missing dependencies and some small suggestions

[wmpauli]

3 changes:

several dependencies were not included in the requirements.txt

also added code to exit ungracefully, if no input file is found (for coco conversion)

argument parsing to have "source" as script arg

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

📊 Key Changes

• Added pandas, pillow, and opencv-python dependencies in requirements.txt.
• Implemented argparse in run.py to parse command-line arguments.
• Added a warning and early exit in run.py if no JSON files are found in specified directory.

🎯 Purpose & Impact

• 🛠 Including missing dependencies ensures the project setup is smoother, preventing errors during execution.
• 📝 The argument parser allows users to specify the source dataset format via command line, improving usability and flexibility.
• ⚠️ Warning message and exit mechanism help users diagnose issues early if JSON files are missing, enhancing the user experience by avoiding potential confusion.

🌟 Summary

This PR enhances the JSON to YOLO conversion tool's usability by adding essential dependencies and user-friendly command-line argument parsing, plus it aids in early error detection. 🛠️🎛️🚀

[glenn-jocher]

@wmpauli thanks for bringing these issues to our attention. We'll review and include any missing dependencies in the requirements.txt file. Regarding the ungraceful exit, we'll consider adding a feature to handle cases where no input file is found during COCO conversion. Additionally, we'll assess the feasibility of incorporating the "source" argument into the script's argument parsing. Thank you for your suggestions!