Module utilities.azure_ocr_integrator

Function call for Microsoft Azure OCR requests

Classes

class AzureOCRIntegrator (max_retries: int = 10, timeout: int = 60)

Extract text from pdf images via calls to Azure Cognitive Services OCR API.

Class variables

var client : azure.cognitiveservices.vision.computervision._computer_vision_client.ComputerVisionClient | None

var last_op_id : str

var last_read_results : list[azure.cognitiveservices.vision.computervision.models._models_py3.ReadResult]

var max_retries : int

var timeout : int

Methods

def create_client(self, azure_secret_name: str) ‑> None

Create an Azure ComputerVisionClient based on the supplied AWS secret.

If the azure_secret_name parameter is empty or does not point to a valid AWS secret, self.client is set to None and no image text will be extracted.

Args

azure_secret_name

the name of the secret defined in AWS for this facility's subscription key.

Environment Variables: AWS_ACCESS_KEY_ID: the access key ID for an account with access to the secret AWS_SECRET_ACCESS_KEY: the secret access key for an account with access to the secret AWS_SESSION_TOKEN (optional): required if the selected account is not allowed the non-interactive login permission. Blank otherwise. AWS_REGION (optional): defaults to "us-east-1" if not set.

def ocr_pages(self, pdf: bytes, pages: list[int], debug_path: pathlib.Path | None = None) ‑> list[str]

Read the text from supplied pdf page indices.

Args

pdf

bytes of a pdf file

pages

list of pdf page indices to OCR

Returns

list[str]

list of strings containing structured text extracted from each supplied page index.