[franktankbank]

The amount of code changes as seen in a git diff is roughly close to the amount of text changes. So reviewing is scaled linear with the change. In the way your are diffing are there any confounding changes? i.e. I make what seems like a small textual diff but it realigns the whole model of the code base that causes your diff to look like a big chunk of your code base is changed? Maybe that's a good thing? Just thinking out loud, I can't say I have an understanding about what your methodology is.

[conartist6]

The methodology is like a codemod or a lint rule. If you've worked with codemods you usually review a combination of the (small) script that describes the intent behind the changes you want and the (larger) diff of (text) changes that were generated by applying the codemod.