File size: 6,342 Bytes
afe68b4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
from __future__ import annotations

import os
from pathlib import Path

from ase import Atoms, units
from ase.calculators.calculator import BaseCalculator
from dotenv import load_dotenv
from prefect import flow, task
from prefect.cache_policies import INPUTS, TASK_SOURCE
from prefect.futures import wait

from mlip_arena.models import MLIPEnum
from mlip_arena.tasks import MD
from mlip_arena.tasks.utils import get_calculator

from .data import get_atoms_from_db

load_dotenv()

HF_TOKEN = os.environ.get("HF_TOKEN", None)


@task(cache_policy=TASK_SOURCE + INPUTS)
def nvt_heat_one(atoms: Atoms, model: MLIPEnum | BaseCalculator, run_dir: Path):
    """Run a 10 ps NVT MD simulation with linear heating schedule.

    This task invokes the MD task (wrapped via Prefect) to perform an NVT
    simulation using the provided calculator or MLIP model. It is intended
    to probe whether the model remains stable when the system is heated
    from 300 K to 3000 K over a short timeframe.

    Parameters
    - atoms: ASE Atoms object representing the system to simulate. A copy
      is typically submitted by the caller.
    - model: either an MLIPEnum entry (selects a registered model) or an
      already-constructed ASE BaseCalculator.

    Returns
    - The result produced by the MD task. On exception, the exception object
      is returned (the calling flow records and filters results).
    """

    calculator = (
        get_calculator(
            model.name,
            calculator_kwargs=None,
        )
        if isinstance(model, MLIPEnum)
        else model
    )

    model_name = model.name if isinstance(model, MLIPEnum) else model.__class__.__name__

    return MD.with_options(
        # timeout_seconds=600,
        # retries=1,
        refresh_cache=True
    )(
        atoms=atoms,
        # wrap get_calculator in task to dynamically assign GPU device
        calculator=calculator,
        ensemble="nvt",
        dynamics="nose-hoover",
        time_step=None,
        dynamics_kwargs=dict(
            ttime=25 * units.fs,
            # pfactor=((75 * units.fs) ** 2) * 1e2 * units.GPa
        ),
        total_time=1e4,  # 10 ps
        temperature=[300, 3000],
        pressure=None,
        traj_file=run_dir
        / f"{model_name}_{atoms.info.get('material_id', 'random')}_{atoms.get_chemical_formula()}_nvt.traj",
        traj_interval=10,
    )


@task(cache_policy=TASK_SOURCE + INPUTS)
def npt_compress_one(atoms: Atoms, model: MLIPEnum | BaseCalculator, run_dir: Path):
    """Run a 10 ps NPT MD simulation with linear pressure ramp.

    This task invokes the MD task (wrapped via Prefect) to perform an NPT
    simulation where the pressure ramps up to probe structural response and
    potential instabilities under compression.

    Parameters
    - atoms: ASE Atoms object representing the system to simulate.
    - model: either an MLIPEnum entry (selects a registered model) or an
      already-constructed ASE BaseCalculator.

    Returns
    - The result produced by the MD task.
    """

    calculator = (
        get_calculator(
            model.name,
            calculator_kwargs=None,
        )
        if isinstance(model, MLIPEnum)
        else model
    )

    model_name = model.name if isinstance(model, MLIPEnum) else model.__class__.__name__

    return MD.with_options(timeout_seconds=600, retries=2, refresh_cache=True)(
        atoms=atoms,
        calculator=calculator,
        ensemble="npt",
        dynamics="nose-hoover",
        time_step=None,
        dynamics_kwargs=dict(
            ttime=25 * units.fs, pfactor=((75 * units.fs) ** 2) * 1e2 * units.GPa
        ),
        total_time=1e4,  # 5e4, # fs
        temperature=[300, 3000],
        pressure=[0, 5e2 * units.GPa],  # 500 GPa / 10 ps = 50 GPa / 1 ps
        traj_file=run_dir
        / f"{model_name}_{atoms.info.get('material_id', 'random')}_{atoms.get_chemical_formula()}_npt.traj",
        traj_interval=10,
    )


@flow
def heating(
    model: MLIPEnum | BaseCalculator, run_dir: Path, hf_token: str | None = HF_TOKEN
):
    """Prefect flow to run NVT heating tasks for many database structures.

    This flow iterates over structures from the 'random-mixture.db' dataset
    and submits nvt_heat_one tasks for each structure. It waits for all
    submitted futures and returns the list of completed results.

    Parameters
    - model: MLIPEnum or BaseCalculator to use for the simulations.

    Returns
    - A list of results from completed tasks. Failed tasks are filtered out.
    """

    futures = []
    # To download the database automatically, `huggingface_hub login` or provide HF_TOKEN
    for i, atoms in enumerate(
        get_atoms_from_db("random-mixture.db", hf_token=hf_token, force_download=False)
    ):
        if i >= 200:
            break
        future = nvt_heat_one.with_options(
            timeout_seconds=600, retries=2, refresh_cache=False
        ).submit(atoms.copy(), model, run_dir)
        futures.append(future)

    wait(futures)

    return [
        f.result(timeout=None, raise_on_failure=False)
        for f in futures
        if f.state.is_completed()
    ]


@flow
def compression(
    model: MLIPEnum | BaseCalculator, run_dir: Path, hf_token: str | None = HF_TOKEN
):
    """Prefect flow to run NPT compression tasks for many database structures.

    This flow iterates over structures from the 'random-mixture.db' dataset
    and submits npt_compress_one tasks for each structure. It waits for
    completion and returns the list of successful results.

    Parameters
    - model: MLIPEnum or BaseCalculator to use for the simulations.

    Returns
    - A list of results from completed tasks. Failed tasks are filtered out.
    """

    futures = []
    # To download the database automatically, `huggingface_hub login` or provide HF_TOKEN
    for i, atoms in enumerate(
        get_atoms_from_db("random-mixture.db", hf_token=hf_token, force_download=False)
    ):
        if i >= 200:
            break
        future = npt_compress_one.with_options(
            timeout_seconds=600, retries=2, refresh_cache=False
        ).submit(atoms.copy(), model, run_dir)
        futures.append(future)

    wait(futures)

    return [
        f.result(timeout=None, raise_on_failure=False)
        for f in futures
        if f.state.is_completed()
    ]