deproject 0.2.0

License

The deproject module is released under the BSD 2-Clause license, available as the file LICENSE in the source distribution.

Requirements

The installation assumes that you are installing deproject into the CIAO environment (CIAO 4.11 or later), since this is the easiest way to get the XSPEC models along with Sherpa. The standalone Sherpa version can be used, but in this case you will need to build Sherpa with XSPEC support.

The following Python packages are required:

• sherpa

• Astropy (restricted to version 3.0 when using CIAO 4.11)

• SciPy.

Installation

The deproject module should install with the following command (assuming CIAO 4.11 is already installed):

echo “numpy==1.12.1” > constraints.txt pip install -c constraints.txt ‘astropy<3.1’ deproject

Example

If you have a set of X-ray PHA spectra called src<n>.pi, where <n> is an integer representing the annulus number, and the files contain the XFLT0001 to XFLT0005 header keywords used by the XSPEC projct model, then a Deproject object can be created using the deproject_from_xflt helper routine with the commands:

>>> from deproject import deproject_from_xflt
>>> from astropy import units as u
>>> dep = deproject_from_xflt('src*.pi', 0.492 * u.arcsec)

where, in this example, the XFLT0001 and XFLT0002 keywords, which specify the inner and outer radii of the annulus, are in ACIS pixels, and so need to be multiplied by 0.492 arcseconds to convert to an angle (the second parameter).

This will automatically load the spectra into separate Sherpa datasets, which can be fitted individually, but it is generally easier to use the object returned by deproject_from_xflt. For instance, the following will set the data range to be fit for each spectra and ensure that the background is subtracted before fitting:

>>> dep.ignore(None, 0.5)
>>> dep.ignore(7.0, None)
>>> dep.subtract()

Sherpa functions are used to change the statistic and optimiser:

>>> from sherpa.astro import ui
>>> ui.set_stat('chi2xspecvar')
>>> ui.set_method('levmar')

The data can be fit, and errors estimated for all the parameter, using the onion-skin deprojection approach, with the following commands:

>>> onion = dep.fit()
>>> errs = dep.conf()

The return value includes the density (and errors, if appropriate), as an Astropy Quantity.

>>> print(onion['density'])
print(onion['density'])
      density
      1 / cm3
--------------------
  0.1100953546292787
 0.07736622021374819
 0.04164827967805805
 0.03630168106524076
0.025221797991301052
0.021845331641349316
                 ...
0.012396857131392835
 0.01336640115325031
0.012303975980575187
0.013631563529090736
0.013996131292837352
0.010843683594144967
0.023067220584935984
Length = 20 rows

The on-line documentation contains more information, including creating the Deproject object directly (without the need for the XFLTxxxx keywords).