Astrometrical errors in modified UHF B1934 model

[bennahugo]

As discussed in the mattermost channel the modified UHF B1934 model is off (even with fixes applied to the DFT as discussed in #34). I have not checked 0407. I will need to recalibrate data to check that.

However, below find the quotient phase and amplitude distributions. The modified model is substantially worse than the model I provided and does not subtract correctly.

compared to my model:

Amp Mean and std 1.0042092 +/- 0.03587294
Amp Skewness 0.08708567172288895
Amp Kurtosis 2.0559853858674737
Phase Mean and std 0.006116044 +/- 2.0292552
Phase Skewness 0.03825875371694565
Phase Kurtosis 2.2872743848713952

This will need to be redone as soon as possible, because the resulting systemic astrometrical errors on the calibrated maps in archive (which uses these models to phase up).

[bennahugo]

I think the modifications to the models all need to be checked and qualified before we proceed to use them in production. Can we please roll back to a single component until we checked this?