Figures from the galpy paper (Bovy 2015)

Jo Bovy - bovy at ias dot edu

galpy and its dependencies.

This repository contains code to produce many of the figures in the paper describing the galpy code (figures up to and including 24, excluding code examples).

Figure 6 can be produced by calling (change the extension of the figure filename to save as a different format)

python figure6.py figure6.png

Figures 7, 8, and 9 and the entries in Table 1 are produced by calling

python fitMWPotential2014.py fitResults.sav --rotcurve=figure8.png --kzcurve=figure7a.png --termcurve=figure7b.png --potname=figure9a.png --densname=figure9b.png --tablename=table1.txt

This routine fits a bulge+disk+halo potential to the data described in section 3.5 in the galpy paper, producing MWPotential2014. Play around with it and add your own data! The labels on some of the plots may be somewhat cut off when using .png, but not when using .ps.

The two panels of figure 10 and figure 12 are produced by calling

python figure10+12.py figure10a.png figure10b.png figure12.ps

(my current version of matplotlib has a weird behavior in that it does not plot anything if I use figure12.png).

Figure 13 takes a while to compute (~20 min.) and is produced by running

python figure13.py figure13.png

The code spits out the time the integration takes for each integrator and the mean dE/E.

Figure 14 can be reproduced by

python figure14.py figure14.png

which again prints the time the integration takes for each integrator and the mean of the Jacobian minus one.

We can generate figures 15 and 16 using

python figure15+16.py figure15.png figure16.png

which prints the mean deviation in the radial and vertical frequencies, and in the corresponding angles.

Similarly, figures 17 and 18 are produced by the following

python figure17+18.py figure17.png figure18.png

which prints the actions and the deviations in the actions, frequencies, and angles. The calculation with actionAngleIsochroneApprox takes a very long time (~10 hours). You can create a version with coarser time sampling by editing the line that says tts= ts[::1] to skip more values in ts (e.g., tts= ts[::20], which only takes about half an hour).

Figure 19, which displays the focal length to use when using the Staeckel approximation of the actions for MWPotential2014, can be reproduced by

python figure19.py figure19.sav figure19.png

The savefile contains a pickle of the 2D array of focal lengths that is displayed. See the figure19.py code for how to read this and what the grid on which it is calculated is. The code prints the radius of a circular orbit for each L grid point. This code also takes a long time: about 2.5 hours.

The two panels of figure 20 can be obtained as

python figure20.py figure20a.ps figure20b.ps

Again, my version of matplotlib has some weird issues with plotting the black points in the top panel for PNG output, which is why this command is written to produce PS figures.

The two panels of figure 21 can be created using

python figure21.py figure21a.png figure21b.png

Figure 22 can in principle be produced by doing

python figure22.py figure22.png

but this will take a very long time, as all of the corrections corresponding to different iterations have to be computed (it does not take forever...).

Figure 24 is created by running

python figure24.py figure24.png

This can also take a very long time to run, especially if the necessary DF corrections for the Dehnen and Shu DFs have not been calculated before; progress in going through the various DFs is printed. This code creates two files containing pickles of the asymmetric drift for all models and the Oort constants. See the code for more information.

The remaining figures require so much computation time to run that it is not particularly interesting to exactly reproduce them. See this gist for code that can be used to reproduce Figures 26, 27, 28, 29, and 30.