# augmentations_collector

Python API: `luxonis_ml.data.utils.augmentations_collector`

## Classes

### AugmentationsCollector

Collect augmentation names that were applied to a sample.

#### Methods

##### init

```python
def __init__(augmentations: object, aug_config: Path | list[dict[str, Any]]):
```

##### collect_applied_transform_paths

```python
def collect_applied_transform_paths(transform: Any, parent_path: tuple[str, ...] = ()) -> list[str]:
```

##### flatten_config_augmentation_paths

```python
def flatten_config_augmentation_paths(config: list[dict], parent_path: tuple[str, ...] = ()) -> list[str]:
```

##### get_applied_augmentations

```python
def get_applied_augmentations(self) -> list[str]:
```

##### get_tracked_transforms

```python
def get_tracked_transforms(augmentations: AugmentationsLike) -> list[Any]:
```

##### get_wrapped_transform

```python
def get_wrapped_transform(fn: Callable[..., object]) -> Any:
```

##### load_augmentation_paths

```python
def load_augmentation_paths(aug_config: Path | list[dict[str, Any]]) -> list[str]:
```

##### reset_transform_params

```python
def reset_transform_params(transform: Any):
```

#### Attributes

##### augmentations

Augmentation engine being instrumented.

##### configured_paths

Set of configured augmentation paths, including nested transform paths.

### AugmentationsLike

Protocol for augmentation engines inspected by the collector.

#### Attributes

##### apply

Callable used to apply augmentations.

##### batch_transform

Batch transform callable.

##### custom_transform

Custom transform callable.

##### pixel_transform

Pixel transform callable.

##### resize_transform

Resize transform callable.

##### spatial_transform

Spatial transform callable.