Docs/formatting parameters and maths by sagar1sharma · Pull Request #1291 · automl/SMAC3

sagar1sharma · 2026-02-10T00:55:57Z

Thanks for contributing a pull request! Please ensure you have taken a look at the contribution guidelines: https://github.com/automl/SMAC3/blob/main/CONTRIBUTING.md

Reference Issues/PRs

Example: Fixes #1234. See also #3456.
Please use keywords (e.g., Fixes) to create link to the issues or pull requests you resolved, so that they will automatically be closed when your pull request is merged. See https://github.com/blog/1506-closing-issues-via-pull-requests

Description, Motivation and Context

Provide a short description of your changes.

Changes

Explain your changes in detail here.

Type Of Changes

What types of changes does your code introduce? Put an x in all the boxes that apply:

Bug fix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (fix or feature that would cause existing functionality to change)

What is Missing?

Please provide any further information that is needed to close this PR.
This could be another PR that is blocking this PR, or any further changes
that need to be addressed in the future.

Checklist

Go over all the following points, and put an x in all the boxes that apply.
Should there be a good argument to differ, please provide a reasoned explanation.

My change is based on the latest stage of the develop branch.
My change required a change of the documentation, which has been done.
I checked that the documentation can be build, visualizes everything as expected, and does not contain any warnings.
I have added/adapted tests to cover my changes.
The tests can be executed successfully.
I have added a description of the changes to CHANGELOG.rst.

Any other comments?

We appreciate your effort, but please be aware that we are not working full-time on this project. We welcome any contribution and value your effort, no matter how minor they are. If we are slow to review, either the pull request needs some benchmarking, tinkering, convincing, etc. or more likely the reviewers are simply busy. In either case, we ask for your understanding during the review process.

Thanks for contributing!

This PR fixes a bug in which integer hyperparameters would not be created correctly in the initial design. This went unnoticed, except in the case where it was part of a condition and the function deactivate_inactive_hyperparameters was called. This PR fixes this bug by mapping the random float that was sampled by, for example and LHD or Sobol sequence, to the float representation of an integer by calling _transform and _inverse_transform of the integer hyperparameter.

Updated description

Fix 531 correctly create integers in initial design

Co-authored-by: René Sass <mail@renesass.de>

Hotfix/early emails stale issues

Add Git-Flow Workflow Specification

…velopment

Implement RunHistoryPercentileEncoder for response value transformation.

* Added ecosystem information in documentation * Changed formatting from bullet points to sentences * Update index.md --------- Co-authored-by: sagar1sharma <sagarsharma.ss329.ss46@gmail.com> Co-authored-by: C. Benjamins <75323339+benjamc@users.noreply.github.com>

* Play With paralelisation * Update Parallelism Doc * Update paralelisation example * Update docs * Update parallelism docs * Update parallelism * Adapt Changelog.md --------- Co-authored-by: Lukas Fehring <l.fehring@ai.uni-hannover.de>

Bumps [actions/checkout](https://github.com/actions/checkout) from 5 to 6. - [Release notes](https://github.com/actions/checkout/releases) - [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md) - [Commits](actions/checkout@v5...v6) --- updated-dependencies: - dependency-name: actions/checkout dependency-version: '6' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

…apts-sampling 1184 prioracquisitionfunction adapts sampling

Update docs

Removed debug print statements and added warnings for non-positive values.

Removed comments that ensure minimal value and cutoff limits.

Fix formatting and indentation in docstring.

966 output trafos

1264 bug reset after initialization

Copilot

Pull request overview

This PR substantially expands SMAC3’s functionality and documentation, including enabling math rendering in MkDocs, adding adaptive capping support (with cutoff propagation to target functions), and switching the default RandomForest surrogate from pyrfr to a scikit-learn based implementation (while keeping an opt-in pyrfr path).

Changes:

Enable MathJax + arithmatex in MkDocs and update math formatting in acquisition-function docs.
Add adaptive capping plumbing: scenario flags, cutoff computation in intensifier, and cutoff forwarding through runners/facade signature handling.
Replace the default RF surrogate backend with scikit-learn, add optional pyrfr RF implementation and related installation/docs updates.

