What is the Right Way to Iteratively Develop a heuristic.py File with Heudiconv?

What I mean by iteration - Let’s say I run heudiconv on a specific subject, getting a .heudiconv folder in my BIDS output. As is the case 95% of the time, let’s also say I don’t get my conversion right on the first try, and some of the mappings in .heudiconv/*/info/*.auto.txt are incorrect.

The issue I’m having is that if I rewrite my heudiconv.py file to change how these mappings are made, and then rerun heudiconv, these .auto.txt files will not update. The --overwrite flag will not change this fact, as it only seems to apply to the actual images in the main BIDS directory.

So what I end up having to do, and have seen others do, is either deleting the .heudiconv directory any subjects I want to rerun, or just creating an entirely new BIDS directory to make sure that the “caching” that heudiconv does with those mappings is not hiding something that should be working.

Is there a better way to iterate over a heudiconv file that I am missing?

If not, what are some solutions you all have come up with?

I also posted this question as an issue in heudiconv’s GitHub here