Source code for queso_cluster.atoms.norm
import numpy as np
import dask.array as da
import numba as nb
[docs]
def normFunc(dataSquare, func):
"""
Normalizes the data with a user-defined function
Parameters
----------
dataSquare : ndarray
2D array containing the spectral data
func : function
user function which accepts an array and outputs an array of the same shape
Returns
--------
ndarray
2D array of normalized spectral data
"""
return(func(dataSquare))
[docs]
def normZ(dataSquare):
"""
Z-Normalization
TODO
-----
I have to write this function.
Parameters
----------
dataSquare : ndarray
2D array containing the spectral data
Returns
-------
ndarray
2D array of normalized spectral data
"""
print("TBD")
return(dataSquare)
[docs]
def normMaximum(dataSquare, windowIndx=None):
"""
Normalizes the data to the maximum value in a given range
Parameters
----------
dataSquare : ndarray
2D array containing the spectral data
windowIndx : list, optional
List containing the beginning and end (inclusive) of the desired range to find maximum.
If not set, this function will use the full avaliable range of the data
Returns
-------
ndarray
2D array of normalized spectral data
"""
if windowIndx is None:
windowIndx = [0, dataSquare.shape[1]]
ii, jj = windowIndx
#norm_func = lambda x: x/np.nanmax(x[: ii:jj+1], axis=1)[:,None]
#normSquare = da.blockwise(norm_func, 'ij', dataSquare, 'ij', dtype=np.float32)
dataMin = np.nanmin(dataSquare[:, ii:jj+1], axis=1) - 0.001
normSquare = (dataSquare - dataMin[:, None])
dataMax = np.nanmax(normSquare[:, ii:jj+1], axis=1)
normSquare /= dataMax[:, None]
return(normSquare)
[docs]
def normContinuum(dataSquare, continuumIndx):
"""
Normalizes the data to the intensity of a reference position
Parameters
----------
dataSquare : ndarray
2D array containing the spectral data
continuumIndx : int
Integer index of the position to normalize with respect to
Returns
-------
ndarray
2D array of normalized spectral data
"""
norm_func = lambda x: x/(x[:, int(continuumIndx)])[:,None]
normSquare = da.blockwise(norm_func, 'ij', dataSquare, 'ij', dtype=np.float32)
return(normSquare)
