"""
Standardized temperature sweep for the 2D Clock model.
Calculates magnetization, energy, susceptibility, and specific heat, and
produces a thermodynamics figure plus a companion diagnostics figure.
The sweep runs the discrete clock model by default; ``--continuous`` selects
planar spins with the anisotropy ``-A cos(q theta)`` instead.
"""
from __future__ import annotations
import argparse
from scripts.clock._model_choice import (
ORDERED_BELOW_DISCRETE,
add_clock_model_arguments,
resolve_clock_model,
)
from utils.sweep_runner import add_temperature_sweep_arguments, run_temperature_sweep
from utils.system import parse_args_compat
# Approximate crossover temperatures of the q=6 clock model, matching the
# phase boundaries used by scripts/clock/correlation_comparison.py.
_T1_CLOCK6_APPROX = 0.68
_T2_CLOCK6_APPROX = 0.92
# Anisotropy of the continuous model when --continuous is given without
# --aniso; it matches the ClockSimulation constructor default.
_DEFAULT_CONTINUOUS_ANISO = 1.0
[docs]
def main() -> None:
"""
Execute the temperature sweep and generate the thermodynamics and
companion diagnostics figures.
"""
parser = argparse.ArgumentParser(description='2D Clock Model Temperature Sweep')
parser.add_argument('--q', type=int, default=6, help='Number of clock states')
add_clock_model_arguments(parser=parser, default_aniso=_DEFAULT_CONTINUOUS_ANISO)
add_temperature_sweep_arguments(
parser=parser,
size=48,
t_min=0.1,
t_max=2.0,
t_points=60,
# As for XY, the clock model needs a high cap for the convergence
# criterion rather than the cap to end equilibration.
eq_max_steps=200_000,
meas_steps=20_000,
output_dir='results/clock',
transition_help=(
'Transition overlay preset for plotting; auto (default) shows the '
'approximate q=6 crossover temperatures (only for q=6), none '
'disables the overlay'
),
)
args = parse_args_compat(parser=parser)
choice = resolve_clock_model(
parser=parser, args=args, default_aniso=_DEFAULT_CONTINUOUS_ANISO,
)
variant = choice.variant
# Crossover overlay: the approximate T1/T2 values belong to the discrete
# q=6 model, so the overlay is skipped for other q and for the continuous
# model, whose transition temperatures depend on A.
transitions = (
{'T1 (approx)': _T1_CLOCK6_APPROX, 'T2 (approx)': _T2_CLOCK6_APPROX}
if args.transition_preset in ('auto', 'theory') and args.q == 6
and choice.is_discrete
else None
)
run_temperature_sweep(
args=args,
model_cls=choice.model_cls,
model_kwargs=choice.model_kwargs,
model_label=f'{args.q}-state Clock',
plot_title=f'{args.q}-state Clock Temperature Sweep ($L={args.size}$)',
metadata_note=(
f'L={args.size}, q={args.q}, {variant}, target_n_seeds={args.n_seeds}'
),
transition_temperatures=transitions,
variant_note=f'{variant}, ',
# The discrete model orders like Ising below its lowest transition, and
# random starts strand in domain-wall states there, so the ordered start
# replaces a stuck random start. The continuous model gets no such
# preference: its transition temperatures depend on A, and between T1
# and T2 the clock model is only quasi-ordered.
ordered_start_below=(
ORDERED_BELOW_DISCRETE.get(args.q) if choice.is_discrete else None
),
)
if __name__ == '__main__':
main()