API Reference¶

Flowcean's public API is imported from its packages, not from the top-level flowcean module. Choose a package below for its exported classes, functions, and signatures. For concepts and workflows, start with the user guide.

Foundations¶

• Core: environments, learners, models, metrics, transforms, callbacks, and learning strategies.
• Hybrid systems: system definitions, simulation, trace I/O, and plotting.
• Hybrid benchmarks: reusable systems and registry metadata.
• HyDRA identification: identification and mode-selector APIs.
• Polars: dataframe environments, time-series processing, and transforms.

Learners and models¶

• scikit-learn: regression learners, model wrappers, and evaluation metrics.
• River: incremental learning.
• PyTorch: neural networks and Lightning training.
• PySR: symbolic regression.
• XGBoost: boosted classifiers and regressors.
• PalaestrAI: Soft Actor-Critic learning for active environments.
• AALpy: passive Mealy and Moore automata learning with RPNI.
• Ensemble: combined and cluster-based learners and models.

See installation for optional backend dependencies.

Integrations and tools¶

• Adapters: dataframe and OPC interfaces for model deployment.
• ROS: loading ROS bag data.
• Mosaik: active energy-system environments.
• Testing: model tests, input domains, and predicates.
• Utilities: experiment initialization, random seeds, and prediction loops.