focal.mlflow_utils module

focal.mlflow_utils.log_classifier_test_results(tester: Any, run_name: str, model_path: str, dataset_path: str, confusion_matrix_path: str, classification_path: str, roc_path: str, true_labels: List[int], pred_labels: List[int], predictions: List[float], image_only: bool | None = False) → None

Logs summary to mlflow for testing cnn.

Parameters:

• tester (Any) – TestCNN class instance

• run_name (str) – name of the mlflow run

• model_path (str) – path to the cnn model

• dataset_path (str) – path to the csv dataset

• confusion_matrix_path (str) – path to save confusion matrix

• classification_path (str) – path to save classification report

• roc_path (str) – path to save roc plot

• true_labels (List[int]) – true prediction labels

• pred_labels (List[int]) – prediction results

• predictions (List[float]) – float values of predictions

• image_only (Optional[bool], optional) – if you want to test on only images

focal.mlflow_utils.log_cnn_hyperparameter(config: dict, best_hp: dict, run_name: str, experiment_name: str | None = 'cnn_hyperparameter') → None

Logs cnn hyperparameter search to mlflow.

Parameters:

• config (json file) – configuration file with parameters

• best_hp (Dict) – best hyperparameters from search

• run_name (str) – name of mlflow run

• experiment_name (str, optional) – Defaults to “cnn_hyperparameter”.

focal.mlflow_utils.log_cnn_training_run(config: dict, model: Model, history: Any, dataset_path: str, artifacts: Dict[str, str | None] | None = None) → None

Logs training run details for CNN models using MLflow.

Parameters:

• config – Config object with hyperparameters.

• model – Trained TensorFlow model.

• history – Training history from model.fit().

• artifacts (dict) – Optional dict with keys like “model”, “history”.

focal.mlflow_utils.log_image_training_run(config: dict, model: Model, history: Any, dataset_path: str, artifacts: Dict[str, str | None] | None = None) → None

Logs training run details for CNN models (image_only) using MLflow.

Parameters:

• config – Config object with hyperparameters.

• model – Trained TensorFlow model.

• history – Training history from model.fit().

• artifacts (dict) – Optional dict with keys like “model”, “history”.

focal.mlflow_utils.log_mlp_hyperparameter(config: dict, best_hp: dict, run_name: str, experiment_name: str | None = 'mlp_hyperparameter') → None

Logs mlp hyperparamter results to mlflow.

Parameters:

• config (Any) – json config file

• best_hp (Any) – best hyperparameters from search

• run_name (str) – name of mlflow run

• experiment_name (Optional[str], optional) – Defaults to “mlp_hyperparameter”.

focal.mlflow_utils.log_mlp_training_run(config: dict, model: Model, history: Any, dataset_path: str, artifacts: Dict[str, str | None] | None = None) → None

Logs training run details for MLP models using MLflow.

Parameters:

• config – Config object with hyperparameters.

• model – Trained TensorFlow model.

• history – Training history from model.fit().

• artifacts (dict) – Optional dict with keys like “model”, “history”.

focal.mlflow_utils.log_regressor_test_results(model_path: str, run_name: str, dataset_path: str, experiment_name: str, tensions: List[float], predicted_delta: List[float], predictions: List[float], true_delta: List[float], mean: float) → None

Log mlp or xgb regression results to mlflow.

Parameters:

• model_path (str) – path to the trained model

• run_name (str) – name of mlflow run

• dataset_path (str) – path to the csv dataset

• experiment_name (str) – name of mlflow experiment

• tensions (List[float]) – list of current tensions

• predicted_delta (List[float]) – list of predicted change in tensions

• predictions (List[float]) – list of predicted absolute tensions

• true_delta (List[float]) – list of true delta in tension

Raises:

Exception – experiment failed to be created

focal.mlflow_utils.log_rl_test(config: dict, run_name: str, info: List[Dict[str, Any]], experiment_name: str | None = 'rl_testing') → None

Logs reinforcement learning test to mlflow.

Parameters:

• config (json file) – configuration file with parameters

• run_name (str) – name of mlflow run

• info (Dict) – dictionary of saved info from training rl

• experiment_name (str, optional) – Defaults to “rl_testing”.

focal.mlflow_utils.log_xgb_training_run(config: dict, model: Model, X_train: Any, y_train: Any, dataset_path: str, evals_result: Dict[str, str | None] | None = None, artifacts: Dict[str, str | None] | None = None) → None

Logs training run details for CNN models using MLflow.

Parameters:

• config – Config object with hyperparameters.

• model – Trained TensorFlow model.

• history – Training history from model.fit().

• artifacts (dict) – Optional dict with keys like “model”, “history”.