Transformed Profiles

Any profile can be modified in a number of ways, including stretching, dilating, rotating, and translating, and even more complicated things like forming the inverse in Fourier space to effect a deconvolution.

Affine Transformations

class galsim.Transformation(obj, jac=None, offset=(0.0, 0.0), flux_ratio=1.0, gsparams=None, propagate_gsparams=True)[source]

Bases: GSObject

A class for modeling an affine transformation of a GSObject instance.

Typically, you do not need to construct a Transformation object explicitly. This is the type returned by the various transformation methods of GSObject such as GSObject.shear, GSObject.rotate, GSObject.shift, GSObject.transform, etc.

All the various transformations can be described as a combination of a jacobian matrix (i.e. GSObject.transform) and a translation (GSObject.shift), which are described by (dudx,dudy,dvdx,dvdy) and (dx,dy) respectively.

Note

The name of the flux_ratio parameter is technically wrong here if the jacobian has a non-unit determinant, since that would also scale the flux. The flux_ratio parameter actually only refers to an overall amplitude ratio for the surface brightness profile. The total flux scaling is actually |det(jac)| * flux_ratio.

Parameters:

obj – The object to be transformed.
jac – A list, tuple or numpy array ( dudx, dudy, dvdx, dvdy ) describing the Jacobian of the transformation. Use None to indicate that the Jacobian is the 2x2 unit matrix. [default: None]
offset – A galsim.PositionD or tuple giving the offset by which to shift the profile. [default: (0.,0.)]
flux_ratio – A factor by which to multiply the surface brightness of the object. (Technically, not necessarily the flux. See above.) [default: 1]
gsparams – An optional GSParams argument. [default: None]
propagate_gsparams – Whether to propagate gsparams to the transformed object. This is normally a good idea, but there may be use cases where one would not want to do this. [default: True]

Attributes:

original – The original object that is being transformed.
jac – The jacobian of the transformation matrix.
offset – The offset being applied.
flux_ratio – The amount by which the overall surface brightness amplitude is multiplied.
gsparams – The usual gsparams attribute that all GSObject classes have.

Note: if gsparams is unspecified (or None), then the Transformation instance inherits the GSParams from obj. Also, note that parameters related to the Fourier-space calculations must be set when initializing obj, NOT when creating the Transform (at which point the accuracy and threshold parameters will simply be ignored).

property flux_ratio: The flux ratio of the transformation.

property jac: The Jacobian of the transforamtion.

property offset: The offset of the transformation.

property original: The original object being transformed.

withGSParams(gsparams=None, **kwargs)[source]: Create a version of the current object with the given gsparams

Note

Unless you set propagate_gsparams=False, this method will also update the gsparams of the wrapped component object.

galsim.Transform(obj, jac=None, offset=(0.0, 0.0), flux_ratio=1.0, gsparams=None, propagate_gsparams=True)[source]

A function for transforming either a GSObject or ChromaticObject.

This function will inspect its input argument to decide if a Transformation object or a ChromaticTransformation object is required to represent the resulting transformed object.

Note: the name of the flux_ratio parameter is technically wrong here if the jacobian has a non-unit determinant, since that would also scale the flux. The flux_ratio parameter actually only refers to an overall amplitude ratio for the surface brightness profile. The total flux scaling is actually |det(jac)| * flux_ratio.

Parameters:

obj – The object to be transformed.
jac – A list or tuple ( dudx, dudy, dvdx, dvdy ) describing the Jacobian of the transformation. Use None to indicate that the Jacobian is the 2x2 unit matrix. [default: None]
offset – A galsim.PositionD or tuple giving the offset by which to shift the profile. [default: (0.,0.)]
flux_ratio – A factor by which to multiply the surface brightness of the object. (Technically, not necessarily the flux. See above.) [default: 1]
gsparams – An optional GSParams argument. [default: None]
propagate_gsparams – Whether to propagate gsparams to the transformed object. This is normally a good idea, but there may be use cases where one would not want to do this. [default: True]

Returns:

a Transformation or ChromaticTransformation instance as appropriate.

galsim._Transform(obj, jac=None, offset=(0.0, 0.0), flux_ratio=1.0)[source]

Approximately equivalent to Transform, but without some of the sanity checks (such as checking for chromatic options) or setting a new gsparams.

For a ChromaticObject, you must use the regular Transform.

