Improve provisional colouration function

Dependencies

• [2] ABITXOUU Model trainer: Add provisional colouration
• [3] 5AMZXFS5 model trainer: function to generate inputs to pass neural network for rendering equirectangular projection
• [4] FYOPJXOX Model trainer: North/South comprise the Z axis in the source data
• [*] 6AXPZL5P Try to offload tight loop to rust (for now it segfaults)

Change contents

• replacement in trainmodel/src/helpers.rs at line 142
  [2.1092]→[2.1092:1151](∅→∅)
  − let o = 0.9 * s;

  − (o,0.9,o)
  [2.1092]

  [2.1151]
  + (s.powf(0.6),0.9,s * 0.5)
• replacement in trainmodel/src/helpers.rs at line 144
  [2.1172]→[2.1172:1200](∅→∅)
  − (0.,0.,0.9)
  [2.1172]

  [2.1200]
  + (0.,0.,s * 4. + 0.4)