Add warning for quantized variables #286
Conversation
…ariable in question is stored as a quantized integer, but loaded as a float, which will cause the calculation of the bitinformation to yield bogus results depending on the data.
for more information, see https://pre-commit.ci
|
Hi @JoelJaeschke, |
|
Hi @observingClouds, I did not run the test suite fully, so CI caught some stuff I failed to account for. I have pushed a new version that is more defensive in the way it does its checks. There are still two errors in the test I added when calling the Julia implementation, but I feel like those may be upstream errors? I don't have too much experience with how the code works yet, so forgive me if I have introduced those. Best |
|
I fixed some upstream issues that were caused by the usage of integers. The remaining failure of the CI can be traced back to milankl/BitInformation.jl#50. You were "lucky" and hit some cases we haven't tested for before. We need to address these and can merge your contribution afterwards (hoping that no other issues occur). Could you, @JoelJaeschke, in the meantime add a short description of your PR to the CHANGELOG.md? Thanks! |
|
@JoelJaeschke it took a few iterations but with the upstream changes that I merged into this branch the tests now succeed 😄 |
|
Congrats @JoelJaeschke on your first contribution to xbitinfo and welcome to the community 🎉 Thank you very much for your time and effort, looking forward to any future inputs! |
This lets a user know that the variable in question is stored as a quantized integer, but loaded as a float, which will cause the calculation of the bit-information to yield bogus results depending on the data. Fixes #202