Releasing v.0.2.0 #31

anchal-physics · 2024-01-17T19:58:38Z

Feature changes

SOLPS mesh can be extended out to fill the whole area upto the device walls.
- Some cells still lie outside limiter surface.
geqdsk_to_imas records X-points to DD after EFIT finds them

Internal changes

Uses GGDUtils (v.0.2.1) for subset tools as they have been moved out of SOLPS2IMAS in v.0.2.0

- Find gridlines by starting at inner boundary of existing mesh and following gradients of psi_N. Stop at regular intervals to get nodes in the new mesh. - Modify steepest descent path near secondary X-point - The path that is probably meant to go through the secondary X-point will miss the secondary X-point if there's any numerical error (so, always) - Instead, find the X-point (a proposed change to EFIT.jl does this, but we can move that function back to SD4SOLPS if EFIT.jl doesn't accept it) and draw a line from the secondary X-point to the closest vertex on the inner edge of the mesh, replacing the whacky curved path that almost hits the X-point. - At this point, some of the cells are outside the limiting surface

- Add documentation to edge mesh extension functions - JSON instead of YAML: turning the matrices back into matrices is easier this way

- See ProjectTorreyPines/IMASggd.jl#25

- Fix indexing of spaces - Clean up loops - Reduce some function interfaces

Mesh extension

eldond

The test is failing and we have to determine if the mesh extension cell connection bug (seen in GGDUtils example notebook plots) is still present, and if so, fix it.

Test failure:

est Summary:          | Pass  Total   Time
core_profile_extension |    9      9  17.8s
Slice #1 already has a rho_tor_norm profile; skipping
Rho has been added to the equilibrium
cleared ext mesh cache: /home/eldond/.julia/dev/SD4SOLPS/src/../data/__home__eldond__.julia__dev__SD4SOLPS__src__..__sample__ITER_Lore_2296_00000__EQDSK__g002296.00200___home__eldond__.julia__dev__SD4SOLPS__src__..__sample__ITER_Lore_2296_00000__EQDSK__g002296.00200.mesh_ext.json
there are 1 secondary x points
# deleted edges = 9636
edge_profile_extension: Test Failed at /home/eldond/.julia/dev/SD4SOLPS/test/runtests.jl:252
  Expression: all(all_indices .== all_indices_reconstruct)

Stacktrace:
 [1] macro expansion
   @ ~/.julia/juliaup/julia-1.9.4+0.x64.linux.gnu/share/julia/stdlib/v1.9/Test/src/Test.jl:478 [inlined]
 [2] macro expansion
   @ ~/.julia/dev/SD4SOLPS/test/runtests.jl:252 [inlined]
 [3] macro expansion
   @ ~/.julia/juliaup/julia-1.9.4+0.x64.linux.gnu/share/julia/stdlib/v1.9/Test/src/Test.jl:1498 [inlined]
 [4] top-level scope
   @ ~/.julia/dev/SD4SOLPS/test/runtests.jl:223
Test Summary:          | Pass  Fail  Total   Time
edge_profile_extension |   20     1     21  21.8s
ERROR: LoadError: Some tests did not pass: 20 passed, 1 failed, 0 errored, 0 broken.
in expression starting at /home/eldond/.julia/dev/SD4SOLPS/test/runtests.jl:220

eldond · 2024-01-18T01:54:58Z

src/supersize_profile.jl

+            add_subset_element!(all_nodes_sub, space_idx, node_dim, node_idx)
+
+            # Edges
+            if (i > 1) & (i != cut)  # i-1 to i  in the npol direction


This, or a place like it, is probably where the error in the mesh extension is entering. The error is visible in the example notebooks for GGDUtils and SynthDiag.

eldond · 2024-01-18T01:56:43Z

src/supersize_profile.jl

+            end
+
+            # Cells
+            if (i > 1) & (i != cut) & (j > 1)


Another possible oopsie location for connecting the extended mesh

eldond · 2024-01-18T01:58:10Z

src/supersize_profile.jl

+    if isfile(cached_ext_name)
+        # md = YAML.load_file(cached_ext_name)
+        md = JSON.parsefile(cached_ext_name)
+        pfr_transition = md["pfr_transition"]


Uh oh, I could have cached an error. I'm not sure if I set an expiration for caches.

On that note, the sample files might've been generated by a buggy version and saved, whereas the current version might be okay. Beyond where the bug is, I'm not sure WHEN the bug is.

Even if the bug is still present, it remains that the sample files were written with the bug active and so are problematic, but we just merged new versions of the affected repos. Whoops.

I don't think it is because the error was cached. I see this issue in fresh clones of the repo. The GGDUtils plotting recipe does not do any calculations. The misconnected cells are present in the ids so it is plotting them.

eldond · 2024-01-18T01:58:45Z

src/supersize_profile.jl

+    # connections should be nice and simple.
+    # The PFR flag needs to be respected; pfr nodes don't connect to non-pfr nodes
+    pfr = rzpi.(mesh_r[:, 1], mesh_z[:, 1]) .< 1
+    pfr_transition = argmax(abs.(diff(pfr)))


There could be an off by one indexing issue here.

anchal-physics · 2024-01-18T02:36:01Z

The test is failing because the EFIT package registered with Julia registry is not up to date with the EFIT git repo where x_point code is present. To resolve this issue, in your SD4SOLPS directory, do following:

% julia --project=./
               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.9.2 (2023-07-05)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

julia> ]
(SD4SOLPS) pkg> rm EFIT
...
(SD4SOLPS) pkg> add "[email protected]:JuliaFusion/EFIT.jl.git"#master
...
(SD4SOLPS) pkg> instantiate
(SD4SOLPS) # press backspace (delete on mac) to get out
julia> exit()

This should fix the issue in the test case.

anchal-physics · 2024-01-18T02:38:13Z

Since this repo is not a dependency of any other repo, all other version upgrades are still valid and this mesh extension issue does not affect them.

eldond · 2024-01-18T03:59:11Z

Since this repo is not a dependency of any other repo, all other version upgrades are still valid and this mesh extension issue does not affect them.

