CLI Helpers

_run_from_project(project_root: Path, verbose: bool, create_report: bool, results_dir: str) → None

Run experiments from a project directory.

Loads algorithms, metrics, and configuration from a project directory and executes the training workflow. This function is used when running experiments from project files.

Parameters:

project_rootpathlib.Path

Path to the project root directory containing algorithms.py, metrics.py, and settings.py files.

verbosebool

Whether to enable verbose logging output.

create_reportbool

Whether to generate an HTML report after experiments complete.

results_dirstr

Directory path where experiment results will be saved.

Raises:

FileNotFoundError

If required project files are missing.

ImportError

If there are issues importing project modules.

AttributeError

If required functions or objects are not found in modules.

ValueError

If there are configuration or data validation errors.

Notes

This function expects the following files to exist in the project root: - algorithms.py: Contains algorithm definitions - metrics.py: Contains metric configurations - settings.py: Contains a create_configuration function

_run_from_config(project_root: Path, verbose: bool, create_report: bool, results_dir: str, config_file: str) → None

Run experiments from a saved configuration file.

Loads a previously saved experiment configuration and reruns the experiments with the same settings. Validates environment compatibility and dataset integrity before execution.

Parameters:

project_rootPath

Path to the project root directory.

verbosebool

Whether to enable verbose logging output.

create_reportbool

Whether to generate an HTML report after experiments complete.

results_dirstr

Directory path where experiment results will be saved.

config_filestr

Name of the results directory (without path) to load the configuration from.

Raises:

RuntimeError

If the configuration was created with a different Brisk version.

FileNotFoundError

If the configuration file or dataset files are missing.

ValueError

If dataset validation fails or there are configuration errors.

Notes

The configuration file should be located at: {project_root}/results/{config_file}/run_config.json

This function performs several validation checks: - Package version compatibility - Dataset file existence and integrity - Environment package compatibility

load_sklearn_dataset(name: str) → dict | None

Load a dataset from scikit-learn.

Parameters:

name{‘iris’, ‘wine’, ‘breast_cancer’, ‘diabetes’, ‘linnerud’}

Name of the dataset to load

Returns:

dict or None

Loaded dataset object or None if not found

_validate_dataset(dataset_path: str, metadata: Dict[str, Any]) → None

Validate dataset file against expected metadata.

Checks that the dataset file exists and matches the expected structure including number of samples, features, and feature names.

Parameters:

dataset_pathstr

Path to the dataset file to validate.

metadatadict

Dictionary containing expected dataset metadata with keys: - table_name: Name of the table in the dataset file - num_samples: Expected number of rows/samples - num_features: Expected number of columns/features - feature_names: List of expected feature column names

Raises:

FileNotFoundError

If the dataset file does not exist.

ValueError

If the dataset dimensions or feature names don’t match expectations.

Notes

This function loads the dataset using the IOService and performs validation to ensure data integrity for experiment reruns.

_handle_incompatible(config_file: str, differences: List[EnvironmentDiff], env_manager: EnvironmentManager) → None

Handle environment compatibility warnings and provide guidance.

Displays warnings about environment differences and provides instructions for recreating the original environment or continuing with potential compatibility issues.

Parameters:

config_filestr

Name of the configuration file being rerun.

differenceslist

List of environment differences detected by EnvironmentManager.

env_managerEnvironmentManager

Environment manager instance used for comparison.

Notes

This function identifies critical package differences and provides detailed instructions for: - Recreating the original environment - Exporting environment requirements - Checking detailed environment differences

The function categorizes differences as critical if they involve: - Missing packages - Incompatible versions of critical packages - Python version changes