MAINT: Prevent OverflowError in multiplication (#230)

* MAINT: Prevent OverflowError in multiplication

In newer numpy versions the multiplication of ndarrays containing uint8 results in uint8. Here this leads to an OverflowError. Therefore changing to an uint64 before the multiplication prevents this.

---------

Co-authored-by: Christian Bespin <cbespin@uni-bonn.de>

Author: magruber

basil/utils/utils.py

@@ -41,7 +41,7 @@ def bitarray_to_byte_array(bitarr):
     ba = bitarray(bitarr, endian=bitarr.endian())
     ba.reverse()  # this flip the byte order and the bit order of each byte
     bs = np.frombuffer(ba.tobytes(), dtype=np.uint8)  # byte padding happens here, bitarray.tobytes()
-    bs = (bs * 0x0202020202 & 0x010884422010) % 1023
+    bs = (bs * np.uint64(0x0202020202) & 0x010884422010) % 1023
     return array('B', bs.astype(np.uint8))