Reviewed changes

Copilot reviewed 71 out of 73 changed files in this pull request and generated 29 comments.

Show a summary per file

File	Description
tests/test_utils/test_configspace.py	Adds test for creating a uniform-transformed ConfigSpace copy.
tests/test_multi_objective/test_combined_function.py	Expands mean-aggregation test to loop over objective weights.
tests/test_model/test_rf.py	Adds correctness checks and new log_y-related RF test.
tests/test_main/test_smbo.py	Formatting + adds SMBO reset test.
tests/test_main/test_config_selector.py	Adds test for exhausted configuration space exception.
tests/test_intensifier/test_intensifier.py	Minor formatting and import changes.
tests/test_intensifier/test_hyperband_utils.py	Updates expected budgets/trials + adds `__main__` block.
tests/test_intensifier/test_adaptive_capping_intensifier.py	Adds adaptive capping behavior test with logging mock.
tests/fixtures/config_selector.py	Updates fixture to renamed `max_new_config_tries` parameter.
smac/utils/configspace.py	Updates continuous design transform + adds `create_uniform_configspace_copy`.
smac/scenario.py	Expands deterministic docs + adds adaptive capping + runtime cutoff params and validation.
smac/runner/target_function_script_runner.py	Adds `additional_info` plumbing and cutoff forwarding.
smac/runner/target_function_runner.py	Adds `additional_info` plumbing and cutoff forwarding.
smac/runner/dask_runner.py	Forwards `additional_info` into the single-worker runner.
smac/runner/abstract_runner.py	Passes `TrialInfo.additional_info` into runner calls.
smac/runhistory/runhistory.py	Adds `n_objectives` to support RUNNING trial costs for multi-objective.
smac/runhistory/encoder/power_transform_encoder.py	Adds Yeo–Johnson power transform encoder.
smac/runhistory/encoder/percentile_encoder.py	Adds percentile-based encoder (currently prints debug output).
smac/runhistory/encoder/gaussian_copula_encoder.py	Adds Gaussian-copula style encoder.
smac/runhistory/encoder/abstract_encoder.py	Adjusts warning string formatting.
smac/runhistory/dataclasses.py	Adds `additional_info` field to `TrialInfo`.
smac/multi_objective/parego.py	Updates ParEGO reference URL.
smac/model/random_forest/random_forest.py	Replaces pyrfr-based RF with scikit-learn-based RF implementation.
smac/model/random_forest/pyrfr/random_forest_pyrfr.py	Introduces optional pyrfr RF implementation in separate module.
smac/model/random_forest/pyrfr/init.py	Package init for pyrfr backend.
smac/main/smbo.py	Allows runner=None for ask/tell; changes validate to use submit/iter_results; adds reset behavior.
smac/main/exceptions.py	Adds `ConfigurationSpaceExhaustedException`.
smac/main/config_selector.py	Renames retries param + adds fallback random sampling and exhaustion exception.
smac/intensifier/successive_halving.py	Adds `uses_cutoffs` property + updates warning text.
smac/intensifier/intensifier.py	Adds `uses_cutoffs` + adaptive capping cutoff computation and propagation.
smac/intensifier/hyperband_utils.py	Fixes total budget computation to budget*configs.
smac/intensifier/abstract_intensifier.py	Adds `uses_cutoffs` abstract property + dominance/rejection logic enhancements.
smac/initial_design/random_design.py	Samples initial configs from uniformized ConfigSpace copy.
smac/initial_design/abstract_initial_design.py	Adjusts n_configs scaling logic for n_configs<=1.
smac/facade/old/multi_fidelity_facade_pyrfr.py	Adds legacy facade for pyrfr RF in multi-fidelity.
smac/facade/old/hyperparameter_optimization_facade_pyrfr.py	Adds legacy facade for pyrfr RF in HPO.
smac/facade/old/init.py	Exposes legacy pyrfr facades via `smac.facade.old`.
smac/facade/algorithm_configuration_facade.py	Formatting-only change.
smac/facade/abstract_facade.py	Makes target_function optional; threads cutoff requirement into signature args; runner meta handles None.
smac/acquisition/maximizer/local_search.py	Changes init-point de-duplication to use a set (order impact).
smac/acquisition/function/thompson.py	Updates math markup formatting.
smac/acquisition/function/probability_improvement.py	Updates math markup formatting.
smac/acquisition/function/expected_improvement.py	Updates math markup formatting.
smac/acquisition/function/confidence_bound.py	Refactors confidence bound into abstract base + LCB/UCB; updates docs.
setup.py	Moves pyrfr to extras; bumps scikit-learn minimum version.
mkdocs.yaml	Enables MathJax + arithmatex; switches docstring style to numpy; updates nav.
examples/2_multi_fidelity/1_mlp_epochs.py	Updates link formatting.
examples/1_basics/9_additional_arguments.py	New example for passing additional arguments to target functions.
examples/1_basics/8_warmstart.py	Updates usage text and argument naming.
examples/1_basics/8_adaptive_capping.py	New adaptive capping example.
examples/1_basics/7_parallelization_cluster.py	Refactors and updates cluster parallelization example.
examples/1_basics/6_priors.py	Fixes configspace reference for prior example.
examples/1_basics/3_ask_and_tell.py	Updates ask/tell guidance and target_function argument naming.
examples/1_basics/2_svm_cv.py	Updates config dict extraction logic.
docs/index.md	Adds ecosystem section + expands feature list.
docs/advanced_usage/9_parallelism.md	Expands and restructures parallelism documentation.
docs/advanced_usage/5_ask_and_tell.md	Adds note about ask/tell-only usage (terminology issue).
docs/advanced_usage/2_multi_fidelity.md	Fixes link to examples.
docs/advanced_usage/1_components.md	Adds warning about initial-design distributions being ignored + links.
docs/advanced_usage/13_adaptive_capping.md	New adaptive capping documentation page.
docs/advanced_usage/11_reproducibility.md	Adds reset-related warning.
docs/3_getting_started.md	Adds deterministic behavior info + SMBO section (includes a runhistory usage mistake).
docs/1_installation.md	Updates Python requirement + adds pyrfr optional installation instructions.
README.md	Updates install guidance + adds pyrfr optional section.
CHANGELOG.md	Adds 2.4.0 entry and additional 2.3.1 notes.
.gitignore	Adds SMAC/Dask log paths and extra artifacts.
.github/workflows/recent_reminder.yml	Bumps action-send-mail version.
.github/workflows/pytest.yml	Updates python matrix and GitHub actions versions.
.github/workflows/pre-commit.yml	Updates actions versions.
.github/workflows/pre-commit-update.yml	Updates actions versions.
.github/workflows/docs.yml	Updates actions versions.
.github/workflows/dist.yml	Updates actions versions.
.github/workflows/citation.yml	Updates checkout action version.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-10T01:05:02Z

