mir-aidj/all-in-one — reverse-engineered prompt
Reverse engineered prompt
Build me a Python music analysis tool that can take one or more song files and figure out the song structure in a useful way. I want it to detect the tempo, beat times, downbeats, and section changes with labels like intro, verse, chorus, bridge, outro, start, and end.
Please make it easy to use from the command line so I can point it at wav files, and mp3 files too if the right audio support is installed, and get one JSON result per track saved in a struct folder by default. I also want a simple Python API so I can call something like an analyze function in code and get back structured results.
If possible, add optional output for visual summaries and audio sonifications so I can inspect what it found. It should use the GPU when available, otherwise fall back to the CPU. Please include sensible defaults, tests, and setup instructions, and check current docs online if you need to fill in details.
Want more depth? Deep Reverse