"""
Module that contains classes to hold information read from
reduced data files.
A good description of the data members can be found in
the CanSAS 1D XML data format:
http://www.smallangles.net/wgwiki/index.php/cansas1d_documentation
"""
#####################################################################
#This software was developed by the University of Tennessee as part of the
#Distributed Data Analysis of Neutron Scattering Experiments (DANSE)
#project funded by the US National Science Foundation.
#See the license text in license.txt
#copyright 2008, University of Tennessee
######################################################################
from __future__ import print_function
#TODO: Keep track of data manipulation in the 'process' data structure.
#TODO: This module should be independent of plottables. We should write
# an adapter class for plottables when needed.
#from sas.guitools.plottables import Data1D as plottable_1D
from sas.sascalc.data_util.uncertainty import Uncertainty
import numpy as np
import math
[docs]class plottable_1D(object):
"""
Data1D is a place holder for 1D plottables.
"""
# The presence of these should be mutually
# exclusive with the presence of Qdev (dx)
x = None
y = None
dx = None
dy = None
## Slit smearing length
dxl = None
## Slit smearing width
dxw = None
## SESANS specific params (wavelengths for spin echo length calculation)
lam = None
dlam = None
# Units
_xaxis = ''
_xunit = ''
_yaxis = ''
_yunit = ''
def __init__(self, x, y, dx=None, dy=None, dxl=None, dxw=None, lam=None, dlam=None):
self.x = np.asarray(x)
self.y = np.asarray(y)
if dx is not None:
self.dx = np.asarray(dx)
if dy is not None:
self.dy = np.asarray(dy)
if dxl is not None:
self.dxl = np.asarray(dxl)
if dxw is not None:
self.dxw = np.asarray(dxw)
if lam is not None:
self.lam = np.asarray(lam)
if dlam is not None:
self.dlam = np.asarray(dlam)
[docs] def xaxis(self, label, unit):
"""
set the x axis label and unit
"""
self._xaxis = label
self._xunit = unit
[docs] def yaxis(self, label, unit):
"""
set the y axis label and unit
"""
self._yaxis = label
self._yunit = unit
[docs]class plottable_2D(object):
"""
Data2D is a place holder for 2D plottables.
"""
xmin = None
xmax = None
ymin = None
ymax = None
data = None
qx_data = None
qy_data = None
q_data = None
err_data = None
dqx_data = None
dqy_data = None
mask = None
# Units
_xaxis = ''
_xunit = ''
_yaxis = ''
_yunit = ''
_zaxis = ''
_zunit = ''
def __init__(self, data=None, err_data=None, qx_data=None,
qy_data=None, q_data=None, mask=None,
dqx_data=None, dqy_data=None):
self.data = np.asarray(data)
self.qx_data = np.asarray(qx_data)
self.qy_data = np.asarray(qy_data)
self.q_data = np.asarray(q_data)
self.mask = np.asarray(mask)
self.err_data = np.asarray(err_data)
if dqx_data is not None:
self.dqx_data = np.asarray(dqx_data)
if dqy_data is not None:
self.dqy_data = np.asarray(dqy_data)
[docs] def xaxis(self, label, unit):
"""
set the x axis label and unit
"""
self._xaxis = label
self._xunit = unit
[docs] def yaxis(self, label, unit):
"""
set the y axis label and unit
"""
self._yaxis = label
self._yunit = unit
[docs] def zaxis(self, label, unit):
"""
set the z axis label and unit
"""
self._zaxis = label
self._zunit = unit
[docs]class Vector(object):
"""
Vector class to hold multi-dimensional objects
"""
## x component
x = None
## y component
y = None
## z component
z = None
def __init__(self, x=None, y=None, z=None):
"""
Initialization. Components that are not
set a set to None by default.
