Returns a TensorStore that reads/writes from an in-memory array.

Parameters:¶

array: ArrayLike¶

Source array.

dtype: DTypeLike | None = None¶

Data type to which array will be converted.

context: Context = None¶

Optional context to use, for specifying Context.data_copy_concurrency.

copy: bool | None = None¶

Indicates whether the returned TensorStore may be a copy of the source array, rather than a reference to it.

• If None (default), the source array is copied only if it cannot be referenced.

• If True, a copy is always made.

• If False, a copy is never made and an error is raised if the source data can’t be referenced.

write: bool | None = None¶

Indicates whether the returned TensorStore is writable.

• If None (default), the returned TensorStore may or may not be writable, depending on whether the source array converts to a writable NumPy array.

• If True, the returned TensorStore is required to be writable, and an error is returned if the source array does not support writing.

• If False, the returned TensorStore is never writable, even if the source array is writable.