Won't opencv-python and opencv-contrib-python conflict with each other? They are both in requirements!

[sisrfeng]

请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem

• 系统环境/System Environment：
• 版本号/Version：Paddle： PaddleOCR： 问题相关组件/Related components：
• 运行指令/Command Code：
• 完整报错/Complete Error Message：

我们提供了AceIssueSolver来帮助你解答问题，你是否想要它来解答(请填写yes/no)?/We provide AceIssueSolver to solve issues, do you want it? (Please write yes/no):no

请尽量不要包含图片在问题中/Please try to not include the image in the issue.

From google's Bard

You're absolutely right! Having both opencv-python and opencv-contrib-python listed in the requirements.txt of the PaddleOCR GitHub repository can indeed lead to conflicts. Here's why:

1. Redundancy: opencv-contrib-python is essentially an extension of opencv-python, containing additional modules and algorithms. Installing both packages installs many of the same files, potentially causing version clashes and unexpected behavior.
2. Version Mismatch: If the specified versions of opencv-python and opencv-contrib-python differ, pip might not be able to resolve the conflict, leading to installation errors or missing functionalities.
3. Dependency Issues: Other dependencies in the project might have specific requirements for either opencv-python or opencv-contrib-python. Having both listed could lead to unexpected dependency conflicts for those libraries.

Solutions:

There are a couple of ways to address this redundancy and potential conflict:

• Use opencv-contrib-python only: Since it includes all functionalities of opencv-python, using only opencv-contrib-python is sufficient. This simplifies the dependency list and avoids potential conflicts.
• Specify version compatibility: If there are specific reasons for needing both packages, ensure their versions are compatible and explicitly specify them in the requirements.txt using version specifiers (e.g., opencv-python==4.5.5, opencv-contrib-python==4.5.5).
• Consider alternatives: Depending on your specific needs, there might be alternative libraries or functionalities within PaddleOCR that can replace the need for both opencv-python and opencv-contrib-python.

It's generally recommended to avoid having both packages listed unless absolutely necessary. Choosing the appropriate approach depends on your specific requirements and the functionalities you need in your project.

I hope this explanation clarifies the potential conflict and provides helpful solutions!

[sisrfeng]

I guess this is relevant: https://github.com/aleju/imgaug/issues/473

[TingquanGao]

Thanks~ Could you please provide the steps for reproducing the conflicts? We would like to check and fix.

[michaelconan]

This is clearly stated on OpenCV's PyPI page:

There are four different packages (see options 1, 2, 3 and 4 below) and you should SELECT ONLY ONE OF THEM. Do not install multiple different packages in the same environment. There is no plugin architecture: all the packages use the same namespace (cv2). If you installed multiple different packages in the same environment, uninstall them all with pip uninstall and reinstall only one package. a. Packages for standard desktop environments (Windows, macOS, almost any GNU/Linux distribution) Option 1 - Main modules package: pip install opencv-python Option 2 - Full package (contains both main modules and contrib/extra modules): pip install opencv-contrib-python (check contrib/extra modules listing from OpenCV documentation)

I don't believe there should be any need to install both and as mentioned, it causes issues.

We're currently seeing: AttributeError: module 'cv2.ximgproc' has no attribute 'niBlackThreshold'. Going to try to change versions to fix for now.