[BUG]Problems encountered in expanding the library(simulation step) #581

I added λx to the original quadratic model (including first-order and second-order terms), that is, I expanded my library to ['x0', 'x1', 'x2', 'x3', 'x4', 'x5', 'x6', 'x7', 'x0^2', 'x0 x1', 'x0 x2', 'x0 x3', 'x0 x4', 'x0 x5', 'x0 x6', 'x0 x7', 'x1^2', 'x1 x2', 'x1 x3', 'x1 x4', 'x1 x5', 'x1 x6', 'x1 x7', 'x2^2', 'x2 x3', 'x2 x4', 'x2 x5', 'x2 x6', 'x2 x7', 'x3^2', 'x3 x4', 'x3 x5', 'x3 x6', 'x3 x7', 'x4^2', 'x4 x5', 'x4 x6', 'x4 x7', 'x5^2', 'x5 x6', 'x5 x7', 'x6^2', 'x6 x7', 'x7^2', '27.550698135319728 x0', '27.30162539091796 x1', '17.823184677763013 x2', '19.962163132722978 x3', '18.583052660862982 x4', '19.606816257956034 x5', '22.26123984319782 x6', '23.015871481922936 x7']
New library has 52 items!

I successfully got the fitted model, for example for x0：
(x0)' = -0.002060 x0 x1 + 0.024776 x0 x2 + 0.001030 x0 x3 + 0.013200 x0 x4 + 0.032148 x0 x5 + -0.011575 x0 x6 + 0.001541 x0 x7 + 0.015141 x1^2 + 0.015628 x1 x2 + 0.046373 x1 x3 + -0.001168 x1 x4 + 0.024892 x1 x5 + 0.001096 x1 x7 + -0.017467 x2^2 + 0.017878 x2 x3 + -0.006292 x2 x4 + 0.024463 x2 x5 + 0.010341 x2 x6 + 0.005851 x2 x7 + 0.009305 x3^2 + -0.010523 x3 x4 + 0.029996 x3 x5 + -0.002509 x3 x6 + 0.011126 x3 x7 + -0.006788 x4^2 + -0.004106 x4 x5 + 0.004906 x4 x7 + 0.001332 x5 x6 + 0.015972 x6^2 + -0.003034 x6 x7 + -0.015874 x7^2 + 0.025424 27.30162539091796 x1 + -0.006484 17.823184677763013 x2 + 0.001398 19.962163132722978 x3 + 0.001691 18.583052660862982 x4

But it is clear that terms like 0.025424 27.30162539091796 x1 are not optimally represented. When I then run simulation, I get the error Simulation error: cannot reshape array of size 44 into shape (12000,newaxis). I don't know why the newly added terms are not recognized during model.simulate, even though they were successfully recognized during the model.fit step.

I tried to manually modify the parameters to make the correlation term a first-order term with only one coefficient (which exists at the same time as the ordinary first-order term, i.e. there may be two x0 terms), but the same problem still exists.

Thank you very much for your guidance and help!

Reproducing code example:

import pysindy
<< your code here >>

Error message:

Simulation error: cannot reshape array of size 44 into shape (12000,newaxis)
Simulation error: [[-1.21098846e-01 1.10845991e-01 -5.11850682e-02 2.38984799e-01......

PySINDy/Python version information: