New densification methods, variable quad length-scales. increased flexibility for defining quad widths, depths #81
Conversation
…function - to define the width, any property other than stream order can be defined through function argument, examples, width based on drainage area, labels in node properties, distance from the outlet, etc.
- this is more robust, especially than the previous method. - Previous method is available as densify_river old. This will be depreciated eventually once new method is test enough times
|
There are some updates in progress for the convexity-enforcing function to make it more robust. Once, that is done, we can go ahead with this PR. |
saubhagya-gatech
changed the title
|
all tests are passing on the local machine. The failing test is due to NHDPlus dataset name change |
For the case when the small segment is one of the tributaries at the junction, we merge it with its child (if there is only one child) and not with its parent.
|
cleaned up code as suggested, and checked all examples. Tests are passing on the local machine. CI is failing at NHD manager due to a property name change. As suggested by Chucho, we can add the patch below, for now; we anyway map all kinds of IDs to "ID" in the properties later in the workflow. Maybe we can bring that mapping at the outset in the source manager itself? In manager_nhd.py after line 254 |
That's fine by me as a temporary fix. Please use "and" over "&" -- they aren't the same thing (one is integer math, the other is boolean comparison operator). |
…om/environmental-modeling-workflows/watershed-workflow into ssr/improve_stream_aligned_meshing
The property codes in NHD Datasets are in lowercase. The code is updated to change property codes to lowercase in cases the new format of the NHD Dataset is encountered. This change is applicable for not only 'NHDFlowline' layer, but also in layers 'NHDPlusCatchment', 'NHDPlusFlowlineVAA', NHDPlusEROMMA'.
Two tests were failing; this could be due to updates in the hydrography in the new NHD dataset. All other tests are passing fine, and all example problems are also running fine. FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties - assert 94 == 97 FAILED watershed_workflow/test/test_hilev.py::test_river_tree_properties_prune - assert 49 == 50
We just keep the node to be merged as self, and , merge it into parent or child based on a flag.
This pull request provides improvements to the stream-alined mixed-polyhedral meshing. The main changes include:
densify_rivers_new, hopefully in a couple of months, once tested on a range of cases, will become the default. Then we can support the old method asdensify_rivers_oldfor reproducibility.