StokesConverter.h

Classes

StokesConverter -- StokesConverter converts any set of polarizations into any other one (full description)

class StokesConverter

Interface

Public Members

StokesConverter()

StokesConverter(const Vector<Int>& out, const Vector<Int>& in, Bool rescale=False)

~StokesConverter()

StokesConverter(const StokesConverter& other)

StokesConverter& operator=(const StokesConverter& other)

void setConversion(const Vector<Int>& out, const Vector<Int>& in, Bool rescale = False)

void convert(Array<Complex>& out, const Array<Complex>& in) const

void convert(Array<Bool>& out, const Array<Bool>& in) const

void convert(Array<Float>& out, const Array<Float>& in, Bool sigma=False) const

void invert(Array<Bool>& out, const Array<Bool>& in) const

Protected Members

void initConvMatrix()

Description

Review Status

Date Reviewed:

yyyy/mm/dd

Programs:

Tests:

• tStokesConverter

Prerequisite

• Stokes

Etymology

Synopsis

Example

    // create converter
       StokesConverter sc;
       Vector<Int> out(7),in(4);
    // set the input polarizations   
       in(0)=Stokes::RR;
       in(1)=Stokes::LL;
       in(2)=Stokes::RL;
       in(3)=Stokes::LR;
    // set the required output  
       out(0)=Stokes::I;
       out(1)=Stokes::Q;
       out(2)=Stokes::U;
       out(3)=Stokes::V;
       out(4)=Stokes::Ptotal;
       out(5)=Stokes::Pangle;
       out(6)=Stokes::PFlinear;
    // initialize the conversion engine   
       sc.setConversion(out,in);
    // set up some test data   
       Vector<Complex> datain(4),dataout(7);
       datain(0)=1.0;
       datain(1)=0.9;
       datain(2)=0.3;
       datain(3)=0.2;
    // convert the data   
       sc.convert(dataout,datain);
    

Motivation

Thrown Exceptions

To Do

• cope with incomplete input polarizations sensibly
• decide what to do about factor 2 between I and RR/LL,XX/YY etc.

Member Description

StokesConverter()

StokesConverter(const Vector<Int>& out, const Vector<Int>& in, Bool rescale=False)

Set up a conversion from in to out. The in and out vectors contain a list of polarization present/wanted in that order. The in vector should match the data to convert. (CORR_TYPE column in SPECTRAL_WINDOW table contains this info) The rescale option will correct for crosscorrelation data that has been scaled to the level Stokes I (common practice in radioastronomy: even though officially I=XX+YY, in practice we need to do I=(XX+YY)/2, set rescale to True to do the latter).

~StokesConverter()

desctructor

StokesConverter(const StokesConverter& other)

Copy constructor

StokesConverter& operator=(const StokesConverter& other)

Assignment,

void setConversion(const Vector<Int>& out, const Vector<Int>& in, Bool rescale = False)

Change or Set the conversion. Arguments are the same as for constructor above.

void convert(Array<Complex>& out, const Array<Complex>& in) const

convert data, first dimension of input must match that of the input conversion vector used to set up the conversion. Output is resized as needed.

void convert(Array<Bool>& out, const Array<Bool>& in) const

convert flags, first dimension of input must match that of the input conversion vector used to set up the conversion. Output is resized as needed. All output depending on a flagged input will be flagged.

void convert(Array<Float>& out, const Array<Float>& in, Bool sigma=False) const

convert weights, first dimension of input must match that of the input conversion vector used to set up the conversion. Output is resized as needed. Set sigma to True when converting sigma's using this routine.

void invert(Array<Bool>& out, const Array<Bool>& in) const

invert flags, first dimension of input must match that of the output conversion vector used to set up the conversion. Output is resized as needed. All output depending on a flagged input will be flagged. This does the inverse operation of convert, allowing flagging of converted data to be transferred back to the original data.

void initConvMatrix()