Also include amount of changed files to the calculation to show complexity of pr instead of size only

[HerrDerb]

Also include amount of changed files to the calculation to show complexity of pr instead of size only

[johnlk]

That's an interesting concept. It might require a new parameter like "size_by_complexity: false" which is defaulted to false.

The complexity though is a bit subjective. Would you judge based on the lines added per file relative to the original files size? Or just the total number of files touched in general?

There are code complexity algorithms that a user might assume are being used if this param is enabled such as cognitive complexity which is used by CodeClimate.

[HerrDerb]

Good thoughts. Well as you probably already have seen, I took a simple approach with the amounts of file being touched, as in my perception, it doesn't matter if there is 1 line change or 30 in a single file. What for me seems to increases the complexity is the context switch from file to file. Might be subjective though. I hear that you'd prefer to use the changed line/total line ratio?
Also the naming should be clear that is not about "code complexity" as it is not a static code analysis.

Disclaimer: I'm surly not that common with shell scripts :)