Core

SpinData object for storing N-dimensional data with coordinates

class spinlab.core.data.SpinData(values=array([], dtype=float64), dims=[], coords=[], attrs={}, spinlab_attrs={'attenuation': 25, 'center_field': 3435.0, 'data_format': 'XEPR', 'data_type': 'EPR', 'frequency': 9625482000.0, 'power': 0.4743, 'pulse_attenuation': 0, 'scans': 1}, proc_attrs=None, **kwargs)

Bases: ABCData

SpinData Class for handling spinlab data

The SpinData class is inspired by pyspecdata nddata object which handles n-dimensional data, axes, and other relevant information together.

This class is designed to handle data and axes together so that performing NMR processing can be performed easily.

values

Numpy Array containing data

Type:: numpy.ndarray

coords

List of numpy arrays containing axes of data

Type:: list

dims

List of axes labels for data

Type:: list

attrs

Dictionary of parameters for data

Type:: dict

add_proc_attrs(proc_attr_name, proc_dict)

Stamp processing step to SpinData object

Parameters:

proc_attr_name (str) -- Name of processing step (e.g. "fourier_transform")
proc_dict (dict) -- Dictionary of processing parameters for this processing step.

exp_info(): Print experiment attributes currently in attrs dictionary

phase()

Return phase of SpinData object

Returns:

phase of data calculated from sum of imaginary: divided by sum of real components

Return type:

phase (float,int)

proc_info(step_name=None): Print processing steps and parameters currently in proc_attrs list

select(selection)

Select subset of 2D data object

Parameters:: selection (int, range, list, tuple) -- list or tuple of slices to keep
Returns:: subset of SpinData object
Return type:: SpinData object

Examples

data.select((1, range(5,10), 15)) # keeps slices: 1, 5, 6, 7, 8, 9, and 15

show_attrs(show_exp_info=False, show_spinlab_info=True, show_proc_info=True): Print experiment attributes, spinlab attributes and processing steps

spinlab_info(): Print parameters currently in used in spinlab

squeeze()

Remove all length 1 dimensions from data

Warning

Axes information is lost

Examples

data.squeeze()

class spinlab.core.base.ABCData(values=array([], dtype=float64), dims=[], coords=[], attrs={}, sl_attrs={}, error=None, **kwargs)

Bases: object

N-Dimensional Data Object

values

Data values in

Type:: numpy.ndarray

dims

List of strings giving dimension labels

Type:: list

coords

Collection of numpy.ndarrays defining the axes

Type:: Coords

attrs

dictionary of parameters

Type:: dict

error

If not None, error for values which are propagated during mathematical operations

Type:: numpy.ndarray

proc_attrs

List of processing steps

Type:: list

property abs

SpinData with absolute part of values

Type:: SpinData

align(b)

Align two data objects for numerical operations

Parameters:: b -- Object to align with self
Returns:: self and b aligned data objects
Return type:: tuple

argmax(dim)

Return value of coord at values maximum for given dim

Parameters:: dim (str) -- Dimension to perform operation along

argmax_index(dim)

Return index of coord at values maximum for given dim

Parameters:: dim (str) -- Dimension to perform operation along

argmin(dim)

Return value of coord at values minimum for given dim

Parameters:: dim (str) -- Dimension to perform operation along

argmin_index(dim)

Return index of coord at values minimum for given dim

Parameters:: dim (str) -- Dimension to perform operation along

chunk(dim, new_dims, new_sizes)

Note

This is a placeholder for a function that's not yet implemented

Parameters:

dim (str) -- Assume that the dimension dim is a direct product of the dimensions given in new_dims, and chunk it out into those new dimensions.
new_dims (list of str) --
The new dimensions to generate. Note that one of the elements of the list can be dim if you like.

It's assumed that the ordering of dim is a direct product given in C-ordering (i.e. the inner dimensions are listed last and the outer dimensions are listed first -- here "inner" means that changes to the index of the inner-most dimension correspond to adjacent positions in memory and/or adjacent indeces in the original dimension that you are chunking)
new_sizes (list of int) -- sizes of the new dimensions`

Returns:

self -- The new nddata object. Note that uniformly ascending or descending coordinates are manipulated in a rational way, e.g. [1,2,3,4,5,6] when chunked to a size of [2,3] will yield coordinates for the two new dimensions: [1,4] and [0,1,2]. Coordinates that are not uniformly ascending or descending will yield and error and must be manually modified by the user.

