Tracking in tangos

Often it’s helpful to track sets of particles or other tracers within a simulation. For example, one might select inflowing or outflowing gas and then inspect its origin at earlier timesteps or destination at later timesteps.

Tangos enables this kind of tracking. In the following introduction we’ll make use of the ChaNGA tutorial dataset. The steps can easily be adapted for other formats.

Selecting particles

The first step is to select particles of interest. Let’s find the particles in a halo with densities higher than 0.1 protons per cubic centimetre at a particular time. Within a python session type the following:

import tangos, pynbody
particles = tangos.get_halo("tutorial_changa/%960/1").load()
high_rho_gas = particles.gas[pynbody.filt.HighPass('rho',"0.1 m_p cm^-3")]

Now we’ll use those particles to generate a tracker:

from tangos import tracking
tracking.new("tutorial_changa", high_rho_gas)

If logging is enabled you may see messages about creation of the required database objects. The tracking.new function returns the identifying number of the tracker which is unique within the simulation. Assuming this is the first tracker you created, the number will be 1, in which case you will find you can now get the corresponding objects throughout time:

particles_step_832 = tangos.get_item("tutorial_changa/%832/tracker_1").load()

In the example above, tangos returns the corresponding particles at step 832. To be clear, that’s not necessarily the particles that exceed the threshold density at step 832 – it’s the particles that will end up satisfying the criterion at our original step 960.

The normal use for trackers is in conjunction with tangos write: we can now use the tracked particles pretty much like any halo, as follows.

Calculating properties for trackers

Let’s calculate the mean gas density for the tracked particles across time. For this we leave the python session and return to the UNIX shell. Run:

tangos write mean_rho --for tutorial_changa --type tracker

Here,

tangos write is the standard command used in the earlier tutorial
mean_rho is a simple built-in property that calculates the mean density of the tracked region in proton masses per cm^3. Of course, the real power is in being able to define your own properties, but this will do for now. You do not need to make any modifications to properties to make them work with trackers; tangos takes responsibility for figuring out and passing in the tracked particles to your calculation.
--type tracker tells tangos to calculate for trackers rather than regular halos

Check the results

Finally, let’s check the results. This can be done from the web interface (where trackers appear next to the main halo list) or in python as follows:

import tangos, pylab
time, rho = tangos.get_halo("tutorial_changa/%960/tracker_1").calculate_for_progenitors("t()","mean_rho")
pylab.plot(time,rho)