richardli
commented
5 years ago "B -" is a historical typo in InterVA-4 software. Since we are using directly their probbase file (that we don't maintain), I kept this check here.
The idea of safeguide is great. But the problem is we need to replicate another software that we don't actively maintain (we can only make suggestions to the maintainers), so it's hard to foresee what can go wrong in their updates. Any ideas to make this more efficient is welcome!
https://github.com/verbal-autopsy-software/InterVA4/blob/792fd94fcad1f143f55865e580a4ea0bc233e5bb/InterVA4_1.7/R/InterVA.R#L224-L225
B- is the only field where we have spaces introduced; is this an artifact of the conditional probabilities file? Since these probabilities are included with the package, could we just fix the issue in the main file and then make this code straightforward? Alternatively, using regular expressions or something akin to that to remove all whitespace across all variables might be a more programmatically-consistent way to do this operation, while also building in a safeguard in case this issue happens for other fields in future iterations.