Return type:

nddata_core

concatenate(b, dim)

Concatenate SpinData objects

Parameters:

b (SpinData) -- Data object to append to current data object
dim (str) -- dimension to concatenate along

copy()

Return deepcopy of sldata object

Returns:: deep copy of data object

cumulative_sum(dim)

Calculate Cumulative sum of sldata object

Returns:: cumulative sum of data object

property dtype

Values type

Type:: type

fold(): Fold 2d data to original ND shape

get_coord(dim)

Return coord corresponding to given dimension name

Parameters:: dim (str) -- Name of dim to retrieve coordinates from
Returns:: array of coordinates
Return type:: numpy.ndarray

property imag

SpinData with imaginary part of values

Type:: SpinData

index(dim)

Find index of given dimension name

Parameters:: dim (str) -- Name of dimension to index
Returns:: Index value of dim
Return type:: int

is_sorted(dim)

Determine if coords corresponding to give dim are sorted in ascending order :param dim: Dimension to check if sorted :type dim: str

Returns:: True if sorted, False otherwise.
Return type:: bool

maximum(dim)

Return max for given dim

Parameters:: dim (str) -- Dimension to take maximum along

merge_attrs(b)

Merge the given dictionaries

Parameters:: b (nddata_core) -- attributes to merge into object

minimum(dim)

Return min for given dim

Parameters:: dim (str) -- Dimension to perform operation along

property ndim

Number of dimensions

Type:: str

new_dim(dim, coord)

Add new dimension with length 1

Parameters:

dim (str) -- Name of new dimension
coord (int, float) -- New coord

property real

SpinData with real part of values

Type:: SpinData

rename(dim, new_name)

Rename dim

Parameters:

dim (str) -- Name of dimension to rename
new_name (str) -- New name for dim

reorder(dims)

Reorder dimensions

Parameters:: dims (list) -- List of strings in new order

property shape

Shape of values

Type:: tuple

property size: Returns values.size. Total number of elements in numpy array.

smoosh(old_dims, new_name): Note

Not yet implemented.

smoosh does the opposite of chunk -- see :func`:~nddata_core.chunk`

sort(dim)

Sort the coords corresponding to the given dim in ascending order

Parameters:: dim (str) -- dimension to sort

sort_dims(): Sort the dimensions

split(dim, new_dim, coord): Split the dimension dim into

squeeze(dim): Remove length 1 axes

sum(dim)

Perform sum down given dimension

Parameters:: dim (str) -- Dimension to perform sum down

unfold(dim)

Unfold ND data to 2d data

Parameters:: dim (str) -- Dimension to make first (length N), all other dimensions unfolded so that values has shape (N x M)

spinlab.core.util.concat(data_list, dim, coord=None, casting='same_kind')

Concatenates list of data objects down another dimension

Parameters:

data_list (list) -- List of SpinData objects to concatenate
dim (str) -- new dimension name
coord -- coords for new dimension

Returns:

concatenated data object

Return type:

data (SpinData)

spinlab.core.util.get_slice(data, dim, slice_index)

Get data slice of SpinData object

Parameters:

data (SpinData) -- Input data object
dim (str) -- Selected dimension
slice_index (int) -- Index of slice to be returned

Returns:

SpinData object with selected slice

Return type:

data (SpinData)

spinlab.core.util.implements_np(np_function): register a numpy function for special handling in SPECIAL_NP_HANDLED

spinlab.core.util.update_axis(data, start_stop, dim=0, new_dims=0, spacing='lin', verbose=False)

Update axis

Update dimensions (dims) and axis (coords) of a slData object. The name of the dims will be replaced with the name giving in new_dims. The variable start_stop defines the values of the new coords. This can be either a tuple (start values, stop value) or a vector with values. If the start and stop value is provided, either a linear axis (spacing = "lin", default) or a logarithmically space (spacing = "log") will be created. The new axis will replace the coords in the sldata object.

The function is currently implemented for 1D objects only.

Parameters:

data (SpinData) -- slData object
start_stop (tuple or vector) -- Coords for new dimension
dim (int) -- Dimension to act on
new_dims (str) -- Name of the new dimension. If None the name will not be changed.
spacing (str) -- "lin" for linear spaced axis or "log" for logarithmically spaced axis

Returns:

concatenated data object

Return type:

data (SpinData)

spinlab.core.ufunc.generate_data(shape)

Generate a SpinData object filled with random normal values.

Parameters:: shape (tuple) -- Shape of the output array. Each element defines the length of one dimension. Dimensions are labeled x0, x1, etc.
Returns:: Data object with random values and integer coordinates.
Return type:: SpinData