luckystarufo / pysindy Goto Github PK

I tried to run examples/example-3-sindypde-reactiondiffusion.ipynb but I get this error:

FileNotFoundError: [Errno 2] No such file or directory: '../datasets/reaction_diffusion.mat'

I noticed that there's a similar file with a .m extension so I renamed it to .mat but then I get a different error:

ValueError: Unknown mat file type, version 95, 116

In the top level README.md the link entitled ISINDy example which points to
https://github.com/luckystarufo/pySINDy/blob/master/examples/example-4-isindy-xxx.ipynb is broken.

Hello!

I found some annoying bug with GeneralizedLibrary(), we may see this into Example 15.
When we use a 'tensor_array' it results in repeated features (see attached picture).

Then, if there are duplicate functions in the library, it may cause misidentification (see example):

The reason why it happens is pretty simple - because we use the bias term when build tensored library. From my point, it's needed to exclude the product of bias term with other terms from another feature library.

There is an obvious workarround: just do not use bias term with default library and use it as a single library.

Thank you!

In pySINDy/linting.py there is a threshold constant defined of .85 but it isn't actually used in printing messages, etc.

Hey there,

Just a tiny inconvenience:
downloading via pip does not download the PySINDy.utils folder. I added it manually

When I tried to run the SINDy: Van Der Pol Oscillator example on my local machine the imports didn't work.
For example from pySINDy.sindy import SINDy should be from pysindy import SINDy
Besides that I couldnt figure out how from pySINDy.utils.generator import van_der_pol_generator

Installed PySINDy with pip install...

In the paper, Discovering governing equations from data, and the Matlab code for the Lorenz Generator it seems there is a noise term added to the derivative matrix. However, I cannot find this in the Python code. Is noise added to the Python derivative matrix as well?

import pysindy as ps
KeyError Traceback (most recent call last)
/usr/local/lib/python3.7/dist-packages/pkg_resources/init.py in requires(self, extras)
2738 try:
-> 2739 deps.extend(dm[safe_extra(ext)])
2740 except KeyError as e:

KeyError: 'alldeps'

The above exception was the direct cause of the following exception:

UnknownExtra Traceback (most recent call last)
6 frames
/usr/local/lib/python3.7/dist-packages/pkg_resources/init.py in requires(self, extras)
2741 raise UnknownExtra(
2742 "%s has no such extra feature %r" % (self, ext)
-> 2743 ) from e
2744 return deps
2745

UnknownExtra: scikit-learn 1.0.2 has no such extra feature 'alldeps'

When developing array in example 1, is it possible to determine all variable names from that array? I understand that fitting the model provides the variables it deems significant but is it possible to determine all variables corresponding to the derivatives in the data developed by the Van der Pol Generator?

1. Why are all of the methods in ISINDY static?
2. You have some non-intuitive variable names, like _n in isinidy.py
3. In generator.py, you have a lot of int constants that should be named constants, such as van_der_pol_generator(mu, x_init, dt=0.01, len_t=100):