v1.0.0

PySR v1.0.0 Release Notes

PySR 1.0.0 introduces new features for imposing specific functional forms and finding parametric expressions. It also includes TensorBoard support, along with significant updates to the core algorithm, including some important bug fixes. The default hyperparameters have also been updated based on extensive tuning, with a maxsize of 30 rather than 20.

Major New Features

Expression Specifications

PySR 1.0.0 introduces new ways to specify the structure of equations through "Expression Specifications", that expose the new backend feature of AbstractExpression:

Template Expressions

TemplateExpressionSpec allows you to define a specific structure for your equations. For example:

expression_spec = TemplateExpressionSpec(["f", "g"], "((; f, g), (x1, x2, x3)) -> sin(f(x1, x2)) + g(x3)")

Parametric Expressions

ParametricExpressionSpec enables fitting expressions that can adapt to different categories of data with per-category parameters:

expression_spec = ParametricExpressionSpec(max_parameters=2)
model = PySRRegressor(
    expression_spec=expression_spec
    binary_operators=["+", "*", "-", "/"],
)
model.fit(X, y, category=category)  # Pass category labels

Improved Logging with TensorBoard

The new TensorBoardLoggerSpec enables logging of the search process, as well as hyperparameter recording, which exposes the AbstractSRLogger feature of the backend:

logger_spec = TensorBoardLoggerSpec(
    log_dir="logs/run",
    log_interval=10,  # Log every 10 iterations
)
model = PySRRegressor(logger_spec=logger_spec)

Features logged include:

• Loss curves over time at each complexity level
• Population statistics
• Pareto "volume" logging (measures performance over all complexities with a single scalar)
• The min loss over time

Algorithm Improvements

Updated Default Parameters

The default hyperparameters have been significantly revised based on testing:

• Increased default maxsize from 20 to 30, as I noticed that many people use the defaults, and this maxsize would allow for more accurate expressions.
• New mutation operator weights optimized for better performance, along the new mutation "rotate tree."
• Improved search parameters tuned using Pareto front volume calculations.
• Default niterations increased from 40 to 100, also to support better accuracy (at the expense of slightly longer default search times).

Core Changes

• New output organization: Results are now stored in outputs/<run_id>/ rather than in the directory of execution.
• Improved performance with better parallelism handling
• Support for Python 3.10+
• Updated Julia backend to version 1.10+
• Fix for aliasing issues in crossover operations

Breaking Changes

• Minimum Python version is now 3.10, and minimum Julia version is 1.10
• Output file structure has changed to use directories
• Parameter name updates:
  • equation_file → output_directory + run_id
  • Added clearer naming for parallelism options, such as parallelism="serial" rather than the old multithreading=False, procs=0 which was unclear

Documentation

The documentation has a new home at https://ai.damtp.cam.ac.uk/pysr/

v0.19.4

What's Changed

• Create load_all_packages to install Julia extensions by @MilesCranmer in #688
• Apptainer definition file for PySR by @wkharold in #687
• JuliaCall 0.9.23 by @MilesCranmer in #703
  • build(deps): bump juliacall from 0.9.21 to 0.9.22 by @dependabot in #695

New Contributors

• @wkharold made their first contribution in #687

Full Changelog: v0.19.3...v0.19.4

v0.19.3

What's Changed

• build(deps): bump juliacall from 0.9.20 to 0.9.21 by @dependabot in #678

Full Changelog: v0.19.2...v0.19.3

v0.19.2

What's Changed

• Avoid automatic upgrade to Julia 1.11 by @MilesCranmer in #671

Full Changelog: v0.19.1...v0.19.2

v0.19.1

What's Changed

• Bump docker/setup-qemu-action from 2 to 3 by @dependabot in #506
• fix: from pysr import * by @MilesCranmer in #670

Full Changelog: v0.19.0...v0.19.1

v0.19.0

What's Changed

• BREAKING: Disable automatic sympy simplification by @MilesCranmer in #658
• Build: update numpy version by @MilesCranmer in #650
• Build: bump docker/build-push-action from 5 to 6 by @dependabot in #652

Full Changelog: v0.18.5...v0.19.0

v0.18.5

What's Changed

New features

• Per-variable custom complexities by @MilesCranmer in #649

  model.fit(X, y, complexity_of_variables=[1, 3])
  # run a search with feature 1 having complexity 1 and feature 2 with complexity 3