Parameters:

obj – The object to be transformed.
jac – A 2x2 numpy array describing the Jacobian of the transformation. Use None to indicate that the Jacobian is the 2x2 unit matrix. [default: None]
offset – The offset to apply. [default (0.,0.)]
flux_ratio – A factor by which to multiply the surface brightness of the object. [default: 1.]

De-convolution of a GSObject

class galsim.Deconvolution(obj, gsparams=None, propagate_gsparams=True)[source]

Bases: GSObject

A class for deconvolving a GSObject.

The Deconvolution class represents a deconvolution kernel. Note that the Deconvolution class, or compound objects (Sum, Convolution) that include a Deconvolution as one of the components, cannot be photon-shot using the ‘phot’ method of GSObject.drawImage method.

You may also specify a gsparams argument. See the docstring for GSParams for more information about this option. Note: if gsparams is unspecified (or None), then the Deconvolution instance inherits the same GSParams as the object being deconvolved.

The normal way to use this class is to use the Deconvolve() factory function:

>>> inv_psf = galsim.Deconvolve(psf)
>>> deconv_gal = galsim.Convolve(inv_psf, gal)

Parameters:

obj – The object to deconvolve.
gsparams – An optional GSParams argument. [default: None]
propagate_gsparams – Whether to propagate gsparams to the deconvolved object. This is normally a good idea, but there may be use cases where one would not want to do this. [default: True]

property orig_obj: The original object that is being deconvolved.

withGSParams(gsparams=None, **kwargs)[source]: Create a version of the current object with the given gsparams

Note

Unless you set propagate_gsparams=False, this method will also update the gsparams of the wrapped component object.

galsim.Deconvolve(obj, gsparams=None, propagate_gsparams=True)[source]

A function for deconvolving by either a GSObject or ChromaticObject.

This function will inspect its input argument to decide if a Deconvolution object or a ChromaticDeconvolution object is required to represent the deconvolution by a surface brightness profile.

Parameters:

obj – The object to deconvolve.
gsparams – An optional GSParams argument. [default: None]
propagate_gsparams – Whether to propagate gsparams to the deconvolved object. This is normally a good idea, but there may be use cases where one would not want to do this. [default: True]

Returns:

a Deconvolution or ChromaticDeconvolution instance as appropriate.

Fourier-space Square Root

class galsim.FourierSqrtProfile(obj, gsparams=None, propagate_gsparams=True)[source]

Bases: GSObject

A class for computing the Fourier-space sqrt of a GSObject.

The FourierSqrtProfile class represents the Fourier-space square root of another profile. Note that the FourierSqrtProfile class, or compound objects (Sum, Convolution) that include a FourierSqrtProfile as one of the components cannot be photon-shot using the ‘phot’ method of GSObject.drawImage method.

You may also specify a gsparams argument. See the docstring for GSParams for more information about this option. Note: if gsparams is unspecified (or None), then the FourierSqrtProfile instance inherits the same GSParams as the object being operated on.

The normal way to use this class is to use the FourierSqrt factory function:

>>> fourier_sqrt = galsim.FourierSqrt(obj)

Parameters:

obj – The object to compute Fourier-space square root of.
gsparams – An optional GSParams argument. [default: None]
propagate_gsparams – Whether to propagate gsparams to the transformed object. This is normally a good idea, but there may be use cases where one would not want to do this. [default: True]

property orig_obj: The original object being Fourier sqrt-ed.

withGSParams(gsparams=None, **kwargs)[source]: Create a version of the current object with the given gsparams

Note

Unless you set propagate_gsparams=False, this method will also update the gsparams of the wrapped component object.

galsim.FourierSqrt(obj, gsparams=None, propagate_gsparams=True)[source]

A function for computing the Fourier-space square root of either a GSObject or ChromaticObject.

The FourierSqrt function is principally used for doing an optimal coaddition algorithm originally developed by Nick Kaiser (but unpublished) and also described by Zackay & Ofek 2015 (http://adsabs.harvard.edu/abs/2015arXiv151206879Z). See the script make_coadd.py in the GalSim/examples directory for an example of how it works.

This function will inspect its input argument to decide if a FourierSqrtProfile object or a ChromaticFourierSqrtProfile object is required to represent the operation applied to a surface brightness profile.

Parameters:

obj – The object to compute the Fourier-space square root of.
gsparams – An optional GSParams argument. [default: None]
propagate_gsparams – Whether to propagate gsparams to the transformed object. This is normally a good idea, but there may be use cases where one would not want to do this. [default: True]

Returns:

a FourierSqrtProfile or ChromaticFourierSqrtProfile instance as appropriate.