Tarsier SOP

Appearance

Install ​

Requirements ​

  • Python 3.11 or 3.12
  • An OCR service credential: Google Cloud Vision service account JSON or Microsoft Azure Computer Vision API key + endpoint

Install the package ​

shell
pip install tarsier

This installs Tarsier along with its core dependencies: playwright, selenium, google-cloud-vision, and azure-ai-vision-imageanalysis.

Install Playwright browsers (if using Playwright) ​

After installing the package, install the Chromium browser binary:

shell
playwright install chromium

Or install all browsers:

shell
playwright install

OCR credential setup ​

Google Cloud Vision ​

Create a service account key by following the instructions at https://stackoverflow.com/a/46290808/1780891. Download the JSON key file and pass the loaded dict to GoogleVisionOCRService.

Microsoft Azure Computer Vision ​

Create a Computer Vision resource in the Azure portal and copy the key and endpoint from the "Keys and Endpoint" section. Store them in a JSON file:

json
{
  "key": "<your_azure_api_key>",
  "endpoint": "<your_azure_endpoint_url>"
}

Pass the loaded dict to MicrosoftAzureOCRService.

Version installed ​

Latest release: 0.8.2

PyPI: https://pypi.org/project/tarsier/