lucabaldini / ixpeobssim Goto Github PK

[ximpol issue #102]

This should be implemented in the pipeline code.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/42
Created on 2018-02-01T13:45:52.190558+00:00

[ximpol issue #68]

For instance, we might have small

stationary_model
achromatic_model

wrappers for models that only depend (via a spline) on energy or time.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/53
Created on 2018-02-01T13:52:45.452521+00:00

[ximpol issue #109]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/40
Created on 2018-02-01T13:44:25.006147+00:00

[ximpol issue #94]

Having some sort of common structure would help readability.

The Crab example is probably the most advanced, and we could take inspiration from that to compile an example script to be used.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/46
Created on 2018-02-01T13:47:54.057011+00:00

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/8
Created on 2017-07-26T06:37:13.777922+00:00

#!python

# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU **GengReral** Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/5
Created on 2017-05-31T10:01:06.617318+00:00

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/7
Created on 2017-07-26T06:32:48.306369+00:00

[ximpol issue #69]

Right now we have a rough zero-order approximation (assuming the period constant in time) in ximpol/srcmodel/roi.py

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/52
Created on 2018-02-01T13:52:11.811678+00:00

Right now we are using 256. Should we increase this value to 4096?

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/4
Created on 2017-05-22T14:05:24.276222+00:00

[ximpol issue# 149]

We have "GengReral" instead of "General" for the GPL license.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/22
Created on 2018-02-01T13:30:35.948633+00:00

[ximpol issue #122]

We should discuss whether we want to add another response function for that.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/37
Created on 2018-02-01T13:41:16.129772+00:00

[ximpol issue #119]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/38
Created on 2018-02-01T13:41:55.171434+00:00

[ximpol issue #65]

Particularly the bit where we decide whether the z values are passed in the form of an array or a callable.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/54
Created on 2018-02-01T13:53:35.507133+00:00

[ximpol issue #104]

The phase cannot be assigned at the level of the source, but should be added in a second step, after the simulation is done. In fact all the sources (also the non periodic one) has to have a phase assigned (we don't know in real life which event comes from a particular source. Right now I have changed the old default value (-1) to random.uniform(0,1) for non periodic sources, but this is a quick (and wrong) fix.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/41
Created on 2018-02-01T13:45:14.599843+00:00

[ximpol issue #132]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/33
Created on 2018-02-01T13:38:11.872484+00:00

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/2
Created on 2017-05-22T13:07:15.656063+00:00

(And upload the static html on the server.)

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/15
Created on 2017-07-29T07:29:31.097468+00:00

[ximpol issue #130]

Particularly, the clear=True thing sounds like it should default to False.

I need to go back and see where we call the function to make sure we're not badly breaking stuff (most likely this is a leftover from the early days where we were not using plt.figure consistently).

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/34
Created on 2018-02-01T13:38:45.599703+00:00

Original message from Herman.

There's a Chandra low-fidelity simulator, called marx, that runs fast and has been in use for about 20 yr. Have you tested the IXPE simulator using this sort of input instead of using an actual Chandra data set. One uses marx2fits to make a fits file, which has quite a few fields in it per event, including those in flight data. To use marx output, it would be useful to know which keywords from the FITS header are used along with which columns from the event list.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/18
Created on 2017-09-20T16:40:00.656308+00:00

[ximpol issue #79]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/48
Created on 2018-02-01T13:49:54.431446+00:00

[ximpol issue #142]

i.e., at least the FILE and LIST options. (I don't believe the LOG options makes sense.)

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/25
Created on 2018-02-01T13:33:24.581733+00:00

[ximpol issue #141]

This could be very useful in analysing extended sources.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/26
Created on 2018-02-01T13:34:00.510578+00:00

[ximpol issue #135]

In the model we're using the photoelectric cross section is defined between 0.03 and 10 keV. Internally we're using a spline interpolated in linear space and therefore the extrapolation is unreliable.

In turn, the transmission factor for any given column density is only defined between 0.03 and 10 keV, and if we extrapolate the spline beyond the boundaries there's no guarantee that the result is physical (i.e., between 0 and 1).

I suspect this is not an issue in most cases, as there's essentially no events past 10 keV, but definitely something to handle properly.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/30
Created on 2018-02-01T13:36:02.243024+00:00

[ximpol issue #126]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/36
Created on 2018-02-01T13:40:16.511505+00:00

xpmdp is a good example.

The synopsis should give a good idea of what the application is doing when the user types

app --help

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/16
Created on 2017-07-31T07:01:19.640364+00:00

I ran the single_point_source example with 100ks and I got this output plot from xpxspec

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/6
Created on 2017-06-01T14:48:44.287211+00:00

[ximpol issue #74]

This is to handle situations (such as when we have polarization maps) when it is more efficient to combine the calculations of the two numbers in one function call.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/50
Created on 2018-02-01T13:51:04.382390+00:00

[ximpol issue #139]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/27
Created on 2018-02-01T13:34:26.461969+00:00

[ximpol issue #100]

e.g., the binned objects are not yet supporting the logx ad logy arguments.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/43
Created on 2018-02-01T13:46:24.768137+00:00

Where do we want to save the example output and the figures in ixpeobssim? Should we keep the same structure as in ximpol?

