kinactive.io module

Files and models’ IO operations.

kinactive.io.load(path: Path) → KinactiveClassifier | KinactiveRegressor | DFGClassifier[source]

Automatically determine which model to load based on bath and load this model.

Parameters:

path – A path to the saved model.

Returns:

The loaded model.

kinactive.io.load_dfg(path: Path = PosixPath('/home/docs/checkouts/readthedocs.org/user_builds/kinactive/checkouts/stable/kinactive/resources/models/DFG_classifier')) → DFGClassifier[source]

Load the kinactive.model.DFGclassifier.

Parameters:

path – A path to the saved model. Must contain four directories, for in, out, other, and meta models.

Returns:

kinactive.io.load_json(path: Path) → dict[str, Any][source]
Parameters:

path – A path to a JSON file.

Returns:

The parsed dictionary.

kinactive.io.load_kinactive(path: Path = PosixPath('/home/docs/checkouts/readthedocs.org/user_builds/kinactive/checkouts/stable/kinactive/resources/models/kinactive_classifier')) → KinactiveClassifier[source]

Load the KinActive model classifying PKs into active/inactive conformations.

Parameters:

path – A path to the saved model.

Returns:

A loaded model.

kinactive.io.load_sklearn(path: Path) → LogisticRegression[source]
Parameters:

path – A path to an sklearn model saved via joblib.save()

Returns:

A model loaded via joblib.load().

kinactive.io.load_txt_lines(path: Path) → list[str][source]
Parameters:

path – A path to a text file.

Returns:

A list of non-empty lines.

kinactive.io.load_xgb(path: Path, xgb_model: _X) → _X[source]
Parameters:

  • path – A path to an XGBoost model saved via save_xgb()

  • xgb_model – The model type.

Returns:

The loaded model.

kinactive.io.save(model: KinactiveClassifier | KinactiveRegressor | LogisticRegression | DFGClassifier, base: Path, name: str, overwrite: bool = False) → Path[source]
Parameters:

  • model – A model from kinactive.model.

  • base – Base dir to save to.

  • name – Model name. Will create base / name dir if it doesn’t exist.

  • overwrite – Overwrite existing model with the same name.

Returns:

The base / name path after successful save.

kinactive.io.save_dfg(model: DFGClassifier, base: Path, name: str, overwrite: bool = False) → Path[source]

Save the DFGclassifier model into four different folders.

This will save each model separately using save() under paths base / name / model_name.

Parameters:

  • model – The model to save.

  • base – Base path to write to.

  • name – A dir name within the base dir.

  • overwrite – Overwrite existing models.

Returns:

base/name path after successful save.

kinactive.io.save_json(data: dict, path: Path) → Path[source]
Parameters:

  • data – Data dictionary.

  • path – A valid path to write to.

Returns:

The path after successful writing.

kinactive.io.save_sklearn(model: LogisticRegression, path: Path) → Path[source]

Save an sklearn model using joblib.dump().

Parameters:

  • model – A scikit-learn model.

  • path – A valid path to write to.

Returns:

The path after successful writing.

kinactive.io.save_txt_lines(lines: Iterable[Any], path: Path) → Path[source]
Parameters:

  • lines – Iterable over any printable elements.

  • path – A valid path to write to.

Returns:

The path after successful writing.

kinactive.io.save_xgb(model: XGBClassifier | XGBRegressor, path: Path) → Path[source]
Parameters:

  • model – XGBoost model with a save_model() method.

  • path – A valid path to write to.

Returns:

The path after successful writing.