:param x: x component
:param y: y component
:param z: z component
"""
self.x = x
self.y = y
self.z = z
def __str__(self):
msg = "x = %s\ty = %s\tz = %s" % (str(self.x), str(self.y), str(self.z))
return msg
[docs]class Detector(object):
"""
Class to hold detector information
"""
## Name of the instrument [string]
name = None
## Sample to detector distance [float] [mm]
distance = None
distance_unit = 'mm'
## Offset of this detector position in X, Y,
#(and Z if necessary) [Vector] [mm]
offset = None
offset_unit = 'm'
## Orientation (rotation) of this detector in roll,
# pitch, and yaw [Vector] [degrees]
orientation = None
orientation_unit = 'degree'
## Center of the beam on the detector in X and Y
#(and Z if necessary) [Vector] [mm]
beam_center = None
beam_center_unit = 'mm'
## Pixel size in X, Y, (and Z if necessary) [Vector] [mm]
pixel_size = None
pixel_size_unit = 'mm'
## Slit length of the instrument for this detector.[float] [mm]
slit_length = None
slit_length_unit = 'mm'
def __init__(self):
"""
Initialize class attribute that are objects...
"""
self.offset = Vector()
self.orientation = Vector()
self.beam_center = Vector()
self.pixel_size = Vector()
def __str__(self):
_str = "Detector:\n"
_str += " Name: %s\n" % self.name
_str += " Distance: %s [%s]\n" % \
(str(self.distance), str(self.distance_unit))
_str += " Offset: %s [%s]\n" % \
(str(self.offset), str(self.offset_unit))
_str += " Orientation: %s [%s]\n" % \
(str(self.orientation), str(self.orientation_unit))
_str += " Beam center: %s [%s]\n" % \
(str(self.beam_center), str(self.beam_center_unit))
_str += " Pixel size: %s [%s]\n" % \
(str(self.pixel_size), str(self.pixel_size_unit))
_str += " Slit length: %s [%s]\n" % \
(str(self.slit_length), str(self.slit_length_unit))
return _str
[docs]class Aperture(object):
## Name
name = None
## Type
type = None
## Size name
size_name = None
## Aperture size [Vector]
size = None
size_unit = 'mm'
## Aperture distance [float]
distance = None
distance_unit = 'mm'
def __init__(self):
self.size = Vector()
[docs]class Collimation(object):
"""
Class to hold collimation information
"""
## Name
name = None
## Length [float] [mm]
length = None
length_unit = 'mm'
## Aperture
aperture = None
def __init__(self):
self.aperture = []
def __str__(self):
_str = "Collimation:\n"
_str += " Length: %s [%s]\n" % \
(str(self.length), str(self.length_unit))
for item in self.aperture:
_str += " Aperture size:%s [%s]\n" % \
(str(item.size), str(item.size_unit))
_str += " Aperture_dist:%s [%s]\n" % \
(str(item.distance), str(item.distance_unit))
return _str
[docs]class Source(object):
"""
Class to hold source information
"""
## Name
name = None
## Radiation type [string]
radiation = None
## Beam size name
beam_size_name = None
## Beam size [Vector] [mm]
beam_size = None
beam_size_unit = 'mm'
## Beam shape [string]
beam_shape = None
## Wavelength [float] [Angstrom]
wavelength = None
wavelength_unit = 'A'
## Minimum wavelength [float] [Angstrom]
wavelength_min = None
wavelength_min_unit = 'nm'
## Maximum wavelength [float] [Angstrom]
wavelength_max = None
wavelength_max_unit = 'nm'
## Wavelength spread [float] [Angstrom]
wavelength_spread = None
wavelength_spread_unit = 'percent'
def __init__(self):
self.beam_size = Vector()
def __str__(self):
_str = "Source:\n"
_str += " Radiation: %s\n" % str(self.radiation)
_str += " Shape: %s\n" % str(self.beam_shape)
_str += " Wavelength: %s [%s]\n" % \
(str(self.wavelength), str(self.wavelength_unit))
_str += " Waveln_min: %s [%s]\n" % \
(str(self.wavelength_min), str(self.wavelength_min_unit))
_str += " Waveln_max: %s [%s]\n" % \
(str(self.wavelength_max), str(self.wavelength_max_unit))
