Weighted concatenation fails due to negative integer powers not allowed about pyxtf HOT 3 CLOSED

I believe this worked back in the day because numpy was more willing to do automatic type coercion. Negative powers are only permitted for floating point numbers. Try to change the type of the exponent to float32, that might work. Not sure if the output array needs to be float as well.

from pyxtf.

Thanks @oysstu, I used np.multiply instead of *= and also used np.power with 2.0 instead of 2; this seems to work ok. I also cast it to dtype of out_array, as original. I think it is as you suggest, the output does not need to be float. I hope this works with other xtf input as-well, do you have some to test it out on?

if weighted:
    weight_factors = [ping.ping_chan_headers[channel].Weight for ping in pings[::-1]]
    out_array = np.multiply(out_array, np.power(2.0, -np.array(weight_factors))[:, np.newaxis]).astype(out_array.dtype)

from pyxtf.

Solved by #41

from pyxtf.