# C-APIs¶

## spg_get_major_version, spg_get_minor_version, spg_get_micro_version¶

New in version 1.8.3

Version number of spglib is obtained. These three functions return integers that correspond to spglib version [major].[minor].[micro].

## spg_get_error_code and spg_get_error_message¶

New in version 1.9.5

Be careful. These are not thread safe, i.e., only safely usable when calling one spglib function per process.

These functions is used to see roughly why spglib failed.

SpglibError spg_get_error_code(void);

char * spg_get_error_message(SpglibError spglib_error);


The SpglibError type is a enum type as shown below.

typedef enum {
SPGLIB_SUCCESS = 0,
SPGERR_SPACEGROUP_SEARCH_FAILED,
SPGERR_CELL_STANDARDIZATION_FAILED,
SPGERR_SYMMETRY_OPERATION_SEARCH_FAILED,
SPGERR_ATOMS_TOO_CLOSE,
SPGERR_POINTGROUP_NOT_FOUND,
SPGERR_NIGGLI_FAILED,
SPGERR_DELAUNAY_FAILED,
SPGERR_ARRAY_SIZE_SHORTAGE,
SPGERR_NONE,
} SpglibError;


The usage is as follows:

SpglibError error;
error = spg_get_error_code();
printf("%s\n", spg_get_error_message(error));


## spg_get_symmetry¶

This function finds a set of representative symmetry operations for primitive cells or its extension with lattice translations for supercells. 0 is returned if it failed.

int spg_get_symmetry(int rotation[][3][3],
double translation[][3],
const int max_size,
const double lattice[3][3],
const double position[][3],
const int types[],
const int num_atom,
const double symprec);


The operations are stored in rotation and translation. The number of operations is return as the return value. Rotations and translations are given in fractional coordinates, and rotation[i] and translation[i] with same index give a symmetry operations, i.e., these have to be used together.

As an exceptional case, if a supercell has the basis vectors of the lattice that break crsytallographic point group, the crystallographic symmetry operations are searched with this broken symmetry, i.e., at most the crystallographic point group found in this case is the point group of the lattice. For example, this happens for the $$2\times 1\times 1$$ supercell of a conventional cubic unit cell. This may not be understandable in crystallographic sense, but is practically useful treatment for research in computational materials science.

## spg_get_international¶

Space group type is found and returned in international table symbol to symbol and also as a number (return value). 0 is returned if it failed.

int spg_get_international(char symbol[11],
const double lattice[3][3],
const double position[][3],
const int types[],
const int num_atom,
const double symprec);


## spg_get_schoenflies¶

Space group type is found and returned in schoenflies to symbol and also as a number (return value). 0 is returned if it failed.

int spg_get_schoenflies(char symbol[7],
const double lattice[3][3],
const double position[][3],
const int types[],
const int num_atom,
const double symprec);


## spg_standardize_cell¶

The standardized unit cell (see Spglib conventions of standardized unit cell) is generated from an input unit cell structure and its space group type determined about a symmetry search tolerance. Usually to_primitive=0 and no_idealize=0 are recommended to set and this setting results in the same behavior as spg_refine_cell. 0 is returned if it failed.

int spg_standardize_cell(double lattice[3][3],
double position[][3],
int types[],
const int num_atom,
const int to_primitive,
const int no_idealize,
const double symprec);


Number of atoms in the found standardized unit (primitive) cell is returned.

to_primitive=1 is used to create the standardized primitive cell with the transformation matricies shown at Transformation to the primitive cell, otherwise to_primitive=0 must be specified. The found basis vectors and atomic point coordinates and types are overwritten in lattice, position, and types, respectively. Therefore with to_primitive=0, at a maximum four times larger array size for position and types than the those size of the input unit cell is required to store a standardized unit cell with face centring found in the case that the input unit cell is a primitive cell.

no_idealize=0 is used to idealize the lengths and angles of basis vectors with adjusting the positions of atoms to nearest exact positions according to crystal symmetry. However the crystal can be rotated in Cartesian coordinates by the idealization of the basis vectors. no_idealize=1 disables this. The detail of the idealization (no_idealize=0) is written at Idealization of unit cell structure. no_idealize=1 may be useful when we want to leave basis vectors and atomic positions in Cartesianl coordinates fixed.

## spg_find_primitive¶

Behavior is changed. This function is now a shortcut of spg_standardize_cell with to_primitive=1 and no_idealize=0.

A primitive cell is found from an input unit cell. 0 is returned if it failed.

int spg_find_primitive(double lattice[3][3],
double position[][3],
int types[],
const int num_atom,
const double symprec);


lattice, position, and types are overwritten. Number of atoms in the found primitive cell is returned. The crystal can be rotated by this function. To avoid this, please use spg_standardize_cell with to_primitive=1 and no_idealize=1 although the crystal structure is not idealized.