_str += " Waveln_spread:%s [%s]\n" % \
(str(self.wavelength_spread), str(self.wavelength_spread_unit))
_str += " Beam_size: %s [%s]\n" % \
(str(self.beam_size), str(self.beam_size_unit))
return _str
"""
Definitions of radiation types
"""
NEUTRON = 'neutron'
XRAY = 'x-ray'
MUON = 'muon'
ELECTRON = 'electron'
[docs]class Sample(object):
"""
Class to hold the sample description
"""
## Short name for sample
name = ''
## ID
ID = ''
## Thickness [float] [mm]
thickness = None
thickness_unit = 'mm'
## Transmission [float] [fraction]
transmission = None
## Temperature [float] [No Default]
temperature = None
temperature_unit = None
## Position [Vector] [mm]
position = None
position_unit = 'mm'
## Orientation [Vector] [degrees]
orientation = None
orientation_unit = 'degree'
## Details
details = None
## SESANS zacceptance
zacceptance = (0,"")
yacceptance = (0,"")
def __init__(self):
self.position = Vector()
self.orientation = Vector()
self.details = []
def __str__(self):
_str = "Sample:\n"
_str += " ID: %s\n" % str(self.ID)
_str += " Transmission: %s\n" % str(self.transmission)
_str += " Thickness: %s [%s]\n" % \
(str(self.thickness), str(self.thickness_unit))
_str += " Temperature: %s [%s]\n" % \
(str(self.temperature), str(self.temperature_unit))
_str += " Position: %s [%s]\n" % \
(str(self.position), str(self.position_unit))
_str += " Orientation: %s [%s]\n" % \
(str(self.orientation), str(self.orientation_unit))
_str += " Details:\n"
for item in self.details:
_str += " %s\n" % item
return _str
[docs]class Process(object):
"""
Class that holds information about the processes
performed on the data.
"""
name = ''
date = ''
description = ''
term = None
notes = None
def __init__(self):
self.term = []
self.notes = []
[docs] def is_empty(self):
"""
Return True if the object is empty
"""
return len(self.name) == 0 and len(self.date) == 0 and len(self.description) == 0 \
and len(self.term) == 0 and len(self.notes) == 0
[docs] def single_line_desc(self):
"""
Return a single line string representing the process
"""
return "%s %s %s" % (self.name, self.date, self.description)
def __str__(self):
_str = "Process:\n"
_str += " Name: %s\n" % self.name
_str += " Date: %s\n" % self.date
_str += " Description: %s\n" % self.description
for item in self.term:
_str += " Term: %s\n" % item
for item in self.notes:
_str += " Note: %s\n" % item
return _str
[docs]class TransmissionSpectrum(object):
"""
Class that holds information about transmission spectrum
for white beams and spallation sources.
"""
name = ''
timestamp = ''
## Wavelength (float) [A]
wavelength = None
wavelength_unit = 'A'
## Transmission (float) [unit less]
transmission = None
transmission_unit = ''
## Transmission Deviation (float) [unit less]
transmission_deviation = None
transmission_deviation_unit = ''
def __init__(self):
self.wavelength = []
self.transmission = []
self.transmission_deviation = []
def __str__(self):
_str = "Transmission Spectrum:\n"
_str += " Name: \t{0}\n".format(self.name)
_str += " Timestamp: \t{0}\n".format(self.timestamp)
_str += " Wavelength unit: \t{0}\n".format(self.wavelength_unit)
_str += " Transmission unit:\t{0}\n".format(self.transmission_unit)
_str += " Trans. Dev. unit: \t{0}\n".format(\
self.transmission_deviation_unit)
length_list = [len(self.wavelength), len(self.transmission), \
len(self.transmission_deviation)]
_str += " Number of Pts: \t{0}\n".format(max(length_list))
return _str
[docs]class DataInfo(object):
"""
Class to hold the data read from a file.
It includes four blocks of data for the
instrument description, the sample description,
the data itself and any other meta data.
