Separate DiscreteSINDy as a new class #654

yb6599 · 2025-10-20T20:12:19Z

This PR separates DiscreteSINDy as a new class that derives from _BaseSINDy following the discussion in #351.
Also fixes #650.

…atibility with multiple trajectories for DiscreteSINDy

yb6599 · 2025-10-20T20:53:33Z

The scikit-time tests fail as the scikit-time API has remained unchanged. How should I proceed with this? @Jacob-Stevens-Haas

Jacob-Stevens-Haas · 2025-10-20T21:46:45Z

Part of the problem with the face-to-face conversations is that I can't remember them.

Refresh my memory. What was the issue removing the parameter discrete_time from the scikit-time classes?

yb6599 · 2025-10-21T16:48:24Z

Removing discrete_time should not be a problem here. However, we had not decided how we would deal with the discrete time implementations of the scikit-time API.

…e overview notebook

codecov · 2025-10-21T18:27:34Z

Codecov Report

❌ Patch coverage is 95.80420% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 94.85%. Comparing base (472e5a2) to head (4fe2aa4).

Files with missing lines	Patch %	Lines
pysindy/_core.py	95.65%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #654      +/-   ##
==========================================
- Coverage   95.03%   94.85%   -0.18%     
==========================================
  Files          38       38              
  Lines        4106     4138      +32     
==========================================
+ Hits         3902     3925      +23     
- Misses        204      213       +9

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

yb6599 · 2025-10-21T18:35:49Z

Do we need _BaseSINDy.print() if we have SINDy.print() and DiscreteSINDy.print()?

Jacob-Stevens-Haas

90% of the way there, thanks for this long work!

Few changes, few nits, few documentation/test things

test/utils/test_utils.py

Jacob-Stevens-Haas · 2025-11-04T18:11:42Z

examples/1_feature_overview/example.py


 # Compute derivatives with a finite difference method, for comparison
-x_dot_test_computed = model.differentiate(x_test, t=dt)
+x_dot_test_computed = model.differentiation_method._differentiate(x_test, t=dt)


_differentiate is private - we can't show that in a tutorial.

Suggested change

x_dot_test_computed = model.differentiation_method._differentiate(x_test, t=dt)

x_dot_test_computed = model.differentiation_method(x_test, t=dt)

Jacob-Stevens-Haas · 2025-11-04T18:11:53Z

examples/1_feature_overview/example.py


 # Compute derivatives with a finite difference method, for comparison
-x_dot_test_computed = model.differentiate(x_test, t=dt)
+x_dot_test_computed = model.differentiation_method._differentiate(x_test, t=dt)


Suggested change

x_dot_test_computed = model.differentiation_method._differentiate(x_test, t=dt)

x_dot_test_computed = model.differentiation_method(x_test, t=dt)

Jacob-Stevens-Haas · 2025-11-04T18:12:04Z

examples/1_feature_overview/example.py


 # Compute derivatives with a finite difference method, for comparison
-x_dot_test_computed = model.differentiate(x_test, t=dt)
+x_dot_test_computed = model.differentiation_method._differentiate(x_test, t=dt)


Suggested change

x_dot_test_computed = model.differentiation_method._differentiate(x_test, t=dt)

x_dot_test_computed = model.differentiation_method(x_test, t=dt)

Jacob-Stevens-Haas · 2025-11-04T18:25:44Z

examples/1_feature_overview/example.py

 for i in range(1, n_steps):
    x_train_map[i] = f(x_train_map[i - 1]) + eps * np.random.randn()
-model = ps.SINDy(discrete_time=True)
+model = ps.DiscreteSINDy()


Remove the two discrete examples from feature overview, keep them in the docstring for DiscreteSINDy.

Use the sphinx extension matplotlib.sphinxext.plot_directive to show the plots in the DiscreteSINDy documentation

pysindy/_core.py

Jacob-Stevens-Haas · 2025-11-04T18:28:15Z

pysindy/_core.py


    feature_library: BaseFeatureLibrary
    optimizer: _BaseOptimizer
+    discrete_time: bool


Jacob-Stevens-Haas · 2025-11-04T18:29:34Z

pysindy/_core.py

        return self.optimizer.coef_

-    def equations(self, precision: int = 3) -> list[str]:
+    def equations(self, precision: int = 3, discrete_time=False) -> list[str]:


Override this function, don't add parameters.

pysindy/_core.py

Jacob-Stevens-Haas · 2025-11-04T18:38:32Z

pysindy/_core.py

+    if sys.version_info.minor < 10:
+        mixed_trajectories = (
+            isinstance(x, Sequence)
+            and (
+                not isinstance(x_dot, Sequence)
+                and x_dot is not None
+                or not isinstance(u, Sequence)
+                and u is not None
+            )
+            or isinstance(x_dot, Sequence)
+            and not isinstance(x, Sequence)
+            or isinstance(u, Sequence)
+            and not isinstance(x, Sequence)
+        )
+    else:


3.9 reached EOL, require 3.10 in pyproject.toml and remove this condition

yb6599 added 10 commits October 6, 2025 15:24

CLN: Separate DiscreteSINDy as its own class

9df5bcc

CLN: ensure consistency of shapes and types of x_next and u, and comp…

bbc6c2c

…atibility with multiple trajectories for DiscreteSINDy

TST: update tests for DiscreteSINDy

7a093b4

DOC: Update documentation for DiscreteSINDy

9cf9d28

CLN: use predict from BaseSINDy for SINDy and DiscreteSINDy models

38868a8

CLN: Remove SINDy.differentiate()

ee7866e

FIX: process trajectories only when x_dot is None

393f3e1

CLN: lint the code

404d6b3

CLN: resolve merge conflicts

2591d3d

FIX: fix _tensordot_to_einsum error

8821a49

yb6599 added 2 commits October 21, 2025 09:55

CLN: remove discrete_time from scikit-time API

194b548

CLN: remove SINDy.differentiate() and update DiscreteSINDy for featur…

37feb2c

…e overview notebook

FIX: remove SINDyPI conditionals from DiscreteSINDy.print()

59ce33f

TST: add test for validate_control_variables function

4fe2aa4

Jacob-Stevens-Haas requested changes Nov 4, 2025

View reviewed changes

	x_dot_test_computed = model.differentiation_method._differentiate(x_test, t=dt)
	x_dot_test_computed = model.differentiation_method(x_test, t=dt)

Separate DiscreteSINDy as a new class #654

Are you sure you want to change the base?

Separate DiscreteSINDy as a new class #654

Uh oh!

Conversation

yb6599 commented Oct 20, 2025

Uh oh!

yb6599 commented Oct 20, 2025

Uh oh!

Jacob-Stevens-Haas commented Oct 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

yb6599 commented Oct 21, 2025

Uh oh!

codecov bot commented Oct 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

yb6599 commented Oct 21, 2025

Uh oh!

Jacob-Stevens-Haas left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Jacob-Stevens-Haas Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Jacob-Stevens-Haas Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Jacob-Stevens-Haas Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Jacob-Stevens-Haas Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Jacob-Stevens-Haas Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Jacob-Stevens-Haas Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Jacob-Stevens-Haas Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Jacob-Stevens-Haas commented Oct 20, 2025 •

edited

Loading

codecov bot commented Oct 21, 2025 •

edited

Loading