search

ReliaSolve / Molprobity2

0 stars 0 forks source link

See Vincent's test program as an example of how to interact with the model/pdb hierarchy in cctbx. #4

Open russell-taylor opened 4 years ago

vbchen commented 4 years ago

Just wanted to double-check, do you want me to keep working on this testing code or do you want to take it over?

russell-taylor commented 4 years ago

That's a decision for Dave and Jane. I figured we'd talk about it at the next meeting. One of the topics is how we do comparisons (exact, "close"), which this will make an excellent case study for...

russell-taylor commented 4 years ago

Email from Vincent: It makes use of a concept in cctbx known as bond_proxies, which contains all the information about bond lengths of a model. One of the first things I thought to test for was to make sure the lengths of hydrogens from cctbx and the lengths of hydrogens generated in reduce are the same. Basically it uses the bond_proxies to get those lengths for the different models and compares them. But when I was testing it out I discovered the issue with the hydrogen names being different, as I mentioned on the zoom meeting. To try to test for this, I started writing the function test_model_is_version_3, but then I remembered that our old "remediator" was available in cctbx (for converting from PDB v2.3 to v3.0 format). However, I discovered issues with the remediator code as well so I'm in the midst of fixing it up now. But once the remediator code is working then we should be able to use that tool to check the hydrogen names of a model, so we won't need the test_model_is_version_3 code any more.