But they have, within them, links to DVC repos at specific commits, right? Or are the DVC repos not set to specific commits and you can update them? Anyway, the sample files that are in there now are wrong.

eldond · 2024-01-18T04:07:05Z

The test is failing because the EFIT package registered with Julia registry is not up to date with the EFIT git repo where x_point code is present. To resolve this issue, in your SD4SOLPS directory, do following:
% julia --project=./
               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.9.2 (2023-07-05)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

julia> ]
(SD4SOLPS) pkg> rm EFIT
...
(SD4SOLPS) pkg> add "[email protected]:JuliaFusion/EFIT.jl.git"#master
...
(SD4SOLPS) pkg> instantiate
(SD4SOLPS) # press backspace (delete on mac) to get out
julia> exit()
This should fix the issue in the test case.

That did not fix it. Maybe it's because I updated to julia 1.9.4

eldond

Ugh. Test passes. All I have to do is check out the right branch.

The other branch I'm developing has the fix to the extended mesh split. I DID fix it, but it hasn't made it to dev yet, so of course you don't have it. All right, go ahead with this for now, unless you want to cherry-pick cd641ae

* makefile has been added to allow easier creation of working environment * README has been updated to guide the user through the process of creating a working environment and added links to setup and demonstration wiki.

Adding documentation with example, installation instructions

eldond and others added 11 commits November 3, 2023 13:04

Make geqdsk_to_imas modify in place

db21a64

bugfix in repair_eq

46d6a57

geqdsk to IMAS function records X-points to DD after EFIT finds them

5a06a39

Break mesh extension function into smaller functions

131e204

Add caching for mesh extensions

8b60028

- Add documentation to edge mesh extension functions - JSON instead of YAML: turning the matrices back into matrices is easier this way

Adapt to GGD-related functions moving from SOLPS2IMAS to GGDUtils

b6e205f

- See ProjectTorreyPines/IMASggd.jl#25

Populate boundaries for extended mesh

41c51d2

Fixup in response to review

6f838da

- Fix indexing of spaces - Clean up loops - Reduce some function interfaces

Merge pull request #26 from ProjectTorreyPines/mesh_ext

ebfdc3a

Mesh extension

Rolling out v.0.2.0

b81facb

anchal-physics requested a review from eldond January 17, 2024 20:13

eldond requested changes Jan 18, 2024

View reviewed changes

Adding a demonstration of whole project

120e240

eldond previously approved these changes Jan 18, 2024

View reviewed changes

anchal-physics added 2 commits January 18, 2024 17:59

Adding makefile and updated README

48d6386

* makefile has been added to allow easier creation of working environment * README has been updated to guide the user through the process of creating a working environment and added links to setup and demonstration wiki.

Merge pull request #34 from ProjectTorreyPines/example

e6ec1da

Adding documentation with example, installation instructions

anchal-physics dismissed eldond’s stale review via e6ec1da January 19, 2024 02:04

anchal-physics merged commit d4f7166 into master Jan 19, 2024
2 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Releasing v.0.2.0 #31

Releasing v.0.2.0 #31

anchal-physics commented Jan 17, 2024 •

edited

Loading

eldond left a comment

eldond Jan 18, 2024

eldond Jan 18, 2024

eldond Jan 18, 2024

eldond Jan 18, 2024

eldond Jan 18, 2024

anchal-physics Jan 18, 2024

eldond Jan 18, 2024

anchal-physics commented Jan 18, 2024

anchal-physics commented Jan 18, 2024

eldond commented Jan 18, 2024

eldond commented Jan 18, 2024

eldond left a comment

Releasing v.0.2.0 #31

Releasing v.0.2.0 #31

Conversation

anchal-physics commented Jan 17, 2024 • edited Loading

Feature changes

Internal changes

eldond left a comment

Choose a reason for hiding this comment

eldond Jan 18, 2024

Choose a reason for hiding this comment

eldond Jan 18, 2024

Choose a reason for hiding this comment

eldond Jan 18, 2024

Choose a reason for hiding this comment

eldond Jan 18, 2024

Choose a reason for hiding this comment

eldond Jan 18, 2024

Choose a reason for hiding this comment

anchal-physics Jan 18, 2024

Choose a reason for hiding this comment

eldond Jan 18, 2024

Choose a reason for hiding this comment

anchal-physics commented Jan 18, 2024

anchal-physics commented Jan 18, 2024

eldond commented Jan 18, 2024

eldond commented Jan 18, 2024

eldond left a comment

Choose a reason for hiding this comment

anchal-physics commented Jan 17, 2024 •

edited

Loading