smac/model/random_forest/random_forest.py

+        super().__init__(
+            DecisionTreeRegressor(),
+            n_estimators,
+            estimator_params=(
+                "criterion",
+                "max_depth",
+                "min_samples_split",
+                "min_samples_leaf",
+                "min_weight_fraction_leaf",
+                "max_features",


EPMRandomForest accepts a splitter argument but it is not included in estimator_params, so the underlying DecisionTreeRegressor instances will keep their default splitter (typically "best") regardless of the provided value. Add "splitter" to estimator_params (and ensure it is forwarded) so the parameter actually takes effect.

Copilot · 2026-02-10T01:05:03Z

smac/model/random_forest/random_forest.py

+        if self.log_y:
+            for tree, samples_idx in zip(self.estimators_, self.estimators_samples_):
+                curX = X[samples_idx]
+                curY = y[samples_idx]
+                preds = tree.apply(curX)
+                for k in np.unique(preds):
+                    tree.tree_.value[k, 0, 0] = np.log(np.exp(curY[preds == k]).mean())


fit() assumes self.estimators_samples_ is available when log_y is True. This attribute is not guaranteed to exist (e.g., when bootstrap=False), which would raise at runtime. Guard this logic (or enforce bootstrap) and ensure leaf-value adjustment works for non-bootstrap training too.

