Vision

Implement api to access google vision API

class pysuite.vision.Vision(service: google.cloud.vision_v1.ImageAnnotatorClient)[source]

Bases: object

Class to interact with Google Vision API.

Parameters

service – an authorized Google Vision service client.

add_request(image_path: Union[str, pathlib.PosixPath], methods: Union[List[str], str])[source]

Add a request to annotate a local or GCS image. Multiple annotation methods can be added at the same time. The request will not be immediately submitted. This method is useful for batch_annotate_image and async_annotate_image.

Example

>>> vision.add_request("/my/image.png", methods=["LABEL_DETECTION", "test_detection"])
Example

>>> vision.add_request("gs://my_gcs_bucket/image/path.jpg", methods=["LABEL_DETECTION", "test_detection"])
Parameters

  • image_path – Local or GCS path to the image file.

  • methods – A list of strings representing supported annotation methods. Please view google.cloud.vision_v1.types.Feature.Type for all supported methods. They are case-insensitive.

Returns

None

annotate_image(image_path: Union[str, pathlib.PosixPath], methods: Union[List[str], str]) → google.cloud.vision_v1.types.image_annotator.AnnotateImageResponse[source]

Submit a request to annotate a local image using specified methods.

Parameters

  • image_path – Path to the image file.

  • methods – A list of strings representing supported annotation methods. Please view google.cloud.vision_v1.types.Feature.Type for all supported methods. They are case-insensitive.

Returns

An AnnotateImageResponse object with annotated content.

async_annotate_image(output_gcs_uri: str, batch_size: int = 0) → Optional[google.api_core.operation.Operation][source]

Annotate images asynchronously and place output in Google Cloud Storage in batches.

Parameters

  • output_gcs_uri – Target output location on Google Cloud Storage.

  • batch_size – Maximum number request processed in each batch. If there are 10 requests submitted together, and batch_size is 2, there will be 5 output files (batches) created. Default is 0. Meaning only create one batch.

Returns

An Operation object. You can use it to wait until the process is completed.

batch_annotate_image() → Optional[google.cloud.vision_v1.types.image_annotator.BatchAnnotateImagesResponse][source]

Submit the prepared requests to annotate images and return a response with annotated content. You must first call add_request to prepare the configurations. If no configurations were prepared, this method will return None.

Returns

An BatchAnnotateImagesResponse object with annotated content. Or None if no requests were prepared.

static load_image(image_path: Union[str, pathlib.PosixPath]) → google.cloud.vision_v1.types.image_annotator.Image[source]

Load a local image as Image class that can be used to submit image annotation requests.

Parameters

image_path – Path to the image file.

Returns

Loaded Image object from the target file

static to_json(response: Union[google.cloud.vision_v1.types.image_annotator.AnnotateImageResponse, google.cloud.vision_v1.types.image_annotator.BatchAnnotateImagesResponse]) → dict[source]

Convert possible image responses to dictionary. If it’s not a supported response, a type error will be raised.

Parameters

response – A response returned from annotate_image or batch_annotate_image.

Returns

A dictionary containing annotated contents.