spectre.file_io.input_parameters

Define the data models for SPEC input parameters using Pydantic. The model uses serializers to generate human-readable files for writing with toml (in toml_io.py), and uses validators to cast incoming data into the format expected by SPEC, which are numpy arrays of much larger sizes.

Module Contents

spectre.file_io.input_parameters.mnvol = 256

spectre.file_io.input_parameters.mmpol = 128

spectre.file_io.input_parameters.mntor = 128

spectre.file_io.input_parameters.mnmax = 33025

spectre.file_io.input_parameters.arrays_of_length_nvol = ['pressure', 'adiabatic', 'mu', 'isurf']

spectre.file_io.input_parameters.arrays_of_length_mvol = ['lrad', 'tflux', 'ivolume', 'pflux', 'helicity', 'pl', 'ql', 'pr', 'qr', 'lp', 'lq', 'rp', 'rq']

spectre.file_io.input_parameters.arrays_of_length_mvol_plus = ['iota', 'oita']

spectre.file_io.input_parameters.arrays_of_length_nvol_minus_1 = []

spectre.file_io.input_parameters.arrays_fourier = ['rbc', 'zbs', 'rbs', 'zbc', 'rwc', 'zws', 'rws', 'zwc', 'vns', 'bns', 'vnc', 'bnc']

class spectre.file_io.input_parameters.PhysicsParameters

Bases: pydantic.BaseModel

igeometry: int

nfp: int = 1

nvol: int = 1

mpol: int = 0

ntor: int = 0

lconstraint: int = -1

mupfits: int = 8

ladiabatic: int = 0

num_modes: int

mupftol: float = 1e-14

rpol: float = 1.0

rtor: float = 1.0

lreflect: int = 0

lbdybnzero: bool = True

lbnszero: bool = False

lfreebound: int = 0

phiedge: float = 1.0

curtor: float = 0.0

curpol: float = 0.0

gamma: float = 0.0

pscale: float = 0.0

enforce_stell_sym: bool = True

lrad: numpy.typing.NDArray[numpy.int32]

pl: numpy.typing.NDArray[numpy.int32]

ql: numpy.typing.NDArray[numpy.int32]

pr: numpy.typing.NDArray[numpy.int32]

qr: numpy.typing.NDArray[numpy.int32]

lp: numpy.typing.NDArray[numpy.int32]

lq: numpy.typing.NDArray[numpy.int32]

rp: numpy.typing.NDArray[numpy.int32]

rq: numpy.typing.NDArray[numpy.int32]

tflux: numpy.typing.NDArray[numpy.float64]

pflux: numpy.typing.NDArray[numpy.float64]

pressure: numpy.typing.NDArray[numpy.float64]

adiabatic: numpy.typing.NDArray[numpy.float64]

mu: numpy.typing.NDArray[numpy.float64]

ivolume: numpy.typing.NDArray[numpy.float64]

iota: numpy.typing.NDArray[numpy.float64]

oita: numpy.typing.NDArray[numpy.float64]

isurf: numpy.typing.NDArray[numpy.float64]

helicity: numpy.typing.NDArray[numpy.float64]

rac: numpy.typing.NDArray[numpy.float64]

zas: numpy.typing.NDArray[numpy.float64]

ras: numpy.typing.NDArray[numpy.float64]

zac: numpy.typing.NDArray[numpy.float64]

rbc: numpy.typing.NDArray[numpy.float64]

zbs: numpy.typing.NDArray[numpy.float64]

rbs: numpy.typing.NDArray[numpy.float64]

zbc: numpy.typing.NDArray[numpy.float64]

rwc: numpy.typing.NDArray[numpy.float64]

zws: numpy.typing.NDArray[numpy.float64]

rws: numpy.typing.NDArray[numpy.float64]

zwc: numpy.typing.NDArray[numpy.float64]

vns: numpy.typing.NDArray[numpy.float64]

bns: numpy.typing.NDArray[numpy.float64]

vnc: numpy.typing.NDArray[numpy.float64]

bnc: numpy.typing.NDArray[numpy.float64]

allrzrz: numpy.typing.NDArray[numpy.float64] = None

mmrzrz: numpy.typing.NDArray[numpy.int32]

nnrzrz: numpy.typing.NDArray[numpy.int32]

model_config

serialize_radial_arrays_nvol(value)

serialize the radial arrays that provide a value for each volume/interface internal to the configuration

serialize_radial_arrays_mvol(value)

serialize the arrays that are not always nvol long but depending on freeboundary can be one longer (are thus defined in the vacuum region enclosing the plasma).

serialize_radial_arrays_mvol_plus(value)

serialize the arrays that are not always nvol long but depending on freeboundary can be one longer (are thus defined in the vacuum region enclosing the plasma).

serialize_radial_arrays_nvol_minus_1(value)

serialize the radial arrays that are only nvol-1 long

serialize_toroidal_arrays(value)

serialize_fourier_arrays(value)

serialize_guess(allrzrz)

The interface guess is stored in allrzrz with index helper arrays mmrzrz and nnrzrz. Our job is to reconstruct a human readable mapping from this.

