Unstructured Scheme - Mesh Fetching #46

stephenworsley · 2021-03-26T17:27:52Z

This PR adds support for fetching meshes and mesh dimensions from iris. Originally we had planned on using a helper function _get_mesh_and_dim to do this, however iris has updated to the point where this functionality is given in one line, so for clarities sake I have replaced the helper function. Instead of testing the helper function, I have tested _regrid_unstructured_to_rectilinear__prepare, where _get_mesh_and_dim would have been called. I have also added a test for regrid_unstructured_to_rectilinear to show that the whole regridder now functions as a result. Full test coverage can be added in a further PR.

codecov · 2021-03-26T17:29:29Z

Codecov Report

Merging #46 (54bbc1f) into unstructured_scheme (f892f5f) will increase coverage by 4.60%.
The diff coverage is 100.00%.

@@                   Coverage Diff                   @@
##           unstructured_scheme      #46      +/-   ##
=======================================================
+ Coverage                93.47%   98.07%   +4.60%     
=======================================================
  Files                       13       16       +3     
  Lines                      521      624     +103     
=======================================================
+ Hits                       487      612     +125     
+ Misses                      34       12      -22

Impacted Files	Coverage Δ
esmf_regrid/experimental/unstructured_scheme.py	`96.82% <100.00%> (+36.16%)`	⬆️
...nstructured_scheme/test_MeshToGridESMFRegridder.py	`100.00% <100.00%> (ø)`
...st__regrid_unstructured_to_rectilinear__prepare.py	`100.00% <100.00%> (ø)`
..._scheme/test_regrid_unstructured_to_rectilinear.py	`100.00% <100.00%> (ø)`

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f892f5f...54bbc1f. Read the comment docs.

for more information, see https://pre-commit.ci

…com/stephenworsley/iris-esmf-regrid into unstructured_scheme_mesh_extraction * 'unstructured_scheme_mesh_extraction' of https://github.com/stephenworsley/iris-esmf-regrid: [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

…ion' into unstructured_scheme_mesh_extraction * origin/unstructured_scheme_mesh_extraction: [pre-commit.ci] auto fixes from pre-commit.com hooks

jamesp

Looking good @stephenworsley 👍 . A few comments inline

jamesp · 2021-03-30T08:06:53Z

esmf_regrid/experimental/unstructured_scheme.py

-    grid_x, grid_y = get_xy_coords(target_grid_cube)
+    grid_x, grid_y = get_xy_dim_coords(target_grid_cube)
+    mesh = src_mesh_cube.mesh
+    assert mesh is not None


Is this to be properly handled / exception raised in the future? If so maybe a TODO added?

jamesp · 2021-03-30T08:09:08Z

esmf_regrid/experimental/unstructured_scheme.py

-        # mesh is probably an iris Mesh object, though it could also be a MeshCoord
-        mesh, mesh_dim = _get_mesh_and_dim(cube)
+        mesh = cube.mesh
+        assert mesh is not None


TODO here if we plan on adding more error handling

jamesp · 2021-03-30T08:16:51Z

esmf_regrid/experimental/unstructured_scheme.py

-    # mesh belongs to (mesh_dim).
-    mesh, mesh_dim = _get_mesh_and_dim(src_mesh_cube)
+    # mesh belongs to.
+    mesh_dim = src_mesh_cube.mesh_dim()


Previous pseudocode version of this (L28) took mesh as a parameter. Is it not necessary to do this anymore? Is there a possibility that a cube contains more than one mesh?

The previous pseudocode version was written before this functionality was implemented in iris. We've discussed the idea of a cube containing more than one mesh and we don't anticipate this possibility. In fact I don't believe it's possible to define such a thing under the current UGRID conventions.

jamesp · 2021-03-30T08:26:29Z

...s/unit/experimental/unstructured_scheme/test__regrid_unstructured_to_rectilinear__prepare.py

+    return mesh
+
+
+def _flat_mesh_cube():


Can you document what the object that this returns looks like? i.e. data is all ones?

jamesp · 2021-03-30T08:29:17Z

...egrid/tests/unit/experimental/unstructured_scheme/test_regrid_unstructured_to_rectilinear.py

+    src = _add_metadata(src)
+    result = regrid_unstructured_to_rectilinear(src, tgt)
+
+    expected_data = np.ones([5, 6])


It's not obvious to me why the expected data is [5,6] from the regrid operation, the test is relying on setup from another module in _flat_mesh_cube. Could you add a comment explaining why this is the expected outcome?

