automationpractice-tests

How to configure OpenCV on M1 Macbook

1. Install OpenCV as described in the article https://caffeinedev.medium.com/building-and-installing-opencv-on-m1-macbook-c4654b10c188

NOTE: you can face several issues in Step #5 while running make -j8 command. To fix the issues, you have to go back to previous step - cmake command and adjust some environment variables. So if you see errors similar to below, navigate to your unzipped OpenCV build directory (/Users/user/test_opencv/opencv-4.5.0/build in my case), clear this directory, add missed environment variable and perform cmake and then make -j8 commands.

• "path to Python is not set" - it's necessary to define PYTHON2_EXECUTABLE together with PYTHON3_EXECUTABLE variable (even if BUILD_opencv_python2 is set to OFF);
• "opencv error: no template named 'integer_sequence' in namespace 'std'; did you mean '__integer_sequence'?" - add CMAKE_CXX_STANDARD=14 variable;
• "opencv error: use of undeclared identifier 'CODEC_ID_H264'; did you mean 'AV_CODEC_ID_H264'?" - the issue with FFMPEG, so the easiest way to turn off it on cmake stage: WITH_FFMPEG=OFF.

Updated cmake command in my case (make sure you have own paths in OPENCV_EXTRA_MODULES_PATH, PYTHON3_EXECUTABLE, PYTHON2_EXECUTABLE):

cmake \                                                                
-DCMAKE_SYSTEM_PROCESSOR=arm64 \
-DCMAKE_OSX_ARCHITECTURES=arm64 \
-DWITH_OPENJPEG=OFF \
-DWITH_IPP=OFF \
-D CMAKE_BUILD_TYPE=RELEASE \
-D CMAKE_INSTALL_PREFIX=/usr/local \
-D OPENCV_EXTRA_MODULES_PATH=/Users/user/test_opencv/opencv_contrib-4.5.0/modules \
-D PYTHON3_EXECUTABLE=/Users/user/miniforge3/envs/testcv/bin/python3 \
-D PYTHON2_EXECUTABLE=/Users/user/miniforge3/envs/testcv/bin/python \
-D BUILD_opencv_python3=ON \
-D BUILD_opencv_python2=OFF \
-D CMAKE_CXX_STANDARD=14 \
-D WITH_FFMPEG=OFF \
-D INSTALL_PYTHON_EXAMPLES=ON \
-D INSTALL_C_EXAMPLES=OFF \
-D OPENCV_ENABLE_NONFREE=ON \
-D BUILD_EXAMPLES=ON ..

Now make -j8 command should be completed successfully. Perform the rest of the commands from the article.

2. If you tested successfully installed OpenCV as described in Step #6, you can proceed with writing autotests. Make sure you activate your conda environment with:

conda activate testcv

3. [For Appium Server 2.0+] In order to support image comparison / find by image, you must install Appium Images plugin:

   appium plugin install images

and activate the plugin while running Appium server:

   appium --use-plugins=images

More information: https://github.com/appium-boneyard/appium-plugins/tree/master/packages/images

4. Now you are ready to write and execute your tests:

• put templates of images in .png format to src/main/resources
• define locator using:

  @ExtendedFindBy(image = "images/ios/cart.png")
  private ExtendedWebElement cartByImageButton;

• execute your test

NOTE: if test fails with NoSuchElementException (especially for iOS), try following:

• make image DPI is smaller (it can be done using default Preview tool in Mac). In my case, test starts working with 100 pixels instead of 144;

OR

• update imageMatchThreshold setting (refer https://appium.io/docs/en/advanced-concepts/image-elements/ for more information)