## spg_refine_cell¶

This function exists for backward compatibility since it is same as spg_standardize_cell with to_primitive=0 and no_idealize=0.

The standardized crystal structure is obtained from a non-standard crystal structure which may be slightly distorted within a symmetry recognition tolerance, or whose primitive vectors are differently chosen, etc. 0 is returned if it failed.

int spg_refine_cell(double lattice[3][3],
double position[][3],
int types[],
const int num_atom,
const double symprec);


The calculated standardized lattice and atomic positions overwrites lattice, position, and types. The number of atoms in the standardized unit cell is returned as the return value. When the input unit cell is a primitive cell and is the face centring symmetry, the number of the atoms returned becomes four times large. Since this function does not have any means of checking the array size (memory space) of these variables, the array size (memory space) for position and types should be prepared four times more than those required for the input unit cell in general.

## spg_get_dataset and spg_get_dataset_with_hall_number¶

Changed in version 1.8.1

For an input unit cell structure, symmetry operations of the crystal are searched. Then they are compared with the crsytallographic database and the space group type is determined. The result is returned as the SpglibDataset structure as a dataset. The default choice of setting of basis vectors in spglib is explained in the manuscript found at http://arxiv.org/abs/1506.01455.

The detail of the dataset is given at Spglib dataset.

Dataset corresponding to the space group type in the standard setting is obtained by spg_get_dataset. If this symmetry search fails, NULL is returned in version 1.8.1 or later (spacegroup_number = 0 is returned in the previous versions). In this function, the other crystallographic setting is not obtained.

SpglibDataset * spg_get_dataset(const double lattice[3][3],
const double position[][3],
const int types[],
const int num_atom,
const double symprec);


To specify the other crystallographic choice (setting, origin, axis, or cell choice), spg_get_dataset_with_hall_number is used.

SpglibDataset * spg_get_dataset_with_hall_number(SPGCONST double lattice[3][3],
SPGCONST double position[][3],
const int types[],
const int num_atom,
const int hall_number,
const double symprec)


where hall_number is used to specify the choice. The possible choices and those serial numbers are found at list of space groups (Seto’s web site). The crystal structure has to possess the space-group type of the Hall symbol. If the symmetry search fails or the specified hall_number is not in the list of Hall symbols for the space group type of the crystal structure, spacegroup_number in the SpglibDataset structure is set 0.

Finally, its allocated memory space must be freed by calling spg_free_dataset.

## spg_free_dataset¶

Allocated memoery space of the C-structure of SpglibDataset is freed by calling spg_free_dataset.

void spg_free_dataset(SpglibDataset *dataset);


## spg_get_spacegroup_type¶

Changed at version 1.9.4: Some members are added and the member name ‘setting’ is changed to ‘choice’.

This function allows to directly access to the space-group-type database in spglib (spg_database.c). To specify the space group type with a specific choice, hall_number is used. The definition of hall_number is found at Space group type. number = 0 is returned when it failed.

SpglibSpacegroupType spg_get_spacegroup_type(const int hall_number)


SpglibSpacegroupType structure is as follows:

typedef struct {
int number;
char international_short[11];
char international_full[20];
char international[32];
char schoenflies[7];
char hall_symbol[17];
char choice[6];
char pointgroup_schoenflies[4];
char pointgroup_international[6];
int arithmetic_crystal_class_number;
char arithmetic_crystal_class_symbol[7];
} SpglibSpacegroupType;


## spg_get_symmetry_from_database¶

This function allows to directly access to the space group operations in the spglib database (spg_database.c). To specify the space group type with a specific choice, hall_number is used. The definition of hall_number is found at Space group type. 0 is returned when it failed.

int spg_get_symmetry_from_database(int rotations[192][3][3],
double translations[192][3],
const int hall_number);


The returned value is the number of space group operations. The space group operations are stored in rotations and translations.

## spg_get_multiplicity¶

This function returns exact number of symmetry operations. 0 is returned when it failed.

int spg_get_multiplicity(const double lattice[3][3],
const double position[][3],
const int types[],
const int num_atom,
const double symprec);


This function may be used in advance to allocate memoery space for symmetry operations.

## spg_get_symmetry_with_collinear_spin¶

This function finds symmetry operations with collinear polarizations (spins) on atoms. Except for the argument of const double spins[], the usage is basically the same as spg_get_symmetry, but as an output, equivalent_atoms are obtained. The size of this array is the same of num_atom. See Wyckoff positions and symmetrically equivalent atoms for the definition equivalent_atoms. 0 is returned when it failed.

int spg_get_symmetry_with_collinear_spin(int rotation[][3][3],
double translation[][3],
int equivalent_atoms[],
const int max_size,
SPGCONST double lattice[3][3],
SPGCONST double position[][3],
const int types[],
const double spins[],
const int num_atom,
const double symprec);


## spg_niggli_reduce¶

Niggli reduction is applied to input basis vectors lattice and the reduced basis vectors are overwritten to lattice. 0 is returned if it failed.

int spg_niggli_reduce(double lattice[3][3], const double symprec);


The transformation from original basis vectors $$( \mathbf{a} \; \mathbf{b} \; \mathbf{c} )$$ to final basis vectors $$( \mathbf{a}' \; \mathbf{b}' \; \mathbf{c}' )$$ is achieved by linear combination of basis vectors with integer coefficients without rotating coordinates. Therefore the transformation matrix is obtained by $$\boldsymbol{P} = ( \mathbf{a} \; \mathbf{b} \; \mathbf{c} ) ( \mathbf{a}' \; \mathbf{b}' \; \mathbf{c}' )^{-1}$$ and the matrix elements have to be almost integers.

