Mark Bakker
It is difficult to find what setting are available and what they do: 'prec', 'evap' (not quite sure that is what they are called). Maybe add a Notebook to find...
Can be reproduced with many datasets. Example: ``` data = pd.read_csv('https://raw.githubusercontent.com/pastas/test-datasets/main/vanderspek_bakker_2017/B09B0124_DE-KOOG.csv', index_col=0, parse_dates=True, sep=";").loc["1980":] head = data.iloc[:,0] prec = data.iloc[:,1].dropna() # m/d evap = data.iloc[:,2].dropna() # m/d # ml =...
`tmax` for the 4-parameter function needs more testing. It goes wrong, for example: ``` fourp = ps.FourParam() fp = fourp.step([2, 1, 50, 100], dt=1, cutoff=0.95) ```
The default method for parameter estimation in Pastas, `scipy.optimize.least_squares` has a bug when the absolute value of a parameter is much smaller than 1. See here: https://github.com/scipy/scipy/issues/11258 We should do...
1. All stressmodels will only accept equidistant Pandas time series only (either a series or dataframe with one column). 2. Stressmodels will only accept instance of response function 3. solve...
In StressModel.init, the meanstress is computed as: if meanstress is None: meanstress = stress.series.std() Why is that? The it should either be called stdstress rather than meanstress. Or the name...
`import timml as tml` no pylab but `import numpy as np` `import matplotlib.pyplot as plt`
experimental solve_mp doesn't seem to work anymore. I wonder if it ever worked since the move to numba.
contour levels has not much of any documentation. trace has no docs.
The following traceline gets stuck and doesn't reach the line-sink ``` z = np.arange(20, -0.1, -0.5) ml = Model3D(z=z, kaq=10) L = 50 x = np.linspace(0, 50, 2) y =...