PR#54 broke projections in RAMSES in yt-3.0

[yt-fido]

Originally reported by: nickolas1 (Bitbucket: nickolas1, GitHub: nickolas1)

---

For revisions post-Octree diet, in projections it appears that alternating rows are shifted in alternating directions, and the shift seems to be equal to the cell size at the smallest refinement level. This is subtle if you have several levels of refinement, but it is obvious once individual cells are resolved in the image.

Attached are two outputs of a low res single-level run using these commands: pf = load('output_00001/info_00001.txt') p = ProjectionPlot(pf,2,'Density') p.save()

using revisions 559bca4 (merging PR#54) and 596e053, a few days prior to that merge. (More recent revisions show the same problem). I've used ProjectionPlot for simplicity, but plotting with a projection using to_frb shows the same thing.

artifacty image: http://imgur.com/nXG5qGO

non-artifacty image: http://imgur.com/NQeWioC

---

• Bitbucket: https://bitbucket.org/yt_analysis/yt/issue/624

[yt-fido]

Original comment by Matt Turk (Bitbucket: MatthewTurk, GitHub: MatthewTurk):

---

Fixed in yt-3.0 repo

[yt-fido]

Original comment by Matt Turk (Bitbucket: MatthewTurk, GitHub: MatthewTurk):

---

I believe that this issue is corrected in this pull request:

https://bitbucket.org/yt_analysis/yt-3.0/pull-request/78/split-the-filling-function-into-c-f/diff

[yt-fido]

Original comment by Douglas Rudd (Bitbucket: drudd, GitHub: drudd):

---

Note projection completely fails in my selector test branch.

[yt-fido]

Original comment by nickolas1 (Bitbucket: nickolas1, GitHub: nickolas1):

---

It looks like the y-axis might be ok.

[yt-fido]

Original comment by Matt Turk (Bitbucket: MatthewTurk, GitHub: MatthewTurk):

---

I'm pretty sure this is related to an ijk ordering of cells, and I feel like I fixed it not too long ago, but it may be inside some other changes and will need to be transplanted. I will take a look as soon as I can, but I am still coming back online and it may be Wednesday before I can.

[yt-fido]

Original comment by Douglas Rudd (Bitbucket: drudd, GitHub: drudd):

---

Also, does the problem occur for all projection axes?

[yt-fido]

Original comment by nickolas1 (Bitbucket: nickolas1, GitHub: nickolas1):

---

35 Mb or so: https://www.dropbox.com/sh/4b0sl3ihykjkrq5/qN-hxuSvGX

loading it requires passing the fields: pf = load('output_00001/info_00001.txt', fields=['Density','x-velocity','y-velocity','z-velocity','Pressure'])

[yt-fido]

Original comment by Kacper Kowalik (Bitbucket: xarthisius, GitHub: xarthisius):

---

Is your dataset small enough so that you could upload it somewhere?