Treat Triangles Embedded in the Stream Mesh

[saubhagya-gatech]

Triangles with all three vertices on the stream mesh can cause issues for applications like stream network connectivity. They often can divert water from corridor through them and act like a braid. To deal with this a two-stage treatment is implemented, with user options to select the level of treatment. These options can be accessed through a optional flag treat_stream_triangles : {'None', 'moderate', 'strict'} in tessalate_river_aligned function:

• moderate: This will deploy refinement function of the triangle library, where stream triangles are identified and passed on for additional refinement. This can take of most of the triangles, however, it is not an strict enforcement as triangulation algorithm will try to balance between multiple constraints and refinement like angles and cell sizes.
• strict: This method will split the stream triangles that could not be refined with the moderate option, by providing additional point (mid point of the off-river-corridor edge) during triangulation.

See image below where a stream triangle was split; red markers depicts the original stream triangle.

[saubhagya-gatech]

These two tests are failing, I wonder if these are due to some change on NHD data?

FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties - assert 97 == 94
FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties_prune - assert 50 == 49

[ecoon]

See some changes/simplifications...

[ecoon]

These two tests are failing, I wonder if these are due to some change on NHD data?

FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties - assert 97 == 94 FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties_prune - assert 50 == 49

I'm not sure why this would be the case -- CI should redownload the data, and the last commit is passing, so the data would have had to change since Rich's last commit. Maybe this needs to be rebased onto master and then would pass?

[saubhagya-gatech]

These two tests are failing, I wonder if these are due to some change on NHD data?
FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties - assert 97 == 94 FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties_prune - assert 50 == 49

I'm not sure why this would be the case -- CI should redownload the data, and the last commit is passing, so the data would have had to change since Rich's last commit. Maybe this needs to be rebased onto master and then would pass?

Ah, the above test were failing on my local machine as I might have older NHD database for this HUC. The CI is failing at downloading Daymet.
ERROR: failed to solve: process "/bin/sh -c pytest --nbmake --nbmake-kernel=python3 examples/get_Daymet.ipynb" did not complete successfully: exit code: 1

[ecoon]

You had some weird merging going on this this branch, so I streamlined the branch, rebased onto master, and ran yapf on the code.

I also simplified the code in triangle_split_points -- I think this should be no change from your previous implementation, but please check and make sure it still works for you. Calls like "intersects" can be expensive -- you don't really want to do things like any(p.intersects(she) for shp in list_of_shapes), and you definitely don't want to do that whole thing in a loop. See my reimplementation for a better strategy.

[ecoon]

Note that since I rebased, you'll have to:

git checkout ssr/conditioning_update
git fetch
git reset --hard origin/ssr/conditioning_update

rather than just git pull (which will give you an error now).

[saubhagya-gatech]

This is ready for merge.