Mapping of Triangular structures (varipeps.mapping.triangular)

class varipeps.mapping.triangular.Triangular_Expectation_Value(nearest_neighbor_gates: Sequence[Array], normalization_factor: int = 1, is_spiral_peps: bool = False, spiral_unitary_operator: Array | None = None)

Bases: Expectation_Model

Class to calculate expectation values for a mapped triangular PESS structure.

Structure of the triangular lattice with smallest possible unit cell marked by dashed lines.

Structure of the triangular lattice with smallest possible unit cell marked by dashed lines.

Parameters:

  • nearest_neighbor_gates (sequence of jax.numpy.ndarray) – Sequence with the gates that should be applied to each nearest neighbor.

  • downward_triangle_gates (sequence of jax.numpy.ndarray) – Sequence with the gates that should be applied to the downward triangles.

  • normalization_factor (int, optional) – Factor which should be used to normalize the calculated values. If for example three sites are mapped into one PEPS site this should be 1.

    Default: 1

  • is_spiral_peps (bool, optional) – Flag if the expectation value is for a spiral iPEPS ansatz.

    Default: False

  • spiral_unitary_operator (jax.numpy.ndarray, optional) – Operator used to generate unitary for spiral iPEPS ansatz. Required if spiral iPEPS ansatz is used.

    Default: None

__call__(peps_tensors: Sequence[Array], unitcell: PEPS_Unit_Cell, spiral_vectors: Array | Sequence[Array] | None = None, *, normalize_by_size: bool = True, only_unique: bool = True, return_single_gate_results: bool = False) → Array | List[Array]

Calculate the expectation value for PEPS unitcell depending on the gates set in the class.

Parameters:

  • peps_tensors (sequence of jax.numpy.ndarray) – The sequence of unique PEPS tensors in the unitcell.

  • unitcell (PEPS_Unit_Cell) – The PEPS unitcell.

  • spiral_vectors (single or sequence of jax.numpy.ndarray, optional) – If the expectation value is for a spiral iPEPS ansatz, in this argument the wavevectors are expected.

    Default: None

Keyword Arguments:

  • normalize_by_size (bool, optional) – Flag if the expectation value should be normalized by the number of tensors in the unitcell.

    Default: True

  • only_unique (bool, optional) – Flag if the expectation value should be calculated just once for each unique PEPS tensor in the unitcell.

    Default: True

Returns:

The expectation values for all gates. Single tensor if only one gate is applied.

Return type:

jax.numpy.ndarray or list of jax.numpy.ndarray

is_spiral_peps: bool = False
nearest_neighbor_gates: Sequence[Array]
normalization_factor: int = 1
spiral_unitary_operator: Array | None = None
class varipeps.mapping.triangular.Triangular_Map_PESS_To_PEPS(unitcell_structure: Sequence[Sequence[int]], chi: int, max_chi: int | None = None)

Bases: Map_To_PEPS_Model

Map a triangular iPESS unit cell to a iPEPS structure.

The simplex are expected to be in the upper triangle and for the mapping to PEPS they are contracted with the site tensor sitting left down of it.

Convention for physical site tensors: [up left, up right, down, phys]

Convention for simplex tensors: [up, down left, down right]

Parameters:

  • unitcell_structure (sequence of sequence of int or 2d array) – Two dimensional array modeling the structure of the unit cell. For details see the description of PEPS_Unit_Cell.

  • chi (int) – Bond dimension of environment tensors which should be used for the unit cell generated.

__call__(input_tensors: Sequence[Array], *, generate_unitcell: bool = True) → List[Array] | Tuple[List[Array], PEPS_Unit_Cell]

Calculate the PEPS unitcell out of a list of input tensors depending on the original systems.

Parameters:

input_tensors (sequence of jax.numpy.ndarray) – The sequence of input tensors that should be converted.

Keyword Arguments:

generate_unitcell (bool, optional) – Flag if the only the PEPS tensors or additionally the unitcell should be calculated.

Default: True

Returns:

The mapped PEPS tensors and if generate_unitcell = True the unitcell for these tensors.

Return type:

list of jax.numpy.ndarray or tuple of list of jax.numpy.ndarray and PEPS_Unit_Cell

classmethod autosave_wrapper(filename: PathLike, tensors: Array, unitcell: PEPS_Unit_Cell, counter: int | None = None, max_trunc_error_list: float | None = None) → None
chi: int
classmethod load_from_file(path: PathLike, *, return_config: bool = False, return_max_trunc_error_list: bool = False) → Tuple[List[Array], PEPS_Unit_Cell] | Tuple[List[Array], PEPS_Unit_Cell, PEPS_AD_Config]

Load unit cell from a HDF5 file.

This function read the group “triangular_pess” from the file and pass this group to the method load_from_group then.

Parameters:

  • path (os.PathLike) – Path of the HDF5 file.

  • return_config (bool, optional) – Return a config object initialized with the values from the HDF5 files. If no config is stored in the file, just the data is returned. Missing config flags in the file uses the default values from the config object.

    Default: False

Returns:

The tuple with the list of the PESS tensors and the PEPS unitcell is returned. If return_config = True. the config is returned as well.

Return type:

tuple(list(jax.numpy.ndarray), PEPS_Unit_Cell) or tuple(list(jax.numpy.ndarray), PEPS_Unit_Cell, PEPS_AD_Config)

static load_from_group(grp: Group, *, return_config: bool = False) → Tuple[List[Array], PEPS_Unit_Cell] | Tuple[List[Array], PEPS_Unit_Cell, PEPS_AD_Config]

Load the unit cell from a HDF5 group which is be passed to the method.

Parameters:

  • grp (h5py.Group) – HDF5 group object to load the data from.

  • return_config (bool, optional) – Return a config object initialized with the values from the HDF5 files. If no config is stored in the file, just the data is returned. Missing config flags in the file uses the default values from the config object.

    Default: False

Returns:

The tuple with the list of the PESS tensors and the PEPS unitcell is returned. If return_config = True. the config is returned as well.

Return type:

tuple(list(jax.numpy.ndarray), PEPS_Unit_Cell) or tuple(list(jax.numpy.ndarray), PEPS_Unit_Cell, PEPS_AD_Config)

max_chi: int | None = None
classmethod random(structure: Sequence[Sequence[int]], d: int, D: int, chi: int | Sequence[int], dtype: Type[number], max_chi: int, *, seed: int | None = None, destroy_random_state: bool = True) → Tuple[List[Array], T_Triangular_Map_PESS_To_PEPS]
classmethod save_to_file(path: PathLike, tensors: List[Array], unitcell: PEPS_Unit_Cell, *, store_config: bool = True, max_trunc_error_list: List[float] | None = None) → None

Save unit cell to a HDF5 file.

This function creates a single group “triangular_pess” in the file and pass this group to the method save_to_group then.

Parameters:

  • path (os.PathLike) – Path of the new file. Caution: The file will overwritten if existing.

  • store_config (bool, optional) – Store the current values of the global config object into the HDF5 file as attrs of an extra group.

    Default: True

static save_to_group(grp: Group, tensors: List[Array], unitcell: PEPS_Unit_Cell, *, store_config: bool = True) → None

Save unit cell to a HDF5 group which is be passed to the method.

Parameters:

  • grp (h5py.Group) – HDF5 group object to store the data into.

  • store_config (bool, optional) – Store the current values of the global config object into the HDF5 file as attrs of an extra group.

    Default: True

unitcell_structure: Sequence[Sequence[int]]