"""
## Title
title = ''
## Run number
run = None
## Run name
run_name = None
## File name
filename = ''
## Notes
notes = None
## Processes (Action on the data)
process = None
## Instrument name
instrument = ''
## Detector information
detector = None
## Sample information
sample = None
## Source information
source = None
## Collimation information
collimation = None
## Transmission Spectrum INfo
trans_spectrum = None
## Additional meta-data
meta_data = None
## Loading errors
errors = None
## SESANS data check
isSesans = None
def __init__(self):
"""
Initialization
"""
## Title
self.title = ''
## Run number
self.run = []
self.run_name = {}
## File name
self.filename = ''
## Notes
self.notes = []
## Processes (Action on the data)
self.process = []
## Instrument name
self.instrument = ''
## Detector information
self.detector = []
## Sample information
self.sample = Sample()
## Source information
self.source = Source()
## Collimation information
self.collimation = []
## Transmission Spectrum
self.trans_spectrum = []
## Additional meta-data
self.meta_data = {}
## Loading errors
self.errors = []
## SESANS data check
self.isSesans = False
[docs] def append_empty_process(self):
"""
"""
self.process.append(Process())
[docs] def add_notes(self, message=""):
"""
Add notes to datainfo
"""
self.notes.append(message)
def __str__(self):
"""
Nice printout
"""
_str = "File: %s\n" % self.filename
_str += "Title: %s\n" % self.title
_str += "Run: %s\n" % str(self.run)
_str += "SESANS: %s\n" % str(self.isSesans)
_str += "Instrument: %s\n" % str(self.instrument)
_str += "%s\n" % str(self.sample)
_str += "%s\n" % str(self.source)
for item in self.detector:
_str += "%s\n" % str(item)
for item in self.collimation:
_str += "%s\n" % str(item)
for item in self.process:
_str += "%s\n" % str(item)
for item in self.notes:
_str += "%s\n" % str(item)
for item in self.trans_spectrum:
_str += "%s\n" % str(item)
return _str
# Private method to perform operation. Not implemented for DataInfo,
# but should be implemented for each data class inherited from DataInfo
# that holds actual data (ex.: Data1D)
def _perform_operation(self, other, operation):
"""
Private method to perform operation. Not implemented for DataInfo,
but should be implemented for each data class inherited from DataInfo
that holds actual data (ex.: Data1D)
"""
return NotImplemented
def _perform_union(self, other):
"""
Private method to perform union operation. Not implemented for DataInfo,
but should be implemented for each data class inherited from DataInfo
that holds actual data (ex.: Data1D)
"""
return NotImplemented
def __add__(self, other):
"""
Add two data sets
:param other: data set to add to the current one
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return a + b
return self._perform_operation(other, operation)
def __radd__(self, other):
"""
Add two data sets
:param other: data set to add to the current one
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return b + a
return self._perform_operation(other, operation)
def __sub__(self, other):
"""
Subtract two data sets
:param other: data set to subtract from the current one
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return a - b
return self._perform_operation(other, operation)
def __rsub__(self, other):
"""
Subtract two data sets
:param other: data set to subtract from the current one
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return b - a
return self._perform_operation(other, operation)
def __mul__(self, other):
"""
Multiply two data sets
:param other: data set to subtract from the current one
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return a * b
return self._perform_operation(other, operation)
def __rmul__(self, other):
"""
Multiply two data sets
:param other: data set to subtract from the current one
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return b * a
return self._perform_operation(other, operation)
def __div__(self, other):
"""
Divided a data set by another
:param other: data set that the current one is divided by
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return a/b
return self._perform_operation(other, operation)
def __rdiv__(self, other):
"""
Divided a data set by another
:param other: data set that the current one is divided by
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
def operation(a, b):
return b/a
return self._perform_operation(other, operation)
def __or__(self, other):
"""
Union a data set with another
:param other: data set to be unified
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
return self._perform_union(other)
def __ror__(self, other):
"""
Union a data set with another
:param other: data set to be unified
:return: new data set
:raise ValueError: raised when two data sets are incompatible
"""
return self._perform_union(other)
[docs]class Data1D(plottable_1D, DataInfo):
"""
1D data class
"""
def __init__(self, x=None, y=None, dx=None, dy=None, lam=None, dlam=None, isSesans=None):
DataInfo.__init__(self)
plottable_1D.__init__(self, x, y, dx, dy,None, None, lam, dlam)
self.isSesans = isSesans
try:
if self.isSesans: # the data is SESANS
self.x_unit = 'A'
self.y_unit = 'pol'
elif not self.isSesans: # the data is SANS
self.x_unit = '1/A'
self.y_unit = '1/cm'
except: # the data is not recognized/supported, and the user is notified
raise TypeError('data not recognized, check documentation for supported 1D data formats')
def __str__(self):
"""
Nice printout
"""
_str = "%s\n" % DataInfo.__str__(self)
_str += "Data:\n"
_str += " Type: %s\n" % self.__class__.__name__
_str += " X-axis: %s\t[%s]\n" % (self._xaxis, self._xunit)
_str += " Y-axis: %s\t[%s]\n" % (self._yaxis, self._yunit)
_str += " Length: %g\n" % len(self.x)
return _str
[docs] def is_slit_smeared(self):
"""
Check whether the data has slit smearing information
:return: True is slit smearing info is present, False otherwise
"""
def _check(v):
if (v.__class__ == list or v.__class__ == np.ndarray) \
and len(v) > 0 and min(v) > 0:
return True
return False
return _check(self.dxl) or _check(self.dxw)
[docs] def clone_without_data(self, length=0, clone=None):
"""
Clone the current object, without copying the data (which
will be filled out by a subsequent operation).
