wasifhaq434701-png/Traffic-Violation-Intelligence — reverse-engineered prompt

Build me a simple Python traffic violation detection app that I can run locally and test with either an image, a video file, or a live camera feed. I want it to spot common road violations like riders without helmets, triple riding on bikes, drivers using a mobile phone, and also identify the vehicle type and try to read the number plate.

When it finds a violation, show the result clearly on the frame with boxes and labels, save a screenshot of the violation, and store the details in a small local database so I can review them later. Make the flow feel clean and practical, with one main entry point so I do not have to guess which file to run. If there is already model code in the project, wire that up instead of reinventing everything.

Please also make sure it works for both images and videos, and keep the setup straightforward with any missing dependency or run instructions cleaned up. If needed, look up current docs online to fill in gaps.