Use own hyperparameters

This example demonstrates how to use custom hyperparameters with Octopus. The key difference from the basic example is the use of the hyperparameters parameter in the Octo configuration, where you can define custom hyperparameter ranges for each model using the Hyperparameter class.

Instead of letting Optuna automatically search the hyperparameter space, you can define your own hyperparameter ranges for the models. We will use the diabetes dataset for this purpose.

### Necessary imports for this example
import os

from octopus.example_data import load_diabetes_data
from octopus.models.hyperparameter import IntHyperparameter
from octopus.modules import Octo
from octopus.study import OctoRegression
from octopus.types import ModelName

### Load the diabetes dataset
df, features, targets = load_diabetes_data()

print("Dataset info:")
print(f"  Features: {len(features)} - {features}")
print(f"  Samples: {df.shape[0]}")
print(f"  Classes: {len(targets)} - {targets}")
print(f"  Target distribution: {df['target'].value_counts().sort_index().to_dict()}")

### Create and run OctoRegression with custom hyperparameters
study = OctoRegression(
    name="use_own_hyperparameters_example",
    path=os.environ.get("STUDIES_PATH", "./studies"),
    target_metric="MAE",
    feature_cols=features,
    target_col="target",
    sample_id_col="index",
    ignore_data_health_warning=True,
    outer_parallelization=False,
    run_single_outersplit_num=0,
    workflow=[
        Octo(
            task_id=0,
            models=[ModelName.RandomForestRegressor],
            n_trials=3,
            hyperparameters={
                ModelName.RandomForestRegressor: [
                    IntHyperparameter(name="max_depth", low=2, high=32),
                    IntHyperparameter(name="min_samples_split", low=2, high=100),
                ]
            },
        ),
    ],
)

study.fit(data=df)

print("Workflow completed")