Logistic_regression.rout outside of the working directory in MCHBP

[MilitaoLucas]

The logistic_regression.rout is not inside the work directory and I can't find an easy way to fix it. I think that it is saved by the API and I could not find in the code how to save it in a custom directory.

[pmbulit]

Hi @Luke2Sky , sorry for not replying earlier. Started looking into it and then got sidetracked. The standard script shows the expected behaviour -saving these files in the REFCODE_# folder- for me but I am wondering if this might be related to you having had to install R separately. Do you have any environment variables set for R? If you open any of the logistic_regression.R files is the path there correct?

[MilitaoLucas]

Oh, for sure it could be that. Should have upgraded my environment to the latest version. Sorry, I did not have the time. Is the R issue fixed in the latest API version?

[MilitaoLucas]

Just tested with 3.0.17 and still have the same problem in the R install (it was reproduced by support). But also the file logistic_regression.Rout was created outside of the installation directory. Are you also using Linux? It can be a platform specific problem. The real problem is that each of these files is specific for each run. I am using the following command:

python multi_component_hydrogen_bond_propensity_report.py -c Paulo/library -d test -f test HXACAN28

The results are generated outside of the working directory and in the same directory as the python file. Furthermore, I would like to emphasize that these files are regression specific, so they change depending on the pair being analyzed. This can look like a nitpick, but it is important for my use case that the regression results are not subscribed.

[pmbulit]

The fix for the R issue with Linux will be out in the next release, so keep an eye for that one. Sorry we couldn't get it out for this release.
I managed to reproduce the error. It comes from R ignoring the R_HOME variable, I'll check if there's any easy fix for this.
The regression results are also saved to the REFCODE_#_regression_results.txt, is there any particular value you want to capture that is not saved there?

[MilitaoLucas]

Oh, no. I just didn't realize they were equal. Will check it and get back to you. Thank you.

[pmbulit]

I've made a bug ticket for this issue as the error is somewhere in the API and not in the MC-HBP script