luis-astro-analysis 0.0.2

A small package assisting in data analysis for Xray astronomy

The fits_util library

Introduction

This package is basically just a collection of useful functions, made for data-analysis of satellite data contained inside fits-files. For now, the scope of this project remains rather compact, featuring support for .evt-files (for lightcurve and CC/CI diagrams) and .pha-files (for spectral analysis). In addition there is support for the qdp-format.

As the data-extraction differs from mission to mission, higher level functions have been prepared in mission-specific files (i.e. ninja.py).

The newest addition to this package is the CCI_Analyzer class, that takes curated data in form of a pandas DataFrame and performs an automatic Colour-Colour and Colour-Intensity analysis. The class further provides the option to an interactive analysis, that lets the user choose certain values (binsize and the in-between colourband-boundaries), to make finding suitable values easier. Examples of how to use these functions can be found below.

Dependencies

This Package depends on the following:

• astropy
• numpy
• pandas
• matplotlib
• seaborn
• ...

Overview

The package has three subcategories, as indicated in the introduction. The first of these is of course fits_util, handling event data, while the other two (not as fleshed out yet), handle spectrum related stuff (pha_util) and qdp-formatted lightcurve-data (qdp_util) respectively.

fits_util

read_mjd_channel expects a path to a event-fits file and returns the event times (using mjd as the unit) and channel info as a pandas DataFrame, with the following columns:

df = {
    "mjd"       : float
    "channel"   : int
}

read_rmf_ebounds expects a path to a rmf-fits file and returns a pandas.a DataFrame containing the energy-bounds for each detector channel: (Tested only for NinjaSat-style rmf-files)

df = {
    "channel"   : int
    "energy"    : float
    "e_lo"      : float
    "e_hi"      : float
}

read_gti_intervals expects a path to a gti-fits file and returns a numpy.ndarray with the following shape:

2D array of shape (N, 2)
- Column 0 : start times
- Column 1 : stop times

The time-unit is mjd

select_good_time expects a 1D numpy.ndarray, containing the mjd-values of all photon events, and a 2D numpy.ndarray, containing gti-data, in the form as returned by read_gti_intervals. The function returns a boolean mask, that excludes all events outside the gti. It can be used like this:

mask = select_good_time(events_all)
events_inside_gti = events_all[mask]

bin_gti expects a numpy.ndarray containing gti-data (see read_gti_intervals), and a float of the binsize in seconds. It then calculates bins using the provided binsize only inside the gti always starting at the left edge. If the binsize is larger than a gti, the whole gti will be considered as one bin. If the binsize fits into the gti n-times fully, but the n+1-th bin would reach outside the gti, it is truncated to the right edge of the gti, producing a smaller bin than the specified binsize. All bins inbetween gti intervals, have the same size as the time gap.

pha_util

TODO

qdp_util

TODO

CCI_Analyzer

As an example, I have provided the method I use to analyse NinjaSat data:

import fits_util as f

# paths
ninja_path = "../../autodata/out/merged.evt"
gti_path = "../../autodata/tmp/CygX2.gti"
rmf_path = "../../autodata/out/CygX2.rmf"

# initial colourbands (the outer limits of the initial bands-dict set the
# outer bounds that the CCI_Analyzer uses as well)
ninja_bands = {
    "soft": (2, 4),
    "med": (4, 8),
    "hard": (8, 20),
}

### preliminary loading and formatting using the fits_util module
events = f.read_mjd_channel(ninja_path)
rmf = f.read_rmf_ebounds(rmf_path)
gti = f.read_gti_intervals(gti_path)
# apply chanel-energy information
df = events.merge(rmf_info[["channel", "energy"]], on="channel", how="left")
# select events inside gti
mask = f.select_good_time(df["mjd"].to_numpy(), gti)
df = df[mask]
# initial binsize (will be used as the reference by CCI_Analyzer)
binsize_sec = 600
binsize_day = binsize_sec / 86400
bins = f.bin_gti(gti, binsize_day)

### initialize CCI_Analyzer:

CCI = CCI_Analyzer()

# load data. The expected formats can be found in the docstrings of the
# fits_util module of the functions above, or below this example.
CCI.load_data(df)
CCI.load_gti(gti)
# set initial bins and colourbands
CCI.set_colorbands(ninja_bands)
CCI.set_bins(600)
CCI.set_bin_bounds(1, 1000)
# interactive selection for binsize and colorband-bounds.
# exiting this plot tells the CCI_Analyzer to use the last enterd values
# for further analysis.
CCI.select_bins_and_colors()
# interactive selection for the different branches.
# to confirm the selection, close the path by clicking the first vertex again.
CCI.select_branches()
# plotting the data, using the new values
CCI.plot_color_intensity()
CCI.plot_color_color()