The data arrays will be initialized to zero.
:param length: length of the data array to be initialized
:param clone: if provided, the data will be copied to clone
"""
from copy import deepcopy
if clone is None or not issubclass(clone.__class__, Data1D):
x = np.zeros(length)
dx = np.zeros(length)
y = np.zeros(length)
dy = np.zeros(length)
lam = np.zeros(length)
dlam = np.zeros(length)
clone = Data1D(x, y, lam=lam, dx=dx, dy=dy, dlam=dlam)
clone.title = self.title
clone.run = self.run
clone.filename = self.filename
clone.instrument = self.instrument
clone.notes = deepcopy(self.notes)
clone.process = deepcopy(self.process)
clone.detector = deepcopy(self.detector)
clone.sample = deepcopy(self.sample)
clone.source = deepcopy(self.source)
clone.collimation = deepcopy(self.collimation)
clone.trans_spectrum = deepcopy(self.trans_spectrum)
clone.meta_data = deepcopy(self.meta_data)
clone.errors = deepcopy(self.errors)
return clone
def _validity_check(self, other):
"""
Checks that the data lengths are compatible.
Checks that the x vectors are compatible.
Returns errors vectors equal to original
errors vectors if they were present or vectors
of zeros when none was found.
:param other: other data set for operation
:return: dy for self, dy for other [numpy arrays]
:raise ValueError: when lengths are not compatible
"""
dy_other = None
if isinstance(other, Data1D):
# Check that data lengths are the same
if len(self.x) != len(other.x) or \
len(self.y) != len(other.y):
msg = "Unable to perform operation: data length are not equal"
raise ValueError(msg)
# Here we could also extrapolate between data points
TOLERANCE = 0.01
for i in range(len(self.x)):
if math.fabs((self.x[i] - other.x[i])/self.x[i]) > TOLERANCE:
msg = "Incompatible data sets: x-values do not match"
raise ValueError(msg)
# Check that the other data set has errors, otherwise
# create zero vector
dy_other = other.dy
if other.dy is None or (len(other.dy) != len(other.y)):
dy_other = np.zeros(len(other.y))
# Check that we have errors, otherwise create zero vector
dy = self.dy
if self.dy is None or (len(self.dy) != len(self.y)):
dy = np.zeros(len(self.y))
return dy, dy_other
def _perform_operation(self, other, operation):
"""
"""
# First, check the data compatibility
dy, dy_other = self._validity_check(other)
result = self.clone_without_data(len(self.x))
if self.dxw is None:
result.dxw = None
else:
result.dxw = np.zeros(len(self.x))
if self.dxl is None:
result.dxl = None
else:
result.dxl = np.zeros(len(self.x))
for i in range(len(self.x)):
result.x[i] = self.x[i]
if self.dx is not None and len(self.x) == len(self.dx):
result.dx[i] = self.dx[i]
if self.dxw is not None and len(self.x) == len(self.dxw):
result.dxw[i] = self.dxw[i]
if self.dxl is not None and len(self.x) == len(self.dxl):
result.dxl[i] = self.dxl[i]
a = Uncertainty(self.y[i], dy[i]**2)
if isinstance(other, Data1D):
b = Uncertainty(other.y[i], dy_other[i]**2)
if other.dx is not None:
result.dx[i] *= self.dx[i]
result.dx[i] += (other.dx[i]**2)
result.dx[i] /= 2
result.dx[i] = math.sqrt(result.dx[i])
if result.dxl is not None and other.dxl is not None:
result.dxl[i] *= self.dxl[i]
result.dxl[i] += (other.dxl[i]**2)
result.dxl[i] /= 2
result.dxl[i] = math.sqrt(result.dxl[i])
else:
b = other
output = operation(a, b)
result.y[i] = output.x
result.dy[i] = math.sqrt(math.fabs(output.variance))
return result
def _validity_check_union(self, other):
"""
Checks that the data lengths are compatible.
