openrpa 0.3.1

Pixel based automation library testbed

Welcome to OpenRPA

Pixel based RPA automation library

OpenRPA is a Robot Framework library that can automate desktop tasks with state of the art AI. OpenRPA is totally free for use in commercial and non-commercial applications. Development is sponsored by Robocorp with a long term agreement.

OpenRPA development started in 2019 and now - 2 years later - we are publishing the results by making all components available as a seamless automation library.

Current components:

• Text detection
• Text recognition
• Image detection
• Controller primitives for mouse and keyboard

Integration roadmap

• Locator query language and NLP
• Extracting and classification of user interface elements

Library is available for

• Windows
• MacOS
• Linux

Keywords

Click Word  pattern
Click Image  filename

Examples of Locator Query Language (LQL)

Type Text  Customer id  12345
Mouse Left Click  Description  To Right
Mouse Left Click  money.png  Close to Submit

Installation

channels:
  - conda-forge
  - pytorch
  - fcakyon
dependencies:
  - python=3.7.5
  - python-mss=6.1.0
  - numpy=1.20.1
  - craft-text-detector=0.3.3
  - tesseract=4.1.1
  - pytesseract=0.3.7
  - opencv=4.5.1
  - regex=2020.11.13
  - pynput=1.7.1
  - pyobjc-core=6.2.2