Hi WLED community,
I’m trying to simulate an automotive brake-light and flowing indicator lights. I’m using ESP32 and four 30-pixel strips.
I haven’t found a preset that looks right, so am looking to code a user mod script. Thinking ChatGPT assistance might be a worth trying as shortcut to establish basic code. Anyone else tried this?
using it all the time. be aware: it is good a writing basic code and give you hints, it sucks at writing code that needs to be integrated into other code, which is exactly what you are trying to do.
If you do not know how to write C code, you will have a hard time.
Definitely can be done. I think you want to write a custom effect, not a usermod, though. It will also depend on your ability to describe in clear, concise terms exactly what you want. If you just throw out “I want brake lights that look like a Lambo.” - it will fail. But if you can describe the layout of the pixels, how they should light up, timeframes, etc - it would likely work.
I know ‘just enough code to be dangerous’. Chat GPT has helped me design several effects which I could ‘probably’ do on my own …after a few hours of research and tinkering with the code. But GPT gave me workable code in a few minutes and I simply told it what needed to be changed, or in one case, an error, and it completely refactored the code in another minute. Just had GPT open in one window and Visual Studio in the other…generate code, copy, paste, compile, add additional input, generate new code, … It goes really quick!
I did a ‘true’ twinkling C9 effect, a fire effect across multiple segments and the National Lampoons ‘25,000 imported Italian twinkle lights’. All these actually required setting up arrays to store values and adjust the lights based on that. Not just ‘random’ flashing. But after a couple rounds, GPT nailed every one.