swtloc.swtlocalizer.SWTLocalizer

class swtloc.swtlocalizer.SWTLocalizer(multiprocessing: Optional[bool] = False, images: Optional[Union[numpy.ndarray, List[numpy.ndarray]]] = None, image_paths: Optional[Union[str, List[str]]] = None)[source]

Bases: object

SWTLocalizer acts as an entry point for performing Transformations and Localizations.

It creates and houses a list of SWTImage objects in swtimages attribute, after sanity checks have been performed on the input given.

The inputs can be a path (string) to an image file or an numpy array of the image. Inputs can also be just a single image filepath (string) or a single pre-loaded image (np.ndarray) or it could be a list of either image filepath or list of np.ndarray.

But both the parameters i.e image_paths and images cannot be provided. Once the inputs provided to the SWTLocalizer class, sanity checks are performed on the input, and in case of images being provided, random numerical names are assigned to each image(s).

Example:

>>> # Import the SWTLocalizer class
>>> from swtloc import SWTLocalizer
>>> from cv2 import cv2
>>>
>>> root_path = 'examples/images/'
>>>
>>> # Single Image Path (NOTE : Use your own image paths)
>>> single_image_path = root_path+'test_image_1/test_img1.jpg'
>>> swtl = SWTLocalizer(image_paths=single_image_path)
>>>
>>> # Multiple Image Paths (NOTE : Use your own image paths)
>>> multiple_image_paths = [root_path+'test_image_2/test_img2.jpg',
>>>                         root_path+'test_image_3/test_img3.jpg',
>>>                         root_path+'test_image_4/test_img4.jpeg' ]
>>> swtl = SWTLocalizer(image_paths=multiple_image_paths)

>>> # Single Pre-Loaded Image - Agnostic to image channels
>>> single_image = cv2.imread(root_path+'test_image_1/test_img1.jpg')
>>> swtl = SWTLocalizer(images=single_image)

>>> # Multiple Pre-Loaded Image
>>> multiple_images = [cv2.imread(each_path) for each_path in [root_path+'test_image_2/test_img2.jpg',
>>>                                                            root_path+'test_image_3/test_img3.jpg',
>>>                                                            root_path+'test_image_4/test_img4.jpeg' ]]
>>> swtl = SWTLocalizer(images=multiple_images)

>>> # Accessing `SWTImage` objects from the `SWTLocalizer`
>>> multiple_images = [cv2.imread(each_path) for each_path in [root_path+'test_image_2/test_img2.jpg',
>>>                                                            root_path+'test_image_3/test_img3.jpg',
>>>                                                            root_path+'test_image_4/test_img4.jpeg' ]]
>>> swtl = SWTLocalizer(images=multiple_images)
>>> print(swtl.swtimages, type(swtl.swtimages[0]))
[Image-SWTImage_982112, Image-SWTImage_571388, Image-SWTImage_866821] <class 'swtloc.abstractions.SWTImage'>

>>> # Empty Initialisation -> Raises SWTLocalizerValueError (from v2.1.0)
>>> swtl = SWTLocalizer()
SWTLocalizerValueError: Either `images` or `image_paths` parameters should be provided.
>>>
>>>
>>> # Mixed input given  -> Raises SWTLocalizerValueError
>>> mixed_input = [root_path+'test_image_1/test_img1.jpg' , cv2.imread(root_path+'test_image_1/test_img1.jpg')]
>>> swtl = SWTLocalizer(images=mixed_input)
SWTLocalizerValueError: If a list is provided to `images`, each element should be an np.ndarray
>>>
>>> # Wrong input type given  -> Raises SWTLocalizerValueError
>>> wrong_input = [True, 1, 'abc', root_path+'test_image_1/test_img1.jpg']
>>> swtl = SWTLocalizer(image_paths=wrong_input)
SWTLocalizerValueError: `image_paths` should be a `list` of `str`
>>>
>>>
>>> # If the file is not present at the location (NOTE : Use your own image paths) -> Raises FileNotFoundError
>>> multiple_image_paths = [root_path+'test_image_2/test_img2.jpg',
>>>                         root_path+'test_image_/image_not_there.jpg',
>>>                         root_path+'test_image_4/test_img4.jpeg' ]
>>> swtl = SWTLocalizer(image_paths=multiple_image_paths)
FileNotFoundError: No image present at ../swtloc/examples/test_images/image_not_there.jpg

>>> # Random Names being assigned to each image when `images` parameter is provided
>>> multiple_images = [cv2.imread(each_path) for each_path in [root_path+'test_image_2/test_img2.jpg',
>>>                                                            root_path+'test_image_3/test_img3.jpg',
>>>                                                            root_path+'test_image_4/test_img4.jpeg' ]]
>>> swtl = SWTLocalizer(images=multiple_images)
>>> print([each_image.image_name for each_image in swtl.swtimages])
['SWTImage_982112', 'SWTImage_571388', 'SWTImage_866821']

Methods

SWTLocalizer.__init__([multiprocessing, ...])

Create a SWTLocalizer object which will house a list of SWTImage objects in swtimage attribute.