Checks that the x vectors are compatible.
Returns errors vectors equal to original
errors vectors if they were present or vectors
of zeros when none was found.
:param other: other data set for operation
:return: bool
:raise ValueError: when data types are not compatible
"""
if not isinstance(other, Data1D):
msg = "Unable to perform operation: different types of data set"
raise ValueError(msg)
return True
def _perform_union(self, other):
"""
"""
# First, check the data compatibility
self._validity_check_union(other)
result = self.clone_without_data(len(self.x) + len(other.x))
if self.dy is None or other.dy is None:
result.dy = None
else:
result.dy = np.zeros(len(self.x) + len(other.x))
if self.dx is None or other.dx is None:
result.dx = None
else:
result.dx = np.zeros(len(self.x) + len(other.x))
if self.dxw is None or other.dxw is None:
result.dxw = None
else:
result.dxw = np.zeros(len(self.x) + len(other.x))
if self.dxl is None or other.dxl is None:
result.dxl = None
else:
result.dxl = np.zeros(len(self.x) + len(other.x))
result.x = np.append(self.x, other.x)
#argsorting
ind = np.argsort(result.x)
result.x = result.x[ind]
result.y = np.append(self.y, other.y)
result.y = result.y[ind]
if result.dy is not None:
result.dy = np.append(self.dy, other.dy)
result.dy = result.dy[ind]
if result.dx is not None:
result.dx = np.append(self.dx, other.dx)
result.dx = result.dx[ind]
if result.dxw is not None:
result.dxw = np.append(self.dxw, other.dxw)
result.dxw = result.dxw[ind]
if result.dxl is not None:
result.dxl = np.append(self.dxl, other.dxl)
result.dxl = result.dxl[ind]
return result
[docs]class Data2D(plottable_2D, DataInfo):
"""
2D data class
"""
## Units for Q-values
Q_unit = '1/A'
## Units for I(Q) values
I_unit = '1/cm'
## Vector of Q-values at the center of each bin in x
x_bins = None
## Vector of Q-values at the center of each bin in y
y_bins = None
## No 2D SESANS data as of yet. Always set it to False
isSesans = False
def __init__(self, data=None, err_data=None, qx_data=None,
qy_data=None, q_data=None, mask=None,
dqx_data=None, dqy_data=None):
DataInfo.__init__(self)
plottable_2D.__init__(self, data, err_data, qx_data,
qy_data, q_data, mask, dqx_data, dqy_data)
self.y_bins = []
self.x_bins = []
if len(self.detector) > 0:
raise RuntimeError("Data2D: Detector bank already filled at init")
def __str__(self):
_str = "%s\n" % DataInfo.__str__(self)
_str += "Data:\n"
_str += " Type: %s\n" % self.__class__.__name__
_str += " X-axis: %s\t[%s]\n" % (self._xaxis, self._xunit)
_str += " Y-axis: %s\t[%s]\n" % (self._yaxis, self._yunit)
_str += " Z-axis: %s\t[%s]\n" % (self._zaxis, self._zunit)
_str += " Length: %g \n" % (len(self.data))
_str += " Shape: (%d, %d)\n" % (len(self.y_bins), len(self.x_bins))
return _str
[docs] def clone_without_data(self, length=0, clone=None):
"""
Clone the current object, without copying the data (which
will be filled out by a subsequent operation).
The data arrays will be initialized to zero.
