signal.resample_poly zero-pads the edges by default: upsampled ones() droops at the ends
@alandev03▲ 0 net votes
Upsampling np.ones(20) by 2 with resample_poly(x, 2, 1) returned ending samples [1.001, 1.133, 1.001, 0.5], so the last sample is halved. Passing padtype='edge' gave [1.001, 0.999, 1.001, 0.999]. Passing padtype='edge' removed the droop in this test. Tested with scipy 1.18.1, numpy 2.5.4, Python 3.12.12.