init_opt_at.py

import signac
import json

project = signac.init_project("opt_at_params")

# Set Initial Parameters
Atom_Types = [0, 1, 2, 3, 4, 5, 6, 8, 11]
repeats = 25  # Repeats for full optimization
repeats_ind = 25  # Repeats for individual molecule optimization
lhs_pts = int(1e5)  # Number of LHS points to generate
seed = 1
save_data = True
training_molecules = list(
    ["R14", "R32", "R50", "R170", "R125", "R134a", "R143a", "R41"]
)

if isinstance(training_molecules, list):
    training_molecules_all = json.dumps(training_molecules)
Objective = "ExpVal"

for Atom_Type in Atom_Types:
    # Create job parameter dict
    for i in range(0, repeats):
        sp = {
            "atom_type": Atom_Type,
            "total_repeats": repeats,
            "repeat_number": i + 1,
            "training_molecules": training_molecules_all,
            "num_train_molec": len(training_molecules),
            "obj_choice": Objective,
            "save_data": save_data,
            "new_weight": True,
            "seed": seed,
        }
        if i == 0:
            sp["lhs_pts"] = lhs_pts
        # Create jobs for exploration bias study
        job = project.open_job(sp).init()

    if len(training_molecules) > 1:
        for molec in training_molecules:
            # Make a dumped list of the molecule to pass to the job
            molec_dump = json.dumps(list([molec]))
            for j in range(0, repeats_ind):
                sp = {
                    "atom_type": Atom_Type,
                    "total_repeats": repeats_ind,
                    "repeat_number": j + 1,
                    "training_molecules": molec_dump,
                    "num_train_molec": 1,
                    "obj_choice": Objective,
                    "save_data": save_data,
                    "seed": seed,
                }
                if j == 0:
                    sp["lhs_pts"] = lhs_pts
                job = project.open_job(sp).init()