:param length: length of the data array to be initialized
:param clone: if provided, the data will be copied to clone
"""
from copy import deepcopy
if clone is None or not issubclass(clone.__class__, Data2D):
data = np.zeros(length)
err_data = np.zeros(length)
qx_data = np.zeros(length)
qy_data = np.zeros(length)
q_data = np.zeros(length)
mask = np.zeros(length)
dqx_data = None
dqy_data = None
clone = Data2D(data=data, err_data=err_data,
qx_data=qx_data, qy_data=qy_data,
q_data=q_data, mask=mask)
clone._xaxis = self._xaxis
clone._yaxis = self._yaxis
clone._zaxis = self._zaxis
clone._xunit = self._xunit
clone._yunit = self._yunit
clone._zunit = self._zunit
clone.x_bins = self.x_bins
clone.y_bins = self.y_bins
clone.title = self.title
clone.run = self.run
clone.filename = self.filename
clone.instrument = self.instrument
clone.notes = deepcopy(self.notes)
clone.process = deepcopy(self.process)
clone.detector = deepcopy(self.detector)
clone.sample = deepcopy(self.sample)
clone.source = deepcopy(self.source)
clone.collimation = deepcopy(self.collimation)
clone.trans_spectrum = deepcopy(self.trans_spectrum)
clone.meta_data = deepcopy(self.meta_data)
clone.errors = deepcopy(self.errors)
return clone
def _validity_check(self, other):
"""
Checks that the data lengths are compatible.
Checks that the x vectors are compatible.
Returns errors vectors equal to original
errors vectors if they were present or vectors
of zeros when none was found.
:param other: other data set for operation
:return: dy for self, dy for other [numpy arrays]
:raise ValueError: when lengths are not compatible
"""
err_other = None
TOLERANCE = 0.01
if isinstance(other, Data2D):
# Check that data lengths are the same
if len(self.data) != len(other.data) or \
len(self.qx_data) != len(other.qx_data) or \
len(self.qy_data) != len(other.qy_data):
msg = "Unable to perform operation: data length are not equal"
raise ValueError(msg)
for ind in range(len(self.data)):
if math.fabs((self.qx_data[ind] - other.qx_data[ind])/self.qx_data[ind]) > TOLERANCE:
msg = "Incompatible data sets: qx-values do not match: %s %s" % (self.qx_data[ind], other.qx_data[ind])
raise ValueError(msg)
if math.fabs((self.qy_data[ind] - other.qy_data[ind])/self.qy_data[ind]) > TOLERANCE:
msg = "Incompatible data sets: qy-values do not match: %s %s" % (self.qy_data[ind], other.qy_data[ind])
raise ValueError(msg)
# Check that the scales match
err_other = other.err_data
if other.err_data is None or \
(len(other.err_data) != len(other.data)):
err_other = np.zeros(len(other.data))
# Check that we have errors, otherwise create zero vector
err = self.err_data
if self.err_data is None or \
(len(self.err_data) != len(self.data)):
err = np.zeros(len(other.data))
return err, err_other
def _perform_operation(self, other, operation):
"""
Perform 2D operations between data sets
:param other: other data set
:param operation: function defining the operation
"""
# First, check the data compatibility
dy, dy_other = self._validity_check(other)
result = self.clone_without_data(np.size(self.data))
if self.dqx_data is None or self.dqy_data is None:
result.dqx_data = None
result.dqy_data = None
else:
result.dqx_data = np.zeros(len(self.data))
result.dqy_data = np.zeros(len(self.data))
for i in range(np.size(self.data)):
result.data[i] = self.data[i]
if self.err_data is not None and \
np.size(self.data) == np.size(self.err_data):
result.err_data[i] = self.err_data[i]
if self.dqx_data is not None:
result.dqx_data[i] = self.dqx_data[i]
if self.dqy_data is not None:
result.dqy_data[i] = self.dqy_data[i]
result.qx_data[i] = self.qx_data[i]
result.qy_data[i] = self.qy_data[i]
result.q_data[i] = self.q_data[i]
result.mask[i] = self.mask[i]
a = Uncertainty(self.data[i], dy[i]**2)
if isinstance(other, Data2D):
b = Uncertainty(other.data[i], dy_other[i]**2)
if other.dqx_data is not None and \
result.dqx_data is not None:
result.dqx_data[i] *= self.dqx_data[i]
result.dqx_data[i] += (other.dqx_data[i]**2)
result.dqx_data[i] /= 2
result.dqx_data[i] = math.sqrt(result.dqx_data[i])
if other.dqy_data is not None and \
result.dqy_data is not None:
result.dqy_data[i] *= self.dqy_data[i]
result.dqy_data[i] += (other.dqy_data[i]**2)
result.dqy_data[i] /= 2
result.dqy_data[i] = math.sqrt(result.dqy_data[i])
else:
b = other
output = operation(a, b)
result.data[i] = output.x
result.err_data[i] = math.sqrt(math.fabs(output.variance))
return result
def _validity_check_union(self, other):
"""
Checks that the data lengths are compatible.
