Fix bug where dtype.itemsize was assumed to be in bits rather than bytes.

Author: hameerabbasi

sparse/numba_backend/_compressed/compressed.py

@@ -894,7 +894,7 @@ def __binsparse__(self) -> tuple[dict, list[np.ndarray]]:
 
         data_dt = str(self.data.dtype)
         if np.issubdtype(data_dt, np.complexfloating):
-            data_dt = f"complex[float{self.data.dtype.itemsize // 2}]"
+            data_dt = f"complex[float{self.data.dtype.itemsize * 4}]"
         descriptor = {
             "binsparse": {
                 "version": "0.1",

sparse/numba_backend/_coo/core.py

@@ -1543,7 +1543,7 @@ def __binsparse__(self) -> tuple[dict, list[np.ndarray]]:
 
         data_dt = str(self.data.dtype)
         if np.issubdtype(data_dt, np.complexfloating):
-            data_dt = f"complex[float{self.data.dtype.itemsize // 2}]"
+            data_dt = f"complex[float{self.data.dtype.itemsize * 4}]"
         descriptor = {
             "binsparse": {
                 "version": "0.1",