Normalization

[FloWsnr]

Hi guys, thanks for the amazing project.
I am missing some documentation about the field normalization functions inside the Dataset and Dataloader classes.
As I understand a yaml file with the layout

mean:
  velocity: 1
  pressure: 7
  density: 88

std:
  velocity: 0.5
  pressure: 1
  density: 0.3

must be supplied? Is this the correct layout? How does it treat the t1 and t2 data?

Thank you!

[rubenohana]

Hello Florian!

Thank you for your question!
The layout you provided is not correct. It should be (you can look at this example):

mean:
  velocity:
  - 1.1 (where this is the value of the first dimension, so in The Well format, for v_x)
  - 2.2 (where this is for v_y)
  pressure: 7
  density: 88

std:
  velocity:
  - 3.1
  - 4.2 
  pressure: 1
  density: 0.3

If the field has 1 level with 2 values, it means it is a t_1 field, and if it has 2 levels (so 4 values, e.g. in this dataset) it should be a t_2 field.

You can use this script to compute statistics on any data formatted for the Well.

Let me know if you have more questions!

[FloWsnr]

Ah, I somehow missed those yaml files. Thank you so much!