lsq=best fails for long timeseries

[guziy]

Hi:

I realized that when lsq=best falls back to lsq=normal which fails when we pass a long timeseries with missing values. Here is the reproducer:

import ttide
import matplotlib.pyplot as plt
import numpy as np

def main():

    N = 1_000_200

    # time is in minutes
    dt = 1.
    T = 12 * 60.

    t = np.arange(0, N * dt, dt)

    h = np.sin(2 * np.pi / T * t)

    h[10:20] = np.nan

    con = ttide.t_tide(h, dt = 1. / 60.)

    # print(con)

    nshow = 100
    plt.figure()
    plt.plot(t[:nshow] / 60., h[:nshow])
    plt.plot(t[:nshow] / 60., con["xout"][:nshow])
    plt.show()


    
if __name__ == "__main__":
    main()

I get the following exception

/home/huziy/Python/anaconda3/envs/py3.10/bin/python /home/huziy/Python/Projects/test-ttide/src/test.py
Traceback (most recent call last):
  File "/home/huziy/Python/Projects/test-ttide/src/test.py", line 35, in <module>
    main()
  File "/home/huziy/Python/Projects/test-ttide/src/test.py", line 22, in main
    con = ttide.t_tide(h, dt = 1. / 60.)
  File "/home/huziy/Python/anaconda3/envs/py3.10/lib/python3.10/site-packages/ttide/t_tide.py", line 493, in t_tide
    xr = tu.fixgaps(xres)
  File "/home/huziy/Python/anaconda3/envs/py3.10/lib/python3.10/site-packages/ttide/t_utils.py", line 168, in fixgaps
    bd[:gd.min()]=False
  File "/home/huziy/Python/anaconda3/envs/py3.10/lib/python3.10/site-packages/numpy/core/_methods.py", line 44, in _amin
    return umr_minimum(a, axis, None, out, keepdims, initial, where)
ValueError: zero-size array to reduction operation minimum which has no identity

ttide version

In [2]: ttide.__version__
Out[2]: '0.3lfk'