SpinToolkit python documentation

Classes

class spintoolkit_py.mc_base

Bases: pybind11_builtins.pybind11_object

Base class of Monte Carlo sampler.

__init__(self)

Default constructor.

__init__(self, *, seed_global: int, seed_local: int, Ti: float, dump_dir: str = 'dump')

Constructor.

Parameters:

  • seed_global – Global random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • seed_local – Local random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • Ti – Initial temperature.

  • dump_dir – Directory for saving/loading states. (Default: “dump”)

Energy(self: spintoolkit_py.mc_base) → float

Return current energy of the system.

change_beta(self: spintoolkit_py.mc_base, beta: SupportsFloat, *, log: bool = True) → None

Change inverse temperature.

Parameters:

  • beta – The new inverse temperature \(\beta=1/T\).

  • log – If true, the new temperature will be recorded in the file “kT.dat”. (Default: True)

change_dump_dir(self: spintoolkit_py.mc_base, dump_dir: str) → None

Change the directory for saving/loading states.

Parameters:

dump_dir – New directory for saving/loading states.

load_state_SUN(self: spintoolkit_py.mc_base, *, lattice: spintoolkit_py.lattice, Z: spintoolkit_py.VecZList, sweeps_per_dump: int) → tuple[int, int]

Load a list of SU(N) coherent states from the last snapshot in the folder dump_dir.

Parameters:

  • lattice – Underlying lattice.

  • Z – List of SU(N) coherent states to be loaded from file (modified in-place).

  • sweeps_per_dump – Number of sweeps per dump interval.

Returns:

  • info (int) – Whether loading succeeded (0) or not (1).

  • n_sweeps (int) – Sweep number from file.

load_state_dipole(self: spintoolkit_py.mc_base, *, lattice: spintoolkit_py.lattice, s: spintoolkit_py.Vec3List, sweeps_per_dump: int) → tuple[int, int]

Load a list of spin dipoles from the last snapshot in the folder dump_dir.

Parameters:

  • lattice – Underlying lattice.

  • s – List of spin dipoles to be loaded from file (modified in-place).

  • sweeps_per_dump – Number of sweeps per dump interval.

Returns:

  • info (int) – Whether loading succeeded (0) or not (1).

  • n_sweeps (int) – Sweep number from file.

save_state_SUN(self: spintoolkit_py.mc_base, *, lattice: spintoolkit_py.lattice, Z: spintoolkit_py.VecZList, n_sweeps: int, sweeps_per_dump: int) → None

Save a list of SU(N) coherent states to a gzip compressed file.

Parameters:

  • lattice – Underlying lattice.

  • Z – List of SU(N) coherent states to be saved to file.

  • n_sweeps – Current number of sweeps.

  • sweeps_per_dump – Number of sweeps per dump interval.

save_state_dipole(self: spintoolkit_py.mc_base, *, lattice: spintoolkit_py.lattice, s: spintoolkit_py.Vec3List, n_sweeps: int, sweeps_per_dump: int) → None

Save a list of spin dipoles to a gzip compressed file.

Parameters:

  • lattice – Underlying lattice.

  • s – List of spin dipoles to be saved to file.

  • n_sweeps – Current number of sweeps.

  • sweeps_per_dump – Number of sweeps per dump interval.

specific_heat_bootstrap(self: spintoolkit_py.mc_base, *, L: int, Energy_list: collections.abc.Sequence[SupportsFloat], offset: int, n: int = 100, p: int = 100) → tuple[float, float]

Calculate specific heat using bootstrap resampling.

Picks n numbers randomly from Energy_list[offset:] to compute specific heat, and repeats this p times to estimate the mean and error.

Parameters:

  • L – Total number of lattice sites.

  • Energy_list – A list of recorded energies.

  • offset – Number of initial energies to skip (for thermalization).

  • n – Number of samples per bootstrap. (Default: 100)

  • p – Number of bootstrap repetitions. (Default: 100)

Returns:

  • c (float) – Mean specific heat.

  • c_err (float) – Error of the specific heat.

property beta

Inverse temperature \(\beta = 1/T\).

property dump_dir

Folder for saving/loading states.

property ratio_accept

Acceptance ratio.

class spintoolkit_py.mc_Ising

Bases: spintoolkit_py.mc_base

Monte Carlo Sampler for Ising spins.

__init__(self)

Default constructor.

__init__(self, *, seed_global: int, seed_local: int, Ti: float, dump_dir: str = 'dump')

Constructor.

Parameters:

  • seed_global – Global random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • seed_local – Local random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • Ti – Initial temperature.

