Saving and loading of SplineModel yields incorrect result.init_fit

[paulmueller]

While working on #985, I noticed that the init_fit property is not properly loaded after loading a saved SplineModel:

import numpy as np
from lmfit.models import SplineModel
from lmfit.model import save_modelresult, load_modelresult

model_file = 'spline_modelresult.sav'
xx = np.linspace(-10, 10, 100)
yy = 0.6*np.exp(-(xx**2)/(1.3**2))

spl_model = SplineModel(xknots=np.linspace(-10, 10, 5))
params = spl_model.guess(yy, xx)
result = spl_model.fit(yy, params, x=xx)

save_modelresult(result, model_file)

result2 = load_modelresult(model_file)

print(result.init_fit)
[3.14000311e-26 3.57373261e-04 1.40993305e-03 3.12840834e-03
 5.48352810e-03 8.44602130e-03 1.19866169e-02 1.60760439e-02
 2.06850312e-02 2.57843078e-02 3.13446027e-02 3.73366449e-02
 4.37311633e-02 5.04988868e-02 5.76105445e-02 6.50368654e-02
 7.27485783e-02 8.07164123e-02 8.89110963e-02 9.73033593e-02
 1.05863930e-01 1.14563538e-01 1.23372912e-01 1.32262781e-01
 1.41203873e-01 1.50166919e-01 1.59122646e-01 1.68041784e-01
 1.76895061e-01 1.85653207e-01 1.94286951e-01 2.02767022e-01
 2.11064148e-01 2.19149059e-01 2.26992483e-01 2.34565150e-01
 2.41837788e-01 2.48781126e-01 2.55365894e-01 2.61562821e-01
 2.67342635e-01 2.72676065e-01 2.77533840e-01 2.81886689e-01
 2.85705342e-01 2.88960527e-01 2.91622973e-01 2.93663409e-01
 2.95052565e-01 2.95761168e-01 2.95761168e-01 2.95052565e-01
 2.93663409e-01 2.91622973e-01 2.88960527e-01 2.85705342e-01
 2.81886689e-01 2.77533840e-01 2.72676065e-01 2.67342635e-01
 2.61562821e-01 2.55365894e-01 2.48781126e-01 2.41837788e-01
 2.34565150e-01 2.26992483e-01 2.19149059e-01 2.11064148e-01
 2.02767022e-01 1.94286951e-01 1.85653207e-01 1.76895061e-01
 1.68041784e-01 1.59122646e-01 1.50166919e-01 1.41203873e-01
 1.32262781e-01 1.23372912e-01 1.14563538e-01 1.05863930e-01
 9.73033593e-02 8.89110963e-02 8.07164123e-02 7.27485783e-02
 6.50368654e-02 5.76105445e-02 5.04988868e-02 4.37311633e-02
 3.73366449e-02 3.13446027e-02 2.57843078e-02 2.06850312e-02
 1.60760439e-02 1.19866169e-02 8.44602130e-03 5.48352810e-03
 3.12840834e-03 1.40993305e-03 3.57373261e-04 3.14000311e-26]

print(result2.init_fit)
[-10.          -9.7979798   -9.5959596   -9.39393939  -9.19191919
  -8.98989899  -8.78787879  -8.58585859  -8.38383838  -8.18181818
  -7.97979798  -7.77777778  -7.57575758  -7.37373737  -7.17171717
  -6.96969697  -6.76767677  -6.56565657  -6.36363636  -6.16161616
  -5.95959596  -5.75757576  -5.55555556  -5.35353535  -5.15151515
  -4.94949495  -4.74747475  -4.54545455  -4.34343434  -4.14141414
  -3.93939394  -3.73737374  -3.53535354  -3.33333333  -3.13131313
  -2.92929293  -2.72727273  -2.52525253  -2.32323232  -2.12121212
  -1.91919192  -1.71717172  -1.51515152  -1.31313131  -1.11111111
  -0.90909091  -0.70707071  -0.50505051  -0.3030303   -0.1010101
   0.1010101    0.3030303    0.50505051   0.70707071   0.90909091
   1.11111111   1.31313131   1.51515152   1.71717172   1.91919192
   2.12121212   2.32323232   2.52525253   2.72727273   2.92929293
   3.13131313   3.33333333   3.53535354   3.73737374   3.93939394
   4.14141414   4.34343434   4.54545455   4.74747475   4.94949495
   5.15151515   5.35353535   5.55555556   5.75757576   5.95959596
   6.16161616   6.36363636   6.56565657   6.76767677   6.96969697
   7.17171717   7.37373737   7.57575758   7.77777778   7.97979798
   8.18181818   8.38383838   8.58585859   8.78787879   8.98989899
   9.19191919   9.39393939   9.5959596    9.7979798   10.        ]

Apparently, there is something going wrong when parsing the parameters from the saved model. After digging in the code, I noticed that the methods result.eval and result2.eval return different results.

[paulmueller]

I would also be interested in working on this one, but maybe you @newville already have an idea where to start exactly.