This helper array has the shape it has because it was dumped at the end of the namelist. In the future this will be deprecated for direct writing of the interface guesses one by one.

classmethod set_guess(values)

prepare the allrzrz array from the input mapping if provided. don’t judge me, this keeps the toml file with a clean, readable guess, and keeps the changes to the interface to the SPEC fortran side minimal. In the future the input_list_mod will (hopefully) be replaced with a interface_mod that passes all relevant information for the SPEC state to the FORTRAN memory, allowing switching between configurations without running all intialization steps…

pad_radial_int_arrays(value)

pad the 1D int arrays to the SPEC expected size

pad_radial_float_arrays_mnvolplus(value)

pad the 1D float arrays to the SPEC expected size

pad_radial_float_arrays_mnvol(value)

pad the 1D float arrays to the SPEC expected size

pad_toroidal_arrays(value)

pad the 1D toroidal arrays to the SPEC expected size

pad_fourier_arrays(value)

pad the 2D fourier arrays to the SPEC expected size.

pad_fourier_arrays_input()

pad the fourier arrays to the spec expected size if they were not suppied as a nested dict (for example if read as a namelist or supplied by a different code). The array is expected to be of shape (ntor_input+1, 2*mpol_input+1) (VMEC convention) with negative m values stored in the upper half of the second axis.

classmethod check_correct_fields_provided(values)

ensure that depending on lconstraint, the necessary info was provided

class spectre.file_io.input_parameters.NumericParameters

Bases: pydantic.BaseModel

linitialize: int = 0

lchangeangle: bool = False

lautoinitbn: int = 1

lzerovac: int = 0

ndiscrete: int = 2

nquad: int = -1

impol: int = -4

intor: int = -4

lsparse: int = 0

lsvdiota: int = 0

lrzaxis: int = 1

ntoraxis: int = 3

lvcgrid: int = 0

class spectre.file_io.input_parameters.GlobalParameters

Bases: pydantic.BaseModel

escale: float = 0.0

opsilon: float = 1.0

pcondense: float = 2.0

epsilon: float = 0.0

wpoloidal: float = 1.0

upsilon: float = 1.0

mfreeits: int = 0

gbntol: float = 1e-06

gbnbld: float = 0.666

vcasingeps: float = 1e-12

vcasingtol: float = 1e-08

vcasingits: int = 8

vcasingper: int = 1

vcnt: int = 256

vcnz: int = 256

class spectre.file_io.input_parameters.DiagnosticsParameters

Bases: pydantic.BaseModel

odetol: float = 1e-07

nppts: int = 0

ppts: float = 0.0

nptrj: int | numpy.typing.NDArray[numpy.int32]

lhevalues: bool = False

lhevectors: bool = False

lhmatrix: bool = False

lerrortype: int = 0

ngrid: int = -1

lcheck: int = 0

ltransform: bool = False

model_config

pad_nptrj(value)

pad the 1D toroidal arrays to the SPEC expected size

serialize_nptrj(value)

Serialize the 1D toroidal arrays to the SPEC expected format

class spectre.file_io.input_parameters.ScreenParameters

Bases: pydantic.BaseModel

wmanual: bool = False

wrzaxis: bool = False

wpackxi: bool = False

wvolume: bool = False

wcompute_coordinates: bool = False

wbasefn: bool = False

wmemory: bool = False

wcompute_metric_quantities: bool = False

wvolume_integrate_chebyshev: bool = False

wmatrix: bool = False

wspsint: bool = False

wsolve_beltrami_system: bool = False

wconstruct_beltrami_field: bool = False

wpackab: bool = False

wcompute_rotational_transform: bool = False

wcompute_plasma_current: bool = False

wdf0ab: bool = False

wlforce: bool = False

wintghs: bool = False

wmtrxhs: bool = False

wlbpol: bool = False

wbroadcast_profile_data: bool = False

wdfp100: bool = False

wdfp200: bool = False

wdforce: bool = False

wnewton: bool = False

wcasing: bool = False

wbnorml: bool = False

wcompute_beltrami_error: bool = False

wconstruct_poincare_plot: bool = False

wpp00ab: bool = False

wget_position_vetor_x: bool = False

whesian: bool = False

wwrite_vector_potentials: bool = False

wnumrec: bool = False

wdcuhre: bool = False

wminpack: bool = False

wiqpack: bool = False

wrksuite: bool = False

wi1mach: bool = False

wd1mach: bool = False

wilut: bool = False

witers: bool = False

wsphdf5: bool = False

winitialize_internal_arrays: bool = False

wglobal: bool = False

wxspech: bool = False

wbuild_vector_potential: bool = False

wreadin: bool = False

wwrtend: bool = False

wmacros: bool = False

class spectre.file_io.input_parameters.MinimizationParameters

Bases: pydantic.BaseModel

gtol: float = 1e-13

xtol: float = 1e-22

ftol: float = 1e-13

max_nfev: int = 10000

max_niter: int = 10000

method: Literal['trf', 'dogbox', 'lm'] = 'dogbox'

rcond_lstsq: float | None = None

save_xin_fixb: bool = True

freeb_max_iter: int = 1

freeb_bnorm_tol: float = 1e-05

nt_dbb: int | None = None

nz_dbb: int | None = None

filter_stellsym: bool = True

objective: Literal['freal', 'freal,len'] = 'freal'

resume: bool = False

mpol_rz: int | None = None

ntor_rz: int | None = None

len_factor: float = 1.0

inface_repr: Literal['fourier', 'fourier red', 'henn'] = 'fourier'

mmax: int = 8

nmax: int = 8

class spectre.file_io.input_parameters.InputParameters

Bases: pydantic.BaseModel

minimization: MinimizationParameters

physics: PhysicsParameters

numeric: NumericParameters

global_: GlobalParameters

diagnostics: DiagnosticsParameters

validate_parameters() → InputParameters

Placeholder for cross-field validation between parameter groups.

classmethod check_henneberg_res(values)

Ensure that resolution used for Henneberg representation is representable (mmax<mpol and nmax<ntor).