# Instrument¶

Instances of Instrument are objects representing a model instrument.

class xpsi.Instrument.Instrument(matrix, energy_edges, channels, channel_edges=None, *args, **kwargs)[source]

Base class for astronomical X-ray instruments on-board space telescopes.

The body of the initialiser must not be changed to ensure inter-module compatibility, but can be extended if appropriate using a call to super().__init__. Specialist constructors can be defined in a subclass using the @classmethod decorator.

Parameters: matrix (ndarray[p,q]) – A $$p \times q$$ matrix which is the product of a redistribution matrix and effective area vector. The input energy intervals must increase along the columns of matrix, and the output channels must increase along the rows of matrix. The units of the elements must be that of an effective area ($$cm^2$$). Generally there will be some available calibration product, and deviations from this nominal response model will be parametrised. So here load some nominal response matrix. energy_edges (ndarray[q+1]) – Energy edges in keV of the instrument energy intervals which must be congruent to the first dimension of the matrix: the number of edges must be $$q + 1$$. The edges must be monotonically increasing. These edges will correspond to the nominal response matrix and any deviation from this matrix (see above). channels (ndarray[p]) – Instrument channel numbers which must be equal in number to the number of rows of the matrix. The number of channels must therefore be $$p$$. These channels will correspond to the nominal response matrix and any deviation from this matrix (see above). In common usage patterns, the channel numbers will increase monotonically with row number, and usually increment by one (but this is not necessary).

Note

That these channel numbers are not used to index the loaded instrument (sub)matrix. The xpsi.Data.index_range property returns bounding row numbers that index the loaded instrument response (sub)matrix in order to operate on an incident signal flux. The channel array contained in xpsi.Data.channels must be a contiguous (ordered) subset of the channel array loaded here.

Note

The dimensions of the response matrix need not be equal, but it is required that the number of input intervals be greater than or equal to the number of output channels – i.e., $$p \leq q$$. If $$p < q$$ then it is implied that subsets of adjacent output channels are effectively grouped together.

Parameters: channel_edges (ndarray[p+1]) – The channel (energy) edges of the instrument, in keV. The array must be congruent to the zeroth dimension of the matrix: the number of edges must be $$p + 1$$. The edges must be monotonically increasing. These edges will correspond to the nominal response matrix and any deviation from this matrix (see above). args (tuple) – Container of parameter instances. kwargs (dict) – If you want to prefix parameters of an instance of this instrument subspace with an identifier, pass it as keyword argument and it will find its way to the base class.
__call__(signal, irange, orange)[source]

Register an incident signal.

Parameters: signal (ndarray[m,n]) – An $$m \times n$$ matrix, where input energy interval increments along rows, and phase increases along columns. The number of rows, $$m$$, must equal the number of columns of matrix: $$m=q$$. irange (array-like) – Indexable object with two elements respectively denoting the indices of the first and last input intervals. The response matrix matrix must be indexable with these numbers, i.e., they must satisfy $$indx < q$$. orange (array-like) – Indexable object with two elements respectively denoting the indices of the first and last output channels. The response matrix matrix must be indexable with these numbers, i.e., they must satisfy $$indx < p$$. ndarray[p,n] containing the registered signal.

Note

The product of the most recent operation is stored as the property cached_signal.

cached_signal

Get the cached registered signal.

channel_edges

Get the channel (energy) edges of the instrument, in keV.

A numpy.ndarray of edges of the registered energy intervals labelled as channels defined in the data space. This is relevant when there is a detector-by-detector gain scale applied to event data (such as for NICER instrument calibration products), meaning that the redistribution matrix is effectively shared by detectors and the channels across detectors can share an energy scale definition.

An incident photon of given energy then has a registered-energy distribution that generally peaks in the vicinity of the true photon energy. The resdistribution matrix will have some energy resolution (along with other features such as shelves). With thanks to Paul S. Ray for explaining the choice to calibrate in this manner.

Note

If you made a channel cut that results in a non-contiguous subset of channels, you will need to overwrite the setter method because the checks will fail.

channels

Get the array of channels corresponding to rows of the matrix.

The matrix being the loaded instrument response (sub)matrix.

construct_matrix()[source]

Construct the response matrix if it is parameterised.

If customising, do operations to calculate a matrix, and return it. You can access parameters (free, fixed, and derived) via the container access self[<name>].

If the instrument operation is fixed, you might not need to subclass, because the default behaviour is to return the nominal response you loaded. If for some reason the matrix you loaded is to be modified in some fixed manner, possibly as a function of some custom fixed parameters that you defined, you would also have to subclass and provide the correct implementation of this method.

energy_edges

Get the energy edges of the instrument, in keV.

A numpy.ndarray of edges of the input energy intervals which map to channels defined in the data space.

matrix

Get the reference response matrix.

In common usage patterns there will be some fiducial or nominal response matrix that either defines fixed instrument operation or is a basis for parametrised deviations. This matrix is usually a calibration product distributed by an instrument calibration team.

A matrix of dimension $$p \times q$$. Here $$p$$ must be the number of input energy intervals, and $$q \geq p$$ the number of output channels.

Note

The attribute matrix must be assigned, and it must be a numpy.ndarray for use with numpy.dot() (even if the matrix is sparse to some degree).

class xpsi.Instrument.ResponseError[source]

Bases: xpsi.global_imports.xpsiError

Raised if there is a problem with the input response matrix.

class xpsi.Instrument.EdgesError[source]

Bases: xpsi.global_imports.xpsiError

Raised if there is a problem with the input energy edges.

class xpsi.Instrument.ChannelError[source]

Bases: xpsi.global_imports.xpsiError

Raised if there is a problem with the input channel numbers.