It seems possible to use numpy.random.choice for bootstrap resampling (https://medium.com/swlh/bootstrap-sampling-using-pythons-numpy-85822d868977), besides sklearn.utils.resample, which is used in timeseries2velocity.py.

Numpy is preferred to avoid extra dependencies of mintpy.

scipy.stats.bootstrap is added to the scipy v1.7.0.