annotate_streamlines fails

[blaizot]

In the tutorial Cosmological Volume, the display of the gravitational potential map does not work very well. The line "q1.annotate_strealines(...)" plainly crashes for me.

Some discussion of what we see would be nice as well (what stands out in the potential map are the voids, i guess, which is perhaps counterintuitive ?).

[cspotz]

managed to get this on my laptop, I let you cross check it and place it in a new section or where it should be I added some words on voids, maybe simpler to say it orally also during the course

[blaizot]

This does not seem to work, does it ? The streamlines neatly ignore the halo ... If you can make it work, let me know.

[mtrebitsch]

I am not sure that this is an error, actually. I played around a bit and I think it's because the halo itself is moving across the box. One solution to isolate the "converging flows" around the halo would be to create a derived field of "velocity - halo velocity", and use that.

See below my tests:

• I created a spherical region to make sure that I only plot the region I'm interested in
• I plotted the "projected x-velocity field" (so the integral of v_x dz, really) in that region, with colour coding such that red = negative, blue = positive
• I used the same streamline annotations as before

As we can see, the velocity along x neatly matches the streamlines.

Now, luckily, yt has a way of removing the "bulk velocity" from a region. In a given region (which could be all_data or a sphere as I did), you can define a field parameter called bulk_velocity to be, for example, the velocity of the halo of interest. Then, if you deposit the "relative" velocities on a grid, you get essentially velocity - bulk_velocity, which you can nicely plot. See below for an example.

[blaizot]

Awesome !! Thanks ! Can you commit that into the tutorial ? 🙏