calc_checksum: handle non-smooth partial records.

calc_checksum is called in order to calculate checksums of the signal
data (d_signal or e_d_signal) in a record.  In particular, if rdrecord
is used to read only part of a record, it will call calc_checksum to
determine the checksums of that part of the record.  However, at that
time, self.n_sig is still equal to the total number of signals in the
input record (not the number of signals stored in d_signal or
e_d_signal, which might be different if a subset of channels are
selected).  Thus, if expanded is true and self.n_sig >
len(self.e_d_signal), this would crash.

For simplicity, and consistency with the expanded=False case, ignore
n_sig and simply calculate the checksums of e_d_signal.

Author: Benjamin Moody

wfdb/io/_signal.py

@@ -750,7 +750,7 @@ def calc_checksum(self, expanded=False):
 
         """
         if expanded:
-            cs = [int(np.sum(self.e_d_signal[ch]) % 65536) for ch in range(self.n_sig)]
+            cs = [int(np.sum(s) % 65536) for s in self.e_d_signal]
         else:
             cs = np.sum(self.d_signal, 0) % 65536
             cs = [int(c) for c in cs]