update caprieval to handle coarse-grained structures

[rversin]

You are about to submit a new Pull Request. Before continuing make sure you read the contributing guidelines.

Checklist

• Tests added for the new code
• Documentation added for the code changes
• Does not break licensing
• Does not add any dependencies, if it does please add a thorough explanation

Summary of the Pull Request

This pull request corresponds to the adaptation of the caprieval module for coarse-grained structures.

Related Issue

Additional Info

[VGPReys]

Is it really needed to add the cgffversion parameter ? Can't we detect that we are already working with coarse-grain ?

There is an issue with the variable cg set in the init.py:

• It is set to be an integer, but the CAPRI class and other functions are defined receiving bool types.
• Also cg integer is used to point a specific position in lists, and therefore cannot be of type bool.

Instead of lists, I would rather use dicts (in libalign.py):

PROT_ATOMS = [
    PROT_ATOMS_AA,
    PROT_ATOMS_MARTINI2,
    PROT_ATOMS_MARTINI2 # The backbone particle is identical between MARTINI 2 and 3
]

to

PROT_ATOMS = {
    "aa": PROT_ATOMS_AA,
    "martini2": PROT_ATOMS_MARTINI2,
    "martini3": PROT_ATOMS_MARTINI2 # The backbone particle is identical between MARTINI 2 and 3
}

(this also applies for DNA and RNA lists)

and make cg of type str, taking possible values aa, martini2 (and in the future martini3)
I think it will be easier to understand what is happening.

[mgiulini]

this needs some tests to make sure that the resulting numbers make sense

[rversin]

For the cgffversion parameter, I chose this option instead of opening and closing the pdb files in the caprieval module directly, which was the only other solution I could think of to check if it was CG or not. This pdb opening or closing felt tedious, especially because it is tricky to differentiate between martini2 and martini3 structures based on the pdb.

[rversin]

For the variable cg , I indeed made a mistake, at first it was a bool then I changed my mind but didn't change the definition of the variable
I will correct it :)

[VGPReys]

For the cgffversion parameter, I chose this option instead of opening and closing the pdb files in the caprieval module directly, which was the only other solution I could think of to check if it was CG or not. This pdb opening or closing felt tedious, especially because it is tricky to differentiate between martini2 and martini3 structures based on the pdb.

Don't you have access to the type of topology based on the content of the io.json ?

[rversin]

Hi,
If it's fine with everyone, I would like to merge this now :)

[mgiulini]

looks good to me! especially since you compared the results to other methods
Side note : the example silently breaks the traceback option (as now two topology modules can be defined within the same workflow). That will have to be fixed before the CG PR gets merged into the main