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/3
Created on 2017-05-22T13:43:55.171888+00:00

[ximpol issue #115]

It is not a simple check, since it does involve looking at size(x) * size (y) points, so we may not want to do this each time we instantiate an object of the class. However, it could be useful i.e. for spotting trivial errors in input models.

If we decide to leave the responsibility of providing meaningful input entirely to the user, another possibility would be to develop a test utility which automatically performs a few of these kind of sanity checks on the input models. At the moment, I have no idea of how to implement that.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/39
Created on 2018-02-01T13:43:47.247756+00:00

Right now in ixpe.py in the detector dir we are using the PSF parameters for the XIPE detector. Need to correct for this.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/1
Created on 2017-05-19T11:45:36.888700+00:00

[ximpol issue #134]

I believe the original problem was a memory error for large simulations.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/31
Created on 2018-02-01T13:36:56.010057+00:00

The first is useful to study the deadtime.

The second would be useful to create event lists to be passed to ixpesim.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/17
Created on 2017-08-05T07:01:02.314378+00:00

[ximpol issue #133]

One quick (temporary) workaround could be to use this
https://heasarc.gsfc.nasa.gov/ftools/caldb/help/ftmerge.html

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/32
Created on 2018-02-01T13:37:40.416658+00:00

[ximpol issue #148]

The approach suggested here
http://adsabs.harvard.edu/abs/2015APh....68...45K
seems like a promising alternative to the fit of the azimuthal distribution, especially for extended sources.

It shouldn't take too long to give it a shot.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/23
Created on 2018-02-01T13:31:18.334754+00:00

[ximpol issue #129]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/35
Created on 2018-02-01T13:39:33.383971+00:00

The PHASE column was originally meant to help producing phasograms for periodic sources, and is set to random values between 0 and 1 for non-period sources. Since in the latter case there is no relationship between the event time and the phase, this creates subtle problems for regions of in interests where we mix multiple periodic and non periodic sources---e.g., the Crab complex.

[Historically, we painfully realizes this while studying the effect of the dead time in the Crab.]

Bottomline: we should get rid of the PHASE column and fold the event lists after the fact with the proper ephemeris when needed.

We should also make sure that we do not break examples by doing this.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/9
Created on 2017-07-26T06:41:23.167675+00:00

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/20
Created on 2017-11-09T16:24:29.939843+00:00

[ximpol issue #78]

We could possibly use less points than a linear spline and/or be more accurate.

Needs to be accurately studied on a simple example.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/49
Created on 2018-02-01T13:50:26.003455+00:00

[ximpol issue #159]

At the moment we have to change the numbers by hand in a few places and this is clearly error-prone.

Ideally DETCHANS, TLMIN and TLMAX should be written in the IRFs and every other piece of code should retrieve the information from there. Need to go through the python modules, understand the ramifications of the change and come up with a strategy.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/21
Created on 2018-02-01T13:29:25.178713+00:00

[ximpol issue #81]

The recent modification to support subplots

fig = aplpy.FITSFigure(self.hdu_list[0], figure=plt.figure(0,figsize=(10subplot[1], 10subplot[0])), subplot=subplot)

is cropping the ylabel. Can we plot the image with show=False and take care of the subplots externally?

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/47
Created on 2018-02-01T13:49:19.561142+00:00

[ximpol issue #147]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/24
Created on 2018-02-01T13:32:23.048281+00:00

[ximpol issue #72]

This requires a small function to be run and save the signatures as part of the release phase.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/51
Created on 2018-02-01T13:51:36.616149+00:00

[ximpol issue #136]

The formalism is all spelled out in
http://adsabs.harvard.edu/abs/2013MNRAS.431..394W
and we just have to decide how to proceed. We might need a map of the reddening as a function of the sky position.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/29
Created on 2018-02-01T13:35:30.470610+00:00

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/14
Created on 2017-07-29T06:52:47.302289+00:00

[ximpol issue #98]

Not sure why binning events with CMAP algorithm is so memory consuming (actually more than simulating those events in the first place!). We should probably do something for this.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/44
Created on 2018-02-01T13:46:58.699186+00:00

At the moment the converter uses the TSTART and TSTOP information in the input file header.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/19
Created on 2017-10-12T12:20:31.598777+00:00

[ximpol issue #95]

This way, if the clobber is set to False we can move on without loading unnecessary data.

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/45
Created on 2018-02-01T13:47:26.181829+00:00

[ximpol issue #138]

Original url: https://bitbucket.org/ixpesw/ixpeobssim/issues/28
Created on 2018-02-01T13:34:52.383128+00:00