• Automatically suggest similar parameters by @MilesCranmer in #620

Other

• Bump julia-actions/cache from 1 to 2 by @dependabot in #621
• Update pysr_demo.ipynb by @VishalJ99 in #624
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #612
• Bump docker/login-action from 2 to 3 by @dependabot in #509
• More extensive typing stubs and associated refactoring by @MilesCranmer in #609

Full Changelog: v0.18.4...v0.18.5

Backend changes

New features

• Allow per-variable complexity (MilesCranmer/SymbolicRegression.jl#324) (@MilesCranmer)

Other

• ci: split up test suite into multiple runners (MilesCranmer/SymbolicRegression.jl#311) (@MilesCranmer)
• chore(deps): bump julia-actions/cache from 1 to 2 (MilesCranmer/SymbolicRegression.jl#315) (https://github.com/dependabot[bot])
• CompatHelper: bump compat for DynamicQuantities to 0.14, (keep existing compat) (MilesCranmer/SymbolicRegression.jl#317) (@github-actions[bot])
• Use DispatchDoctor.jl to wrap entire package with @stable (MilesCranmer/SymbolicRegression.jl#321) (@MilesCranmer)
• CompatHelper: bump compat for MLJModelInterface to 1, (keep existing compat) (MilesCranmer/SymbolicRegression.jl#322) (@github-actions[bot])
• Mark more functions as stable (MilesCranmer/SymbolicRegression.jl#323) (@MilesCranmer)
• Refactor tests to use TestItems.jl (MilesCranmer/SymbolicRegression.jl#325) (@MilesCranmer)

Full Changelog: MilesCranmer/SymbolicRegression.jl@v0.24.4...v0.24.5

New Contributors

• @VishalJ99 made their first contribution in #624

v0.18.4

Frontend changes

• Add dimensionless constants mode; update Python version constraints; upgrade juliacall to 0.9.20 (#608) (@MilesCranmer)
• Fix sign typo in example docs (#611) (@hvaara)

Full Changelog: v0.18.3...v0.18.4

Backend changes

• Up to 40% speedup for default settings via more parallelism inside workers (MilesCranmer/SymbolicRegression.jl#304) (@MilesCranmer)
• feat: use ? for wildcard units instead of ⋅ (MilesCranmer/SymbolicRegression.jl#307) (@MilesCranmer)
• refactor: fix some more type instabilities (MilesCranmer/SymbolicRegression.jl#308) (@MilesCranmer)
• refactor: remove unused Tricks dependency (MilesCranmer/SymbolicRegression.jl#309) (@MilesCranmer)
• Add option to force dimensionless constants (MilesCranmer/SymbolicRegression.jl#310) (@MilesCranmer)

Full Changelog: MilesCranmer/SymbolicRegression.jl@v0.24.2...v0.24.4

New Contributors

• @hvaara made their first contribution in #611

v0.18.3

Frontend changes

• Automated update to backend: v0.24.3 by @github-actions in #605

Full Changelog: v0.18.2...v0.18.3

Backend changes

Full Changelog: MilesCranmer/SymbolicRegression.jl@v0.24.1...v0.24.2

v0.18.2

Frontend changes

• Add missing greater operator in sympy mapping by @MilesCranmer in #590
• Bump julia-actions/setup-julia from 1 to 2 by @dependabot in #591
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #537
• Automated update to backend: v0.24.2 by @MilesCranmer in #598

Full Changelog: v0.18.1...v0.18.2

Backend changes

Merged pull requests:

• Bump julia-actions/setup-julia from 1 to 2 (MilesCranmer/SymbolicRegression.jl#300) (@dependabot[bot])
• [pre-commit.ci] pre-commit autoupdate (MilesCranmer/SymbolicRegression.jl#301) (@pre-commit-ci[bot])
• A small update on examples.md for 1-based indexing (MilesCranmer/SymbolicRegression.jl#302) (@liuyxpp)
• Fixes for Julia 1.11 (MilesCranmer/SymbolicRegression.jl#303) (@MilesCranmer)

Closed issues:

• API Overhaul (MilesCranmer/SymbolicRegression.jl#187)
• [Feature]: Training on high dimensions X (MilesCranmer/SymbolicRegression.jl#299)

Full Changelog: MilesCranmer/SymbolicRegression.jl@v0.24.1...v0.24.2