helix.pages package

Submodules

helix.pages.1_New_Experiment module

helix.pages.1_New_Experiment.get_last_column_name(file) → str | None

Returns the name of the last column in the uploaded file.

Parameters:

file – A Streamlit UploadedFile object or None.

Returns:

The name of the last column as a string, or None if unavailable.

helix.pages.2_Data_Preprocessing module

helix.pages.2_Data_Preprocessing.build_config() → PreprocessingOptions

Build the configuration object for preprocessing.

helix.pages.2_Data_Preprocessing.run_preprocessing_pipeline(data, config, experiment_dir: Path, data_opts, path_to_data_opts, path_to_preproc_opts, logger) → None

Run the preprocessing pipeline and save results.

Parameters:

  • data – Input data to preprocess

  • config – Preprocessing configuration

  • experiment_dir – Path to experiment directory

  • data_opts – Data options

  • path_to_data_opts – Path to data options file

  • path_to_preproc_opts – Path to preprocessing options file

  • logger – Logger instance

helix.pages.2_Data_Preprocessing.validate_data(data) → tuple[list, bool]

Validate data for preprocessing.

Parameters:

data – The input data to validate

Returns:

tuple containing list of non-numeric columns and whether y has non-numeric values

helix.pages.3_Data_Visualisation module

helix.pages.3_Data_Visualisation.load_dataset(path_to_raw_data: Path, path_to_preproc_data: Path, logger) → tuple

Load raw and preprocessed data if available.

Parameters:

  • path_to_raw_data – Path to raw data file

  • path_to_preproc_data –

    Path to preprocessed data file

    logger: Logger instance

Returns:

(raw_data, preprocessed_data, data_for_tsne)

Return type:

tuple

helix.pages.3_Data_Visualisation.visualisation_view(data, data_tsne, prefix: str | None = None)

Display visualisation of data.

helix.pages.4_Train_Models module

Train Models page for Helix.

This page allows users to configure and train machine learning models on their data.

helix.pages.4_Train_Models.build_configuration() → tuple[MachineLearningOptions, ExecutionOptions, PlottingOptions, DataOptions, str]

Build the configuration options to run the Machine Learning pipeline.

Returns:

The machine learning options, general execution options, plotting options, data options, experiment name.

Return type:

tuple[MachineLearningOptions, ExecutionOptions, PlottingOptions, DataOptions, str]

helix.pages.4_Train_Models.pipeline(ml_opts: MachineLearningOptions, exec_opts: ExecutionOptions, plotting_opts: PlottingOptions, data_opts: DataOptions, experiment_name: str, data: TabularData)

This function actually performs the steps of the pipeline. It can be wrapped in a process it doesn’t block the UI.

Parameters:

helix.pages.5_Feature_Importance module

Feature Importance page for Helix.

This page provides options for analyzing feature importance using various methods: - Global methods (Permutation, SHAP) - Local methods (LIME, SHAP) - Ensemble methods - Fuzzy feature importance

helix.pages.5_Feature_Importance.build_configuration() → tuple[FuzzyOptions | None, FeatureImportanceOptions, ExecutionOptions, PlottingOptions, DataOptions, str, list]

Build the configuration objects for the pipeline.

Returns:

tuple[ FuzzyOptions | None, FeatureImportanceOptions, ExecutionOptions, PlottingOptions, DataOptions, str, list]: - The options for fuzzy, - The options for feature importance - The options for pipeline execution - The plotting options - The data options - The experiment name - The list of models to explain.

helix.pages.5_Feature_Importance.display_feature_importance_plots(experiment_dir: Path) → None

Display feature importance plots from the given experiment directory.

Parameters:

experiment_dir – Path to the experiment directory

helix.pages.5_Feature_Importance.display_fuzzy_plots(experiment_dir: Path) → None

Display fuzzy plots from the given experiment directory.

Parameters:

experiment_dir – Path to the experiment directory

helix.pages.5_Feature_Importance.pipeline(fuzzy_opts: FuzzyOptions, fi_opts: FeatureImportanceOptions, exec_opts: ExecutionOptions, plot_opts: PlottingOptions, experiment_name: str, explain_models: list, data: TabularData)

This function actually performs the steps of the pipeline. It can be wrapped in a process it doesn’t block the UI.

Parameters:

  • fuzzy_opts (FuzzyOptions) – Options for fuzzy feature importance.

  • fi_opts (FeatureImportanceOptions) – Options for feature importance.

  • exec_opts (ExecutionOptions) – Options for pipeline execution.

  • plot_opts (PlottingOptions) – Options for plotting.

  • experiment_name (str) – The experiment name.

  • explain_models (list) – The models to analyse.

  • data (TabularData) – The data that will be used in the pipeline.

helix.pages.6_View_Experiments module

helix.pages.6_View_Experiments.display_experiment_logs(experiment_path: Path) → None

Display all available logs for the experiment.

Parameters:

experiment_path – Path to the experiment directory

helix.pages.6_View_Experiments.display_experiment_plots(experiment_path: Path) → None

Display all available plots for the experiment.

Parameters:

experiment_path – Path to the experiment directory

Module contents