Copilot · 2026-02-10T01:05:03Z

smac/runner/target_function_runner.py

+        if "cutoff" in self._required_arguments and additional_info is not None:
+            kwargs["cutoff"] = additional_info["cutoff"]
+


Accessing additional_info["cutoff"] can raise KeyError when additional_info is provided but does not contain a cutoff entry. Consider checking for the key (or using .get) and raising a clearer error if it is required.

Copilot · 2026-02-10T01:05:03Z

docs/advanced_usage/5_ask_and_tell.md

    into the intensification process.

+Notice: if you are exclusively using the ask-and-tell interface and do not use `smac.optimize()`, then smac no longer
+is responsible for the evaluation of the trials and therefore the Facade no longer will require a specified `target_algorithm` argument.


The docs say the facade does not require a target_algorithm argument, but the constructor parameter is named target_function (and that's what users pass throughout the docs/examples). Please rename to target_function here to avoid confusion.

Suggested change

is responsible for the evaluation of the trials and therefore the Facade no longer will require a specified `target_algorithm` argument.

is responsible for the evaluation of the trials and therefore the Facade no longer will require a specified `target_function` argument.

Copilot · 2026-02-10T01:05:04Z

smac/main/smbo.py

+            self._runner.submit_trial(trial_info=trial)
+        costs = [trial_value.cost for _, trial_value in self._runner.iter_results()]
        np_costs = np.array(costs)


validate() submits trials and immediately drains iter_results() without waiting for completion. For asynchronous runners (e.g. Dask), this can return an empty/partial costs list and yield nan or incorrect averages. Consider waiting until all submitted validation trials are finished (e.g. loop on is_running()/wait() or track expected trial count) before computing the mean, and ensure only the validation results are aggregated.

Copilot · 2026-02-10T01:05:08Z

examples/1_basics/8_adaptive_capping.py

+
+import warnings
+
+from ConfigSpace import Categorical, Configuration, ConfigurationSpace, Float


Import of 'Categorical' is not used.

Suggested change

from ConfigSpace import Categorical, Configuration, ConfigurationSpace, Float

from ConfigSpace import Configuration, ConfigurationSpace, Float

Copilot · 2026-02-10T01:05:09Z

tests/test_intensifier/test_adaptive_capping_intensifier.py

@@ -0,0 +1,150 @@
+import logging


Import of 'logging' is not used.

Copilot · 2026-02-10T01:05:09Z

tests/test_intensifier/test_adaptive_capping_intensifier.py

@@ -0,0 +1,150 @@
+import logging
+import shutil
+from logging import Logger


Import of 'Logger' is not used.

Copilot · 2026-02-10T01:05:09Z

tests/test_intensifier/test_intensifier.py

 from smac.initial_design.random_design import RandomInitialDesign
 from smac.intensifier.intensifier import Intensifier
 from smac.main.config_selector import ConfigSelector
+from smac.runhistory import TrialInfo, TrialKey, TrialValue


Import of 'TrialInfo' is not used.
Import of 'TrialKey' is not used.
Import of 'TrialValue' is not used.

Copilot · 2026-02-10T01:05:09Z

tests/test_intensifier/test_adaptive_capping_intensifier.py

+    # start smac run
+    try:
+        smac.optimize()
+    except ConfigurationSpaceExhaustedException:


'except' clause does nothing but pass and there is no explanatory comment.

renesass and others added 30 commits December 30, 2022 12:05

Updated changelog

35e4f2d

Make isort

0324b0e

Fix typos

58f0ea2

Removed new line

044beac

Added alpha2 to docs selection

69820e9

update description

851e321

Add contact paragraph to README.

8c2f1b2

Increase days until stale issues are closed by stalebot (#904)

469ca05

Edit changelog

78488e7

Merge pull request #899 from automl/description

cd4ec30

Updated description

Add issue number

292e53d

Merge pull request #900 from automl/fix_531

9b98770

Fix 531 correctly create integers in initial design

Closes #907

fc07e9b

Clarify origin of configurations (#908)

c247853

Random forest speed up (#903)

