File size: 5,392 Bytes
79edee4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
from itertools import product
from pathlib import Path

import numpy as np
import pandas as pd
from dask.distributed import Client
from dask_jobqueue import SLURMCluster
from mlip_arena.models import MLIPEnum
from mlip_arena.tasks import ELASTICITY, OPT, PHONON
from mlip_arena.tasks.optimize import run as OPT
from mlip_arena.tasks.utils import get_calculator
from numpy import linalg as LA
from prefect import flow, task
from prefect_dask import DaskTaskRunner
from tqdm.auto import tqdm

from ase.db import connect

select_models = [
    "ALIGNN",
    "CHGNet",
    "M3GNet",
    "MACE-MP(M)",
    "MACE-MPA",
    "MatterSim",
    "ORBv2",
    "SevenNet",
]

def elastic_tensor_to_voigt(C):
    """
    Convert a rank-4 (3x3x3x3) elastic tensor into a rank-2 (6x6) tensor using Voigt notation.

    Parameters:
    C (numpy.ndarray): A 3x3x3x3 elastic tensor.

    Returns:
    numpy.ndarray: A 6x6 elastic tensor in Voigt notation.
    """
    # voigt_map = {
    #     (0, 0): 0, (1, 1): 1, (2, 2): 2,  # Normal components
    #     (1, 2): 3, (2, 1): 3,  # Shear components
    #     (0, 2): 4, (2, 0): 4,
    #     (0, 1): 5, (1, 0): 5
    # }
    voigt_map = {
        (0, 0): 0,
        (1, 1): 1,
        (2, 2): -1,  # Normal components
        (1, 2): -1,
        (2, 1): -1,  # Shear components
        (0, 2): -1,
        (2, 0): -1,
        (0, 1): 2,
        (1, 0): 2,
    }

    C_voigt = np.zeros((3, 3))

    for i in range(3):
        for j in range(3):
            for k in range(3):
                for l in range(3):
                    alpha = voigt_map[(i, j)]
                    beta = voigt_map[(k, l)]

                    if alpha == -1 or beta == -1:
                        continue

                    factor = 1
                    # if alpha in [3, 4, 5]:
                    if alpha == 2:
                        factor = factor * (2**0.5)
                    if beta == 2:
                        factor = factor * (2**0.5)

                    C_voigt[alpha, beta] = C[i, j, k, l] * factor

    return C_voigt


# -


@task
def run_one(model, row):
    if Path(f"{model.name}.pkl").exists():
        df = pd.read_pickle(f"{model.name}.pkl")

        # if row.key_value_pairs.get('uid', None) in df['uid'].unique():
        #     pass
    else:
        df = pd.DataFrame(columns=["model", "uid", "eigenvalues", "frequencies"])

    atoms = row.toatoms()
    # print(data := row.key_value_pairs)

    calc = get_calculator(model)

    result_opt = OPT(
        atoms,
        calc,
        optimizer="FIRE",
        criterion=dict(fmax=0.05, steps=500),
        symmetry=True,
    )

    atoms = result_opt["atoms"]

    result_elastic = ELASTICITY(
        atoms,
        calc,
        optimizer="FIRE",
        criterion=dict(fmax=0.05, steps=500),
        pre_relax=False,
    )

    elastic_tensor = elastic_tensor_to_voigt(result_elastic["elastic_tensor"])
    eigenvalues, eigenvectors = LA.eig(elastic_tensor)

    outdir = Path(f"{model.name}") / row.key_value_pairs.get(
        "uid", atoms.get_chemical_formula()
    )
    outdir.mkdir(parents=True, exist_ok=True)

    np.savez(outdir / "elastic.npz", tensor=elastic_tensor, eigenvalues=eigenvalues)

    result_phonon = PHONON(
        atoms,
        calc,
        supercell_matrix=(2, 2, 1),
        outdir=outdir,
    )

    frequencies = result_phonon["phonon"].get_frequencies(q=(0, 0, 0))

    new_row = pd.DataFrame(
        [
            {
                "model": model.name,
                "uid": row.key_value_pairs.get("uid", None),
                "eigenvalues": eigenvalues,
                "frequencies": frequencies,
            }
        ]
    )

    df = pd.concat([df, new_row], ignore_index=True)
    df.drop_duplicates(subset=["model", "uid"], keep="last", inplace=True)

    df.to_pickle(f"{model.name}.pkl")


@flow
def run_all():
    import random

    random.seed(0)

    futures = []
    with connect("c2db.db") as db:
        random_indices = random.sample(range(1, len(db) + 1), 1000)
        for row, model in tqdm(
            product(db.select(filter=lambda r: r["id"] in random_indices), MLIPEnum)
        ):
            if model.name not in select_models:
                continue
            future = run_one.submit(model, row)
            futures.append(future)
    return [f.result(raise_on_failure=False) for f in futures]


# +


if __name__ == "__main__":
    nodes_per_alloc = 1
    gpus_per_alloc = 1
    ntasks = 1

    cluster_kwargs = dict(
        cores=1,
        memory="64 GB",
        processes=1,
        shebang="#!/bin/bash",
        account="matgen",
        walltime="00:30:00",
        # job_cpu=128,
        job_mem="0",
        job_script_prologue=[
            "source ~/.bashrc",
            "module load python",
            "source activate /pscratch/sd/c/cyrusyc/.conda/dev",
        ],
        job_directives_skip=["-n", "--cpus-per-task", "-J"],
        job_extra_directives=[
            "-J c2db",
            "-q regular",
            f"-N {nodes_per_alloc}",
            "-C gpu",
            f"-G {gpus_per_alloc}",
        ],
    )

    cluster = SLURMCluster(**cluster_kwargs)
    print(cluster.job_script())
    cluster.adapt(minimum_jobs=25, maximum_jobs=50)
    client = Client(cluster)
    # -

    run_all.with_options(
        task_runner=DaskTaskRunner(address=client.scheduler.address), log_prints=True
    )()