ArrayContainer

class cubie.batchsolving.arrays.BaseArrayManager.ArrayContainer[source]

Bases: ABC

Store per-array metadata and references for CUDA managers.

array_names() → List[str][source]: Return array labels managed by this container.

attach(label: str, array: ndarray[Any, dtype[_ScalarType_co]]) → None[source]: Attach an array to this container.

delete(label: str) → None[source]: Delete reference to an array.

delete_all() → None[source]: Delete all array references.

get_array(label: str) → ndarray[Any, dtype[_ScalarType_co]] | FakeCUDAArray | None[source]: Return the stored array for label.

get_managed_array(label: str) → ManagedArray[source]: Retrieve the metadata wrapper for label.

iter_managed_arrays() → Iterator[tuple[str, ManagedArray]][source]: Yield (label, managed) pairs for each array.

property memory_type: str: Return the memory type of the first managed array.

set_array(label: str, array: ndarray[Any, dtype[_ScalarType_co]] | FakeCUDAArray | None) → None[source]: Attach an array reference to label.

set_memory_type(memory_type: str) → None[source]: Apply memory_type to all managed arrays.