Checks that the x vectors are compatible.
Returns errors vectors equal to original
errors vectors if they were present or vectors
of zeros when none was found.
:param other: other data set for operation
:return: bool
:raise ValueError: when data types are not compatible
"""
if not isinstance(other, Data2D):
msg = "Unable to perform operation: different types of data set"
raise ValueError(msg)
return True
def _perform_union(self, other):
"""
Perform 2D operations between data sets
:param other: other data set
:param operation: function defining the operation
"""
# First, check the data compatibility
self._validity_check_union(other)
result = self.clone_without_data(np.size(self.data) + \
np.size(other.data))
result.xmin = self.xmin
result.xmax = self.xmax
result.ymin = self.ymin
result.ymax = self.ymax
if self.dqx_data is None or self.dqy_data is None or \
other.dqx_data is None or other.dqy_data is None:
result.dqx_data = None
result.dqy_data = None
else:
result.dqx_data = np.zeros(len(self.data) + \
np.size(other.data))
result.dqy_data = np.zeros(len(self.data) + \
np.size(other.data))
result.data = np.append(self.data, other.data)
result.qx_data = np.append(self.qx_data, other.qx_data)
result.qy_data = np.append(self.qy_data, other.qy_data)
result.q_data = np.append(self.q_data, other.q_data)
result.mask = np.append(self.mask, other.mask)
if result.err_data is not None:
result.err_data = np.append(self.err_data, other.err_data)
if self.dqx_data is not None:
result.dqx_data = np.append(self.dqx_data, other.dqx_data)
if self.dqy_data is not None:
result.dqy_data = np.append(self.dqy_data, other.dqy_data)
return result
[docs]def combine_data_info_with_plottable(data, datainfo):
"""
A function that combines the DataInfo data in self.current_datainto with a
plottable_1D or 2D data object.
:param data: A plottable_1D or plottable_2D data object
:return: A fully specified Data1D or Data2D object
"""
final_dataset = None
if isinstance(data, plottable_1D):
final_dataset = Data1D(data.x, data.y, isSesans=datainfo.isSesans)
final_dataset.dx = data.dx
final_dataset.dy = data.dy
final_dataset.dxl = data.dxl
final_dataset.dxw = data.dxw
final_dataset.x_unit = data._xunit
final_dataset.y_unit = data._yunit
final_dataset.xaxis(data._xaxis, data._xunit)
final_dataset.yaxis(data._yaxis, data._yunit)
elif isinstance(data, plottable_2D):
final_dataset = Data2D(data.data, data.err_data, data.qx_data,
data.qy_data, data.q_data, data.mask,
data.dqx_data, data.dqy_data)
final_dataset.xaxis(data._xaxis, data._xunit)
final_dataset.yaxis(data._yaxis, data._yunit)
final_dataset.zaxis(data._zaxis, data._zunit)
else:
return_string = ("Should Never Happen: _combine_data_info_with_plottabl"
"e input is not a plottable1d or plottable2d data "
"object")
return return_string
if hasattr(data, "xmax"):
final_dataset.xmax = data.xmax
if hasattr(data, "ymax"):
final_dataset.ymax = data.ymax
if hasattr(data, "xmin"):
final_dataset.xmin = data.xmin
if hasattr(data, "ymin"):
final_dataset.ymin = data.ymin
final_dataset.isSesans = datainfo.isSesans
final_dataset.title = datainfo.title
final_dataset.run = datainfo.run
final_dataset.run_name = datainfo.run_name
final_dataset.filename = datainfo.filename
final_dataset.notes = datainfo.notes
final_dataset.process = datainfo.process
final_dataset.instrument = datainfo.instrument
final_dataset.detector = datainfo.detector
final_dataset.sample = datainfo.sample
final_dataset.source = datainfo.source
final_dataset.collimation = datainfo.collimation
final_dataset.trans_spectrum = datainfo.trans_spectrum
final_dataset.meta_data = datainfo.meta_data
final_dataset.errors = datainfo.errors
return final_dataset