers.test_function(
+        input_dtypes=input_dtype,
+        as_variable_flags=as_variable,
+        with_out=with_out,
+        num_positional_args=num_positional_args,
+        native_array_flags=native_array,
+        container_flags=container,
+        instance_method=instance_method,
+        fw=fw,
+        fn_name="moveaxis",
+        a=np.asarray(a[0], dtype=input_dtype[0]),
+        source=source,
+        destination=destination,
+    )