Writing a formatted text file
I really like pandas for reading and manipulating largish (~tens of thousands objects) astronomy catalogs in python. The pandas read_cvs function makes it very very easy to ingest these tables into a “dataframe”, with headers and correct data type for each column.
However, the default functions for saving panda dataframes to a file are not really built for formatted output. df.to_cvs will basically spit out what you put in, with or without a header or indices, and optionally for a subset of columns.
One of the things I often want to do is actually LOOK at my galaxies, to make sure that what the catalog says about them makes sense and there not some kind measurement error in the catalog. I mostly work with distant galaxies from Hubble Space Telescope survey images, which might have thousands of galaxies in a single image.
The quick and easy way to check out the galaxies I’m interested in is to display the Hubble image in the SAO ds9 software and plot circles around them using a “regions” file. This is a text file with galaxy positions (Right Ascension and Declination) and maybe a galaxy label/id number. However, this text file needs to be formatted in a particular way.
To make a ds9 regions file from a catalog read in with pandas:
import numpy as np
import pandas as pd
#read in a catalog with columns ID, RA, DEC, Z_SPEC, CLASS_STAR
#and probably lots of other things.
galcat = pd.read_csv(galaxyfile, delim_whitespace=True, header=0, comment='#')
#select quasar candidates: objects that look like stars but aren't.
quasars = galcat[(galcat['Z_SPEC'] >=0.5) & (galcat['CLASS_STAR'] > 0.85)]
#open output regions file (needs .reg extension)
f=open('quasars.reg', "w")
#for each quasar candidate, write regions file format for drawing 1" circle
#at RA,DEC (in degrees) with text label ID
for r,d,x in zip(quasars['RA'], quasars['DEC'], quasars['ID']) :
f.write('fk5;circle(%12.7f, %12.7f, 1\") #text=\"%6i\" \n' % (r, d, x)
f.close()
The output file looks like this:
fk5;circle( 53.1220463, -27.9387409, 1”) #text=” 273”
fk5;circle( 53.0788207, -27.8840976, 1”) #text=” 3073”
fk5;circle( 53.1743757, -27.8673487, 1”) #text=” 4936”
fk5;circle( 53.1648153, -27.8143715, 1”) #text=” 10560”
fk5;circle( 53.0393615, -27.8018869, 1”) #text=” 12122”
And, when displayed with the ds9 software, you get this:
