BayesicFitting

Model Fitting and Evidence Calculation

View project on GitHub



class ImageAssistant( object )Source

ImageAssistant contains 2 methods to assist with more dimensional fitting.

  1. getIndices Generates indices for data arrays of any dimension.
        To be used as input in the Fitter classes.
  2. resizeData Resizes the data arrays into a 1-dimensional array.
        To be used as data in the Fitter.

Example

ymap = numpy.arange( 6, dtype=float ).reshape( 2, 3 )
ias = ImageAssistant()
ky = ias.getIndices( ymap )
print( ky.shape )
(6,2)
print( ky[4,0], ky[4,1], ymap[ ky[4,0], ky[4,1] ] )
1 0 4
ias = ImageAssistant( order='F')
ky = ias.getIndices( ymap )
print( ky.shape )
(6,2)
print( ky[4,0], ky[4,1], ymap[ ky[4,1], ky[4,0] ] )
0 1 4
## Suppose y is a 2-dimensional map of something
aass = ImageAssistant( )
input = aass.getIndices( y )
fitter = Fitter( input, some2dModel )
pars = fitter.fit( aass.resizeData( y ) )
yfit = some2dModel.result( input )                  # Double1d
yfit2d = aass.resizeData( yfit, shape=y.shape )     # Double2d

Author Do Kester

ImageAssistant( order='C' )

Helper class to construct from an image, the input arrays needed for the Fitters.

Parameters

  • order : 'C' or 'F'
         set index view according to character
         'C' orders from slow to fast
         'F' orders from fast to slow

getIndices( ya, order='C' )
Generates indices for data arrays of any dimension.

To be used as input in the Fitter classes.

Parameters

  • ya : map
         array of y ( data ) values for which an indexed array
  • order : 'C' or 'F'
         set index view according to character

Returns

  • numpy.array of ints : the indices of the pixels

getPositions( ymap, order='C', center=True, deproject=None )
Return the (x,y) positions of the pixels in the map.

Parameters

  • ya : map
         array of y ( data ) values for which an indexed array
  • order : 'C' or 'F'
         set index view according to character
  • center : bool
         if True, return the positions of the center of the pixels.
         otherwise the (left,lower) corner
  • deproject : callable
         Deprojection method: from projected map to sky position,
         returning (x,y,...) position given the map indices (ix,iy,...)
         Default: returning the indices as floats (+0.5 if center)

Returns

numpy.array of floats : the positions of the pixels

getydata( ya )
Return a copy of ya as a 1 dim array.

Parameters

  • ya : array_like
         map to be reshaped

resizeData( res, shape=None )
Reshape the data (res) into the same shape as the map (ya)

Parameters

  • res : array_like
         result of the fit as a 1-dim array
  • shape : tuple of int
         dimensional lengths of the reconstructable map
         default remembered from a call to getIndices