# Python Machine Learning for Data Generation

I am trying to learn machine learning with Python for a specific application to see if it's doable, right now it's just an idea. The goal is to generate or improve meshes for CFD simulations using ML, or DL, or whichever you recommend me.

The problem is that I found a lot of information about these things but mainly for classification, identification, prediction, and those kinds of things, but I couldn't find information relevant to my problem.

For example, I have this input data:

##INPUT##
# 2 blocks on each cartesian direction (x,y,z)
blocks = (2, 2, 2)

# points that define the limits of the geometry (vertices of a square in this case)
extremes = [(-1, -1, -1),
( 1, -1, -1),
( 1,  1, -1),
(-1,  1, -1),
(-1, -1,  1),
( 1, -1,  1),
( 1,  1,  1),
(-1,  1,  1)]


with this info, I can generate this mesh (using the blockMesh tool from OpenFOAM).

blockMesh generates a set of files, that I have parsed. The output is something like this:

points=[[-1., -1., -1.],
[ 0., -1., -1.],
[ 1., -1., -1.],
[-1.,  0., -1.],
[ 0.,  0., -1.],
....
]

faces = [[ 4.,  1.,  4., 13., 10.],
[ 4.,  3., 12., 13.,  4.],
[ 4.,  9., 10., 13., 12.],
[ 4.,  4., 13., 14.,  5.],
[ 4., 10., 11., 14., 13.],
......
]
neighbour = [1, 2, 4, 3, 5, 3, 6, 7, 5, 6, 7, 7]
owners = [0, 0, 0, 1, 1, 2, 2, 3, 4, 4, ...]


and then I can create the files again with this info.

This is pretty much what I want to do, so I hope someone can recommend something to me. Even if the recommendation is to leave it and do something else. I hope not :-D