  • dump_dir – Directory for saving/loading states. (Default: “dump”)

set_random(self: spintoolkit_py.mc_Ising, *, model: spintoolkit_py.model_spin, s: spintoolkit_py.Vec3List) → None

Randomly flip a list of spin dipoles along input Ising axes.

Parameters:

  • model – Underlying model.

  • s – List of spin dipoles to be flipped (normalization \(|\vec{s}_i| = 1\) required and fulfilled) (modified in-place).

update_heatbath(self: spintoolkit_py.mc_Ising, *, model: spintoolkit_py.model_spin, s: spintoolkit_py.Vec3List) → None

Perform L steps of heat-bath updates.

Parameters:

  • model – Underlying model.

  • s – List of spin dipoles to be updated (normalization \(|\vec{s}_i| = 1\) required and fulfilled) (modified in-place).

update_metropolis(self: spintoolkit_py.mc_Ising, *, model: spintoolkit_py.model_spin, s: spintoolkit_py.Vec3List) → None

Perform L steps of Metropolis updates.

Parameters:

  • model – Underlying model.

  • s – List of spin dipoles to be updated (normalization \(|\vec{s}_i| = 1\) required and fulfilled) (modified in-place).

class spintoolkit_py.mc_dipole

Bases: spintoolkit_py.mc_base

Monte Carlo Sampler for spin dipoles.

__init__(self)

Default constructor.

__init__(self, *, seed_global: int, seed_local: int, Ti: float, dump_dir: str = 'dump')

Constructor.

Parameters:

  • seed_global – Global random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • seed_local – Local random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • Ti – Initial temperature.

  • dump_dir – Directory for saving/loading states. (Default: “dump”)

set_random(self: spintoolkit_py.mc_dipole, *, model: spintoolkit_py.model_spin, s: spintoolkit_py.Vec3List) → None

Randomize a list of spin dipoles on the \(S^2\) sphere uniformly.

Parameters:

  • model – Underlying model.

  • s – List of spin dipoles to be randomized (normalization \(|\vec{s}_i| = 1\) fulfilled) (modified in-place).

update_heatbath(self: spintoolkit_py.mc_dipole, *, model: spintoolkit_py.model_spin, s: spintoolkit_py.Vec3List) → None

Perform L steps of heat-bath updates (currently does not support single-ion anisotropy yet).

Parameters:

  • model – Underlying model.

  • s – List of spin dipoles to be updated (normalization \(|\vec{s}_i| = 1\) required and fulfilled) (modified in-place).

update_metropolis(self: spintoolkit_py.mc_dipole, *, model: spintoolkit_py.model_spin, s: spintoolkit_py.Vec3List) → None

Perform L steps of Metropolis updates.

Parameters:

  • model – Underlying model.

  • s – List of spin dipoles to be updated (normalization \(|\vec{s}_i| = 1\) required and fulfilled) (modified in-place).

update_overrelax(self: spintoolkit_py.mc_dipole, *, model: spintoolkit_py.model_spin, s: spintoolkit_py.Vec3List, random_visit: bool = True) → None

Perform L steps of over-relaxation updates (reflect a spin wrt to its local field, i.e., precess by \(\pi\)).

Parameters:

  • model – Underlying model.

  • s – List of spin dipoles to be updated (normalization \(|\vec{s}_i| = 1\) required and fulfilled) (modified in-place).

  • random_visit – If true, visit sites randomly; otherwise, visit sites sequentially. (Default: True)

class spintoolkit_py.mc_SUN

Bases: spintoolkit_py.mc_base

Monte Carlo Sampler for SU(N) coherent states.

__init__(self)

Default constructor.

__init__(self, *, seed_global: int, seed_local: int, Ti: float, dump_dir: str = 'dump')

Constructor.

Parameters:

  • seed_global – Global random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • seed_local – Local random seed. Random sequences are guaranteed to be independent as long as the pair (seed_global, seed_local) is unique.

  • Ti – Initial temperature.

  • dump_dir – Directory for saving/loading states. (Default: “dump”)

set_random(self: spintoolkit_py.mc_SUN, *, model: spintoolkit_py.model_spin, Z: spintoolkit_py.VecZList) → None

Randomize a list of SU(N) coherent states on the \(S^{2N-2}\) sphere uniformly.

Parameters:

  • model – Underlying model.

  • Z – List of SU(N) coherent states to be randomized (normalization \(|\vec{Z}_i|=1\) fulfilled) (modified in-place).

update_metropolis(self: spintoolkit_py.mc_SUN, *, model: spintoolkit_py.model_spin, Z: spintoolkit_py.VecZList) → None

Perform L steps of Metropolis updates.

Parameters:

  • model – Underlying model.

  • Z – List of SU(N) coherent states to be updated (normalization \(|\vec{Z}_i|=1\) required and fulfilled) (modified in-place).