ProteinFoldingResult, XYZ files and plotting of folded Proteins for ProteinFoldingProblem.

[MarcDrudis]

Summary
Implementation of the interpret method in ProteinFoldingProblem. ProteinFoldingResult created by this method is now functional and can return the turns encoding the shape of the protein, the Cartesian coordinates of each aminoacid, and generate a plot of the structure.
This PR solves issues #235 , #236 and #313.

Details and comments
Three new classes have been created and are arguments of ProteinFoldingResult.

• ProteinDecoder: Takes the raw bitstring result and converts it into sequences of turns describing the shape of the protein.
• ProteinXYZ: Takes the turn sequences from ProteinDecoder and the peptide from ProteinFoldingResult and generates a xyz file for the protein.
• ProteinPlotter: Takes the xyz information from ProteinXYZ and generates a plot of the protein.

Usecase of how the class would work:

from qiskit_nature.problems.sampling.protein_folding.protein_folding_problem import (
    ProteinFoldingProblem,
)
from qiskit_nature.results.protein_folding_result import ProteinFoldingResult
from qiskit_nature.problems.sampling.protein_folding.interactions.miyazawa_jernigan_interaction import (
    MiyazawaJerniganInteraction,
)
from qiskit_nature.problems.sampling.protein_folding.peptide.peptide import Peptide

from qiskit_nature.problems.sampling.protein_folding.penalty_parameters import PenaltyParameters
from qiskit.utils import algorithm_globals

main_chain = "APRLR"
side_chains =  ["", "", "F", "Y", ""]

algorithm_globals.random_seed = 23
peptide = Peptide(main_chain, side_chains)
mj_interaction = MiyazawaJerniganInteraction()

penalty_back = 10
penalty_chiral = 10
penalty_1 = 10

penalty_terms = PenaltyParameters(penalty_chiral, penalty_back, penalty_1)

protein_folding_problem = ProteinFoldingProblem(peptide, mj_interaction, penalty_terms)
protein_folding_problem.qubit_op()

protein_folding_result = ProteinFoldingResult(protein_folding_problem, "0011011")


#Getting the turn sequences of the protein
print("This are the turns describing the main chain:",protein_folding_result.protein_decoder.get_main_turns())
print("This are the turns describing the side chain:",protein_folding_result.protein_decoder.get_side_turns())

#Generating the xyz file
protein_folding_result.protein_xyz.get_xyz_file( name = "NameForMyFile", output_data = True)

#Plotting the structure
protein_folding_result.plotstructure(ticks = True, grid = False)

This is how the plot looks like:

Note: Right now the xyz file is generated in the working directory.

[dlasecki]

Leaving a partial review.

[woodsp-ibm]

Do you think this is enough for people to understand? Would an example of the string here with some further explanation help? You have more explanation in the tutorial right along with the example there. I just wonder if we are exposing this as part of the public API if things can be understood well enough from just from the docs here.

[MarcDrudis]

I changed the documentation. Now it looks like so:

[MarcDrudis]

So, apparently I had to review the changes in order for my comments to be posted. You might see a lot of comments all of the sudden. Don't be concerned about them and just look at the last ones.

[MarcDrudis]

I changed the documentation. Now it looks like so:

[mrossinek]

Other than that this seems pretty much good to go, once the CI is fixed 👍

[woodsp-ibm]

In regards of the depth shading the Axes 3DPlot has a parameter as to whether to do that or not.

depthshade: Whether or not to shade the scatter markers to give the appearance of depth. Default is True.

Is this what the query was about "I am not sure I understand. The plot needs to be "3D" even if it doesn't rotate since it represents the shape of the protein. How would we turn that off?"

I guess I was wondering if it would be worth on the plotter we have to allow a user to change the value via that logic. We allow grids to be on or off, this could be another parameter that was passed on down. But then I guess we don't expose a way to change colors that its uses either right, so maybe things are fine as they are and if the user wants/needs more they can plot it directly themselves.

[MarcDrudis]

In regards of the depth shading the Axes 3DPlot has a parameter as to whether to do that or not.

depthshade: Whether or not to shade the scatter markers to give the appearance of depth. Default is True.

Is this what the query was about "I am not sure I understand. The plot needs to be "3D" even if it doesn't rotate since it represents the shape of the protein. How would we turn that off?"

I guess I was wondering if it would be worth on the plotter we have to allow a user to change the value via that logic. We allow grids to be on or off, this could be another parameter that was passed on down. But then I guess we don't expose a way to change colors that its uses either right, so maybe things are fine as they are and if the user wants/needs more they can plot it directly themselves.

Well, since we are retuning a matplotlip figure the user can change any of those features of the plot. Same way I have manually rotated the plots in the notebook.

[woodsp-ibm]

Thanks for the work doing this. All looks good to me. I made a very minor last minute change to the reno (an and to a but).