Development

The refactored codebase is organized by pipeline phase. Each phase generally contains:

a CLI module, for example p6_cli.py
a runner module, for example p6_runner.py
implementation modules and registries
optional job-submission helpers for cluster modes

Source Layout

streamline/
  p1_data_process/
  p2_impute_scale/
  p3_feature_learning/
  p4_feature_importance/
  p5_feature_selection/
  p6_modeling/
  p7_ensembles/
  p8_summary_statistics/
  p9_compare_datasets/
  p10_replication/
  p11_reporting/
  pipeline/
  utils/

Adding Methods

Prefer the existing registry patterns when adding new components:

Area	Location
Imputers/scalers	`streamline/p2_impute_scale/registry`
Feature learners	`streamline/p3_feature_learning/registry`
Feature importance methods	`streamline/p4_feature_importance/registry`
Feature selectors	`streamline/p5_feature_selection/registry`
Models	`streamline/p6_modeling/models`
Ensembles	`streamline/p7_ensembles/registry`

Tests

The default pytest configuration collects only the current main end-to-end tests. Legacy and phase-level subtests were removed from the maintained v1.0.0 test path so routine testing stays focused on the binary, multiclass, and regression demo pipelines.

Run one main test while developing:

pytest streamline/tests/test_complete_binary.py
pytest streamline/tests/test_complete_multiclass.py
pytest streamline/tests/test_complete_regression.py

Run the default main suite when changing shared phase behavior:

pytest

Run the optional TabPFN smoke test directly when changing TabPFN handling:

pytest streamline/tests/subtests/tabpfn_smoke.py -q -rs

STREAMLINE supports Python 3.10 and newer. The CI pytest workflow exercises the main suite on Python 3.10, 3.11, 3.12, and 3.13.

Documentation

Build the docs locally with:

pip install -r docs/requirements.txt
sphinx-build -b html docs/source docs/build/html

Keep examples synchronized with run_configs/, sample_runcommands.txt, and the notebook parameter names.