* Unstructured scheme skeleton (#31) * create skeleton of unstructured regridding code This code is primarily based on Iris's current Area weighted regridding scheme, with the required adaptions for accepting unstructred UGRID data. * Unstructured Scheme - Basic GridInfo Handling (#32) Add basic `GridInfo` function and associated tests. It returns an ESMF regridding object for gridded data. * Unstructured scheme iris source (#33) * refresh cirrus-ci and nox * add iris artifact support * deal with special-case cirrus-ci quoting * review actions * Unstructured Scheme - Basic MeshInfo Handling (#36) * provide iris Mesh to MeshInfo conversion * test against iris feature branch * lint fixes and documentation * fix typo * address review comments. * change tests to better reflect UGRID orientation * address review comments. * address review comments * ASCII art fix * Unstructured Scheme - Cube Creation 2D (#39) * support creation of 2D cubes * lint fixes * rewrite test * remove mesh from _create_cube arguments * add TODO * add TODO * Unstructured Scheme - Mesh Fetching (#46) * support mesh fetching * fix docstring * lint fix * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * add docstrings * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * add and restructure tests * fix test * fix test * address review comments * address review comments Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Bill Little <bill.little@metoffice.gov.uk> Co-authored-by: Anna Booton <anna.booton@metoffice.gov.uk> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* Unstructured scheme skeleton (SciTools#31) * create skeleton of unstructured regridding code This code is primarily based on Iris's current Area weighted regridding scheme, with the required adaptions for accepting unstructred UGRID data. * Unstructured Scheme - Basic GridInfo Handling (SciTools#32) Add basic `GridInfo` function and associated tests. It returns an ESMF regridding object for gridded data. * Unstructured scheme iris source (SciTools#33) * refresh cirrus-ci and nox * add iris artifact support * deal with special-case cirrus-ci quoting * review actions * Unstructured Scheme - Basic MeshInfo Handling (SciTools#36) * provide iris Mesh to MeshInfo conversion * test against iris feature branch * lint fixes and documentation * fix typo * address review comments. * change tests to better reflect UGRID orientation * address review comments. * address review comments * ASCII art fix * Unstructured Scheme - Cube Creation 2D (SciTools#39) * support creation of 2D cubes * lint fixes * rewrite test * remove mesh from _create_cube arguments * add TODO * add TODO * Unstructured Scheme - Mesh Fetching (SciTools#46) * support mesh fetching * fix docstring * lint fix * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * add docstrings * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * add and restructure tests * fix test * fix test * address review comments * address review comments Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Bill Little <bill.little@metoffice.gov.uk> Co-authored-by: Anna Booton <anna.booton@metoffice.gov.uk> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

stephenworsley added 2 commits March 26, 2021 17:04

support mesh fetching

0f327f1

fix docstring

acb4baf

stephenworsley and others added 9 commits March 29, 2021 09:29

lint fix

f820810

[pre-commit.ci] auto fixes from pre-commit.com hooks

a438c48

for more information, see https://pre-commit.ci

add docstrings

e2e148d

[pre-commit.ci] auto fixes from pre-commit.com hooks

57a3842

for more information, see https://pre-commit.ci

add and restructure tests

615c5e9

Merge remote-tracking branch 'origin/unstructured_scheme_mesh_extract…

6f18416

…ion' into unstructured_scheme_mesh_extraction * origin/unstructured_scheme_mesh_extraction: [pre-commit.ci] auto fixes from pre-commit.com hooks

fix test

615da60

fix test

51a2f89

stephenworsley requested a review from jamesp March 29, 2021 10:27

jamesp reviewed Mar 30, 2021

View reviewed changes

stephenworsley added 2 commits March 30, 2021 11:31

address review comments

3081066

address review comments

54bbc1f

jamesp approved these changes Mar 30, 2021

View reviewed changes

jamesp merged commit a33e376 into SciTools:unstructured_scheme Mar 31, 2021

trexfeathers mentioned this pull request Feb 21, 2022

update changelog #151

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Unstructured Scheme - Mesh Fetching #46

Unstructured Scheme - Mesh Fetching #46

stephenworsley commented Mar 26, 2021

codecov bot commented Mar 26, 2021 •

edited

Loading

jamesp left a comment

jamesp Mar 30, 2021

jamesp Mar 30, 2021

jamesp Mar 30, 2021

stephenworsley Mar 30, 2021

jamesp Mar 30, 2021

jamesp Mar 30, 2021

Unstructured Scheme - Mesh Fetching #46

Unstructured Scheme - Mesh Fetching #46

Conversation

stephenworsley commented Mar 26, 2021

codecov bot commented Mar 26, 2021 • edited Loading

Codecov Report

jamesp left a comment

Choose a reason for hiding this comment

jamesp Mar 30, 2021

Choose a reason for hiding this comment

jamesp Mar 30, 2021

Choose a reason for hiding this comment

jamesp Mar 30, 2021

Choose a reason for hiding this comment

stephenworsley Mar 30, 2021

Choose a reason for hiding this comment

jamesp Mar 30, 2021

Choose a reason for hiding this comment

jamesp Mar 30, 2021

Choose a reason for hiding this comment

codecov bot commented Mar 26, 2021 •

edited

Loading