[jcapriot]

This essentially moves the code that was internal to plot_slice to a user-facing utility.

discretize.utils.slice_tree_mesh builds the 2D TreeMesh of an axis-aligned slice through a 3D TreeMesh (one 2D cell per 3D cell cut by the plane), and can also return the index of the 3D cell each 2D cell comes from, so that values_2d = values_3d[indices]. A plane lying on cell faces takes the cells on its high side, or its low side at the top of the mesh, which makes slicing exactly at a boundary well defined.

TreeMesh.plot_slice now uses it instead of building the slice itself.

[codecov]

Codecov Report

❌ Patch coverage is 89.28571% with 18 lines in your changes missing coverage. Please review.
✅ Project coverage is 87.06%. Comparing base (f6e7d8a) to head (3a75be4).

Files with missing lines	Patch %	Lines
discretize/mixins/mpl_mod.py	5.88%	16 Missing ⚠️
discretize/utils/mesh_utils.py	95.91%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #420      +/-   ##
==========================================
+ Coverage   86.86%   87.06%   +0.20%     
==========================================
  Files          93       93              
  Lines       19239    19348     +109     
  Branches     2154     2157       +3     
==========================================
+ Hits        16712    16846     +134     
+ Misses       1840     1814      -26     
- Partials      687      688       +1     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[santisoler]

Looks nice! I like the fact to be able to slice a mesh and not necessarily just plotting it.

One idea: what if instead of defining a slice_tree_mesh we define a slice_mesh function that can take either a TreeMesh or a TensorMesh?

At least for now we could raise a NonImplementedError if the passed mesh is a TensorMesh, and we can add that functionality later.

[jcapriot]

Looks nice! I like the fact to be able to slice a mesh and not necessarily just plotting it.

One idea: what if instead of defining a slice_tree_mesh we define a slice_mesh function that can take either a TreeMesh or a TensorMesh?

At least for now we could raise a NonImplementedError if the passed mesh is a TensorMesh, and we can add that functionality later.

Easy enough to add it for TensorMesh, but it will still have to fail for all of the other mesh types as well

[santisoler]

Yes, that should be fine. If we want to make it clearer, maybe we could name it to slice_rectangular_mesh, but I think slice_mesh should be enough. We can explain it in the docs that it expects either a TreeMesh or a TensorMesh.

[jcapriot]

@santisoler I implemented it for the tensor mesh path as well since it let us get rid of redundant code within the plotting scripts in the same way

[santisoler]

Amazing! Thanks for taking care of this!