2927921

Co-authored-by: René Sass <mail@renesass.de>

Add: workflow to update pre-commit versions (#874)

a0a8fd1

Co-authored-by: René Sass <mail@renesass.de>

Add Git-Flow Workflow Specification

4eb6528

[DOC] acquisition function maximizers

da841e2

Add stale reminder email action

f7209de

Merge pull request #919 from automl/hotfix/early_emails_stale_issues

ef4653d

Hotfix/early emails stale issues

Increase version number of ConfigSpace

871d656

Merge branch 'feature/improve_documentation' into development

e6c77a8

Add guideline for updating copyright notice years to CONTRIBUTING.md

9207944

Merge pull request #911 from automl/feature/specify_git-flow_workflows

cf6ee29

Add Git-Flow Workflow Specification

Update requirement for pyrfr (#920)

e5f9875

Update pre-commit: isort,flake8 version numbers

83b3dc4

Merge branch 'development' of https://github.com/automl/SMAC3 into de…

a6c60c8

…velopment

Format black

14ddf7a

Fix mypy

c861ae4

lhennig0103 and others added 25 commits November 13, 2025 17:07

Add RunHistoryPercentileEncoder class

a30ca72

Implement RunHistoryPercentileEncoder for response value transformation.

Implement Gaussian Copula Encoder for RunHistory

88d4fac

Ask and tell tp reproduce issue

4c84c07

1226 paralilisation (#1234)

70b57e2

* Play With paralelisation * Update Parallelism Doc * Update paralelisation example * Update docs * Update parallelism docs * Update parallelism * Adapt Changelog.md --------- Co-authored-by: Lukas Fehring <l.fehring@ai.uni-hannover.de>

Update docs

483daf0

Minor docs update

aabee10

Merge pull request #1274 from automl/1184-prioracquisitionfunction-ad…

fab2a06

…apts-sampling 1184 prioracquisitionfunction adapts sampling

Merge pull request #1278 from automl/1269-link-broken-in-docs

5f47524

Update docs

Clean up transform_response_values method

fad48bd

Removed debug print statements and added warnings for non-positive values.

Fix formatting issues in power_transform_encoder.py

8f39b8b

Remove unnecessary comments from GaussianCopulaEncoder

70cd4b1

Removed comments that ensure minimal value and cutoff limits.

Fix docstring formatting in transform_response_values

e1183dc

Fix docstring formatting in transform_response_values

f73b9e7

Correct docstring formatting in percentile_encoder.py

22aba86

Fix formatting and indentation in docstring.

Fix formatting

a878522

Merge pull request #1276 from automl/966_output_trafos

12d0945

966 output trafos

Fix smbo.reset

f4e896a

Add a warning to reproducibility

0302aaf

Remove unused code

ceb5235

Merge pull request #1279 from automl/1264-bug-reset-after-initialization

ac7da77

1264 bug reset after initialization

Fixed formatting for scenario parameters section

63b7592

Fixed parameters and math formula formatting

adde5c2

Scenario parameters final push

d60cb8d

Copilot AI review requested due to automatic review settings February 10, 2026 00:55

Copilot started reviewing on behalf of sagar1sharma February 10, 2026 00:56 View session

sagar1sharma closed this Feb 10, 2026

sagar1sharma deleted the docs/formatting_parameters_and_maths branch February 10, 2026 01:02

Copilot AI reviewed Feb 10, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Docs/formatting parameters and maths#1291

Docs/formatting parameters and maths#1291
sagar1sharma wants to merge 720 commits intomainfrom
docs/formatting_parameters_and_maths

sagar1sharma commented Feb 10, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Copilot AI Feb 10, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

20 participants

		if "cutoff" in self._required_arguments and additional_info is not None:
		kwargs["cutoff"] = additional_info["cutoff"]

	is responsible for the evaluation of the trials and therefore the Facade no longer will require a specified `target_algorithm` argument.
	is responsible for the evaluation of the trials and therefore the Facade no longer will require a specified `target_function` argument.


		import warnings

		from ConfigSpace import Categorical, Configuration, ConfigurationSpace, Float

	from ConfigSpace import Categorical, Configuration, ConfigurationSpace, Float
	from ConfigSpace import Configuration, ConfigurationSpace, Float

Uh oh!

Conversation

sagar1sharma commented Feb 10, 2026

Reference Issues/PRs

Description, Motivation and Context

Changes

Type Of Changes

What is Missing?

Checklist

Any other comments?

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 10, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

20 participants