Open luiscamilosc opened 1 year ago
I can't tell if there is an actual bug here just yet, but could you provide your data in a format other than a .png
image? It looks like you've linked an image to the CSV file over at SO here: https://i.stack.imgur.com/hbZTX.png
But if the data is that small, it shouldn't be too much work to copy-paste it into your original comment above I think.
It may also be helpful to assert
on some property of the data you believe should be true, but that is not, to facilitate with reproduction outside of some other plotting/program.
Dear @tylerjereddy thank you so much for your help. The data is long so I will load it here ;)
My real problem is described in the following steps:
I appreciate it if you can support me in identifying the bug or guide me with a good solution using your methods.
I attached: to. [ 'new_mesh.csv', 'old_mesh.csv'] #the *.cvs files with X,Y coordinates of my vertices b. Meshing methodology.pdf a presentation to show you my goal and problem c. remesh_csv_vtu_feflow.py the scripting created for the execution of the process d. the polygon mesh based (old_mesh_antamine_poly.shp) to do a comparison with the output and the new one in the area of interest.
I appreciate to don't sharing this information with third parties for a confidential thing with the company that I'm envolved Please to download the information use this link, that will be available some days: https://bluewhale.us.dhigroup.com/rs?key=DxRTefT9Y8STMSE3Dz2GkLhpyiD9jAbM&encKey=V5RigXyx2AtJHFCdAGXDNveAxD2gBXQW
If the file is confidential, then please delete it. This is a public website.
Please to download the information use this link, that will be available some days: https://bluewhale.us.dhigroup.com/rs?key=DxRTefT9Y8STMSE3Dz2GkLhpyiD9jAbM&encKey=V5RigXyx2AtJHFCdAGXDNveAxD2gBXQW
It looks like this might benefit from a much smaller/focused/concise reproducer, or at least one that can generate its own data programmatically rather than needing to download a bunch of stuff.
Dear @tylerjereddy I shared with you the files that reproduce the error. My code merges two *csv files, then, create a .vtu file to see the consistency of the mesh and also, plots the mesh too.
I shared to you the code complete, if you prefer you can use the upper version or the attached version... then, you can reproduce the complete problem...
Also, I added a short presentation that shows the problem with small triangles generated for the Delaunay method producing bad quality and deformed elements for the triangulation (for me a bug in the execution)... If you can't reproduce the error tell me to explain you better... but the problem persists when the nodes are very close and the points probably have a colinearity between them.
Maybe we'll take a look, thanks. My point was that it would be really nice if you could generate everyting programmatically in a small Python script without needing external file assets/downloads/me to read a presentation deck.
colinearity between them
There's probably not much we can do about degeneracies, sometimes changing the Qhull flags can help a little. Also, such degeneracies are almost always possible to produce creatively in a Python/NumPy-based reproducer, using i.e., cooked up example meshes with intentional colinearity, or whatever degenenarcy you may have.
Delaunay
uses the Qhull library underneath. Here is its FAQ on the subject. There is, unfortunately, not a whole lot that can be done about collinear points and still get a triangulation out.
Describe your issue.
Hi, I have a code in Python that read the points of the coordinates of my nodes and then, create the mesh expected in *.vtu format. However, some of the elements with small sizes (local refinements in my previous mesh) generated triangles with zero angles inside of the mesh. Do you know how can I fix it, is you need a short meeting to show you my problem lcsc@dhigroup.com
See this link with my problem and script: https://stackoverflow.com/questions/76082920/problem-mesh-creation-vtu-files-with-zero-angles-in-the-elements-created-base
Thanks for your amazing contribution!! Regards, lcsc
Reproducing Code Example
Error message
SciPy/NumPy/Python version and system information