Could not find model for point with AffineMap + ASAGI about easi

[uphoffc]

What are the bounds of the data in your netcdf file (x,y,z vectors)?

[Thomas-Ulrich]

x = 0, 500, 1000, 1500, 2000, 2500, 3000, 3500, 4000, 4500, 5000, 5500, 
    6000, 6500, 7000, 7500, 8000, 8500, 9000, 9500, 10000 ;
 y = 0 ;
 z = 0, 500, 1000, 1500, 2000, 2500, 3000, 3500, 4000, 4500, 5000, 5500, 
    6000, 6500, 7000, 7500, 8000, 8500, 9000, 9500, 10000, 10500, 11000, 
    11500, 12000, 12500, 13000, 13500, 14000 ;
dimensions:
	x = 21 ;
	y = 1 ;
	z = 29 ;
ncdump norciax_210fault_nncia.nc -h
netcdf norciax_210fault_nncia {
types:
  compound initial_condition {
    float s_zz ;
    float s_yy ;
    float s_yz ;
    float s_xx ;
    float s_xz ;
    float s_xy ;
    float d_c ;
    float mu_s ;
  }; // initial_condition
dimensions:
	x = 21 ;
	y = 1 ;
	z = 29 ;
variables:
	float x(x) ;
	float y(y) ;
	float z(z) ;
	initial_condition data(z, y, x) ;
}

[uphoffc]

What are you trying to do? AffineMap mapping point (0,0,0) to (348441,4.76021e+06,0) is correct. The error is thrown because latter point is not part of your netcdf file. So in principle everything works as it should work.

[Thomas-Ulrich]

The setup is not designed by me, I m just debugging it.
I agree that the affine map does not map onto the netcdf.
But then the error itself is not very clear, because it may suggest that easi tries to assess model parameters at a point of the mesh which does not exist.
It does not say that it is a point transformed by the affine map.
Also, why ASAGI is not fixing in this case (e.g. Warn: ASAGI: Coordinate in dimension 2 is out of range. Fixing), but throwing an error? Is it consistent?

[uphoffc]

I agree that the error message is not very intuitive but it's unambiguous:

fault2.yaml@2: Could not find model for point [ 348441 4.76021e+06 0 ] in group 1.

@2 tells that the component at Line 2 is throwing the error (the AffineMap). Then, AffineMap is complaining that it's output point is not accepted by any of its child components.

How would you formulate the error message to make it more clear?

The error is not thrown by ASAGI but by easi. ASAGI would in principle allow points that are out of bounds, but then the value is clamped to the nearest boundary value. In this case, one would get completely wrong values.

[Thomas-Ulrich]

Hi,

I agree that the message was not ambiguous once you know how to interpret it.
Neither I or Taufiq and Emanuelle from INGV could not understand it when they faced it.
There is, therefore, room for improvement in this error message.
Nevertheless, I don't know how to make it clearer.
A minimum would be to add a discussion about this example in:
https://seissol.readthedocs.io/en/latest/easi.html#debugging-easi-script
I will do that if you agree.
Thomas.