SLTimer Tutorial

Milan Williams and Phil Marshall, August 2016

This notebook shows you how to find the estimatation of a lens time delays from example light curve data using the PyCS code. You can work through a detailed tutorial notebook introducing the PyCS code here.

First, we'll import SLTimer, as well as a few other important commands.



In [1]:

    
from __future__ import print_function
%matplotlib inline

import sys
sys.path.append('../python')
import desc.sltimer

%load_ext autoreload
%autoreload 2

Next, let's start a timer instance, and put it to work.



In [2]:

    
timer = desc.sltimer.SLTimer()

The timer can download the 'trialcurves.txt' dataset, which should contain 192 lines of data.



In [3]:

    
url = "https://raw.githubusercontent.com/COSMOGRAIL/PyCS/master/demo/demo1/data/trialcurves.txt"
timer.download(url, and_read=True)









    



Downloaded datafile: https://raw.githubusercontent.com/COSMOGRAIL/PyCS/master/demo/demo1/data/trialcurves.txt
Checking header of "trialcurves.txt"...
Reading "trialcurves.txt"...
[Trial/A] with 192 points imported (0 of them masked).
Checking header of "trialcurves.txt"...
Reading "trialcurves.txt"...
[Trial/B] with 192 points imported (0 of them masked).
Checking header of "trialcurves.txt"...
Reading "trialcurves.txt"...
[Trial/C] with 192 points imported (0 of them masked).
Checking header of "trialcurves.txt"...
Reading "trialcurves.txt"...
[Trial/D] with 192 points imported (0 of them masked).



In [4]:

    
timer.display_light_curves()

Let's use the free-knot spline models for both the intrinsic (AGN) and extrinsic (microlensing) variability on your data, using the PyCS code and following the PyCS tutorial.

You can also try microlensing='polynomial'.

By default, error=None when calling estimate_time_delays: To find only the intrinsic variance: error='intrinsic variance'. To find the full error bars, the relationship between the points, and the instrinsic variance, use: error='complete'

Before we run the pycs code, let's decide whether we want to see the output or not - by default, pycs is very verbose. We do this using a simple redirection class.



In [7]:

    
timer.estimate_time_delays(method='pycs', microlensing='spline', agn='spline', error=None, quietly=True)



In [8]:

    
timer.report_time_delays()









    



Time Delays:
AB   -4.78
AC  -20.89
AD  -68.09
BC  -16.10
BD  -63.31
CD  -47.21



In [9]:

    
timer.display_light_curves()

While we could have estimated the time delay uncertainties above, with a command like

timer.estimate_time_delays(method='pycs',microlensing='spline',agn='spline',error='complete')

we can still do the calculation, with the estimate_uncertainties method:



In [5]:

    
# timer.estimate_uncertainties(n=1,npkl=4)