## spg_delaunay_reduce¶

Delaunay reduction is applied to input basis vectors lattice and the reduced basis vectors are overwritten to lattice. 0 is returned if it failed.

int spg_delaunay_reduce(double lattice[3][3], const double symprec);


The transformation from original basis vectors $$( \mathbf{a} \; \mathbf{b} \; \mathbf{c} )$$ to final basis vectors $$( \mathbf{a}' \; \mathbf{b}' \; \mathbf{c}' )$$ is achieved by linear combination of basis vectors with integer coefficients without rotating coordinates. Therefore the transformation matrix is obtained by $$\boldsymbol{P} = ( \mathbf{a} \; \mathbf{b} \; \mathbf{c} ) ( \mathbf{a}' \; \mathbf{b}' \; \mathbf{c}' )^{-1}$$ and the matrix elements have to be almost integers.

## spg_get_ir_reciprocal_mesh¶

Irreducible reciprocal grid points are searched from uniform mesh grid points specified by mesh and is_shift.

int spg_get_ir_reciprocal_mesh(int grid_address[][3],
int map[],
const int mesh[3],
const int is_shift[3],
const int is_time_reversal,
const double lattice[3][3],
const double position[][3],
const int types[],
const int num_atom,
const double symprec)


mesh stores three integers. Reciprocal primitive vectors are divided by the number stored in mesh with (0,0,0) point centering. The center of grid mesh is shifted +1/2 of a grid spacing along corresponding reciprocal axis by setting 1 to a is_shift element. No grid mesh shift is made if 0 is set for is_shift.

The reducible uniform grid points are returned in fractional coordinates as grid_address. A map between reducible and irreducible points are returned as map as in the indices of grid_address. The number of the irreducible k-points are returned as the return value. The time reversal symmetry is imposed by setting is_time_reversal 1.

Grid points are stored in the order that runs left most element first, e.g. (4x4x4 mesh).:

[[ 0  0  0]
[ 1  0  0]
[ 2  0  0]
[-1  0  0]
[ 0  1  0]
[ 1  1  0]
[ 2  1  0]
[-1  1  0]
....      ]


where the first index runs first. k-qpoints are calculated by (grid_address + is_shift / 2) / mesh. A grid point index is recovered from grid_address by numpy.dot(grid_address % mesh, [1, mesh[0], mesh[0] * mesh[1]]) in Python-numpy notation, where % always returns non-negative integers. The order of grid_address can be changed so that the last index runs first by setting the macro GRID_ORDER_XYZ in kpoint.c. In this case the grid point index is recovered by numpy.dot(grid_address % mesh, [mesh[2] * mesh[1], mesh[2], 1]).

## spg_get_stabilized_reciprocal_mesh¶

The irreducible k-points are searched from unique k-point mesh grids from direct (real space) basis vectors and a set of rotation parts of symmetry operations in direct space with one or multiple stabilizers.

int spg_get_stabilized_reciprocal_mesh(int grid_address[][3],
int map[],
const int mesh[3],
const int is_shift[3],
const int is_time_reversal,
const int num_rot,
const int rotations[][3][3],
const int num_q,
const double qpoints[][3])


The stabilizers are written in fractional coordinates. Number of the stabilizers are given by num_q. Symmetrically equivalent k-points (stars) in fractional coordinates are stored in map as indices of grid_address. The number of reduced k-points with the stabilizers are returned as the return value.

This function can be used to obtain all mesh grid points by setting num_rot = 1, rotations = {{1, 0, 0}, {0, 1, 0}, {0, 0, 1}}, num_q = 1, and qpoints = {0, 0, 0}.

## spg_get_hall_number_from_symmetry¶

experimental

hall_number is obtained from the set of symmetry operations. The definition of hall_number is found at Space group type and the corresponding space-group-type information is obtained through spg_get_spacegroup_type.

This is expected to work well for the set of symmetry operations whose distortion is small. The aim of making this feature is to find space-group-type for the set of symmetry operations given by the other source than spglib. symprec is in the length of the fractional coordinates and should be small like 1e-5.

int spg_get_hall_number_from_symmetry(SPGCONST int rotation[][3][3],
SPGCONST double translation[][3],
const int num_operations,
const double symprec)