increment_path() robustness improvements

[glenn-jocher]

Improved robustness to filename edge cases like data/images/zidane.23.jpg that currently fail. May resolve #7432

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Optimization of the path incrementation method in the Ultralytics YOLOv5 codebase.

📊 Key Changes

• Introduced a new way to increment file/directory names to prevent conflicts.
• Deprecated the older method that used globbing and regular expression matching.
• The new method iteratively checks for the existence of a file/directory and only breaks the loop when a non-existent path is found.

🎯 Purpose & Impact

• Efficiency Improvement: The replaced method is more straightforward and possibly more efficient as it avoids globbing and regex matching over potentially large directories.
• Code Readability: The new code is more readable, and might reduce the potential for bugs that are difficult to trace.
• User Experience: Users may expect slightly faster execution when dealing with file operations involving numerous similarly named files, especially in directories with a large number of files.