Bug and compatibility fixes

[Labib-Bin-Salam]

I cleaned the code by organizing the imports and docstring properly, using PEP 8 naming like snake_case, and replacing hardcoded values with clear constants, which makes the code easier to read and maintain. I also fixed compatibility with newer Gym/Gymnasium-style APIs because reset() and step() can now return extra values like info, terminated, and truncated, so the updated version handles both old and new formats safely. Finally, I removed unused code and kept the frame display logic simple with clear_output(wait=True), which is a common notebook approach for showing animations more smoothly.

[Labib-Bin-Salam]

I renamed variables and functions to clearer PEP 8 names, moved fixed values to uppercase module-level constants, and removed unused imports like numpy.random as rd because PEP 8 favors readable naming and clean imports. I also replaced np.float with float, because np.float was deprecated in NumPy 1.20 and can now raise errors in newer versions. Finally, I kept the multiprocessing setup inside the main guard, used a safer random generator, copied arrays before assigning, and added a zero-check before normalizing the image so the OpenCV display does not crash on division by zero.