simple uses of tekigo

A direct adaptation

The simplest adaptation is controlled by the following script:

import numpy as np
from tekigo import (TekigoSolution, adaptation_pyhip)

tekigo_sol = TekigoSolution(
    mesh='../../GILGAMESH/trapvtx/trappedvtx.mesh.h5',
    solution='combu.init.h5',
    out_dir='./Results')

x_coor = tekigo_sol.load_qoi('/Mesh/coord_x')
metric_field = np.where(x_coor<0.1,1,0.7)

tekigo_sol.evaluate_metric(metric_field)

adaptation_pyhip(tekigo_sol, edge_low=1.e-3)

See the tutorial on TekigoSolution() for a complet introduction. Once it is created, you can get a quantity of interest as a numpy array with .qoi_load(). Nympy operations allows any combinations between the quantities of interest. For a direct adaptation, the metric will be in ]0,1] for refinement, and >1 for coarsening.

The method .evaluate_metric() gives an estimation of the potential size of the adapted mesh, and stores the metric and the target_edge in the solution. It can be read from the file tekigo_init.sol.h5.

You can then use the function adaptation_pyhip() to use the built-in sequential mesh adaptation.

An indirect adaptation

An indirect adaptation means that, instead of one metric, the user can built several criterias and pass them to calibrate_metric(). This function will :

mix the criterias
transform into a metric
rescale the metric to ensure a global number of nodes
clip locally the metric if the future edge is too small

The usage is:

import numpy as np
from tekigo import (TekigoSolution, calibrate_metric, adaptation_pyhip)

tekigo_sol = TekigoSolution(
    mesh='../../GILGAMESH/trapvtx/trappedvtx.mesh.h5',
    solution='combu.init.h5',
    out_dir='./Results')

x_coor = tekigo_sol.load_qoi('/Mesh/coord_x')

criteria = {
       "crt_1" :  np.where(x_coor<0.1,1,0.0),
       "crt_2" :  np.where(x_coor>0.15,-1,0.0)
}
metric_field = calibrate_metric(
    tekigo_sol, 
    criteria, 
    target_ncells=400000, 
    edge_min=1.e-3,
    met_mix="abs_max",
)

tekigo_sol.evaluate_metric(metric_field)
adaptation_pyhip(tekigo_sol, edge_low=1.e-3)