# `sklearn_wrapper`

Convert a trained scikit-learn model into a callable compatible with URANIE Relauncher.

```python
uranie_evaluator(model_sklearn)
```

**Parameters:**

```{list-table}
:header-rows: 1
:widths: 18 20 62

* - Parameter
  - Type
  - Description
* - `model_sklearn`
  - `sklearn.base.BaseEstimator`
  - A trained scikit-learn estimator (e.g., `RandomForestRegressor`, `LogisticRegression`, `SVC`). Must have been fitted using `.fit()` method.
```

**Returns:**

| Type | Description |
|------|-------------|
| `callable` or `None` | A Python callable that accepts `*args` (floats) and returns a 1-D list of floats. Returns `None` if the model is not trained or arguments are invalid. |

**Exceptions/Errors:**

- Returns `None` and prints error message if `model_sklearn` is not a valid scikit-learn estimator
- Returns `None` and prints error message if model has not been fitted (checked via `check_is_fitted`)

## Examples

**Regression example:**

```python
from sklearn.ensemble import RandomForestRegressor
from sklearn.datasets import load_diabetes
from sklearn.model_selection import train_test_split

from uratools.sklearn_wrapper import uranie_evaluator

# Load data
X, y = load_diabetes(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Train model
model = RandomForestRegressor(n_estimators=100, random_state=42)
model.fit(X_train, y_train)

# Wrap for URANIE
wrapped = uranie_evaluator(model)

if wrapped:
    # Test with first sample
    result = wrapped(*X_test[0])
    print(result)  # [predicted_value]
```

**Classification example:**

```python
from sklearn.linear_model import LogisticRegression
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split

from uratools.sklearn_wrapper import uranie_evaluator

# Load data
iris = load_iris()
X_train, X_test, y_train, y_test = train_test_split(
    iris.data, iris.target, test_size=0.2, random_state=42
)

# Train model
model = LogisticRegression(max_iter=200, random_state=42)
model.fit(X_train, y_train)

# Wrap for URANIE
wrapped = uranie_evaluator(model)

if wrapped:
    result = wrapped(*X_test[0])
    print(f"Predicted class: {result[0]}")
```

## Implementation Details

- **Input conversion**: Arguments are passed to `.predict()` as a single-row NumPy array.
- **Output**: For regression models, returns the predicted value. For classifiers with `predict_proba`, returns the class prediction (not probabilities).
- **Fitting check**: Model is considered fitted if `check_is_fitted()` returns True (does not raise `NotFittedError`).

## Supported Model Types

Works with all scikit-learn estimators that implement the `.predict()` method and pass `check_is_fitted()`:

- `LinearRegression`, `Ridge`, `Lasso`, `ElasticNet`
- `SVR`, `KNeighborsRegressor`, `RandomForestRegressor`, `GradientBoostingRegressor`
- `LogisticRegression`, `SVC`, `RandomForestClassifier`, `GradientBoostingClassifier`
- `KNeighborsClassifier`, `DecisionTreeClassifier`, `GaussianNB`, `MLPClassifier`
- `KMeans`, `DBSCAN`, `AgglomerativeClustering`, `SpectralClustering`

## Supported Output Types

- Scalar outputs (int, float)
- 1-D arrays, tuples, and dictionaries
- Multi-dimensional tensors
- Using `outputkey` parameter for multi-output models