Tweaked values for gf* in the lo gf ancillary

[vineetbansal]

Closes #3472

A diff against the v004 version of the ancillary file in the SDC pipeline will show that only the precision of 3 columns has changed to match the numbers I find in https://github.com/vineetbansal/imap-lo/blob/main/3S5_l1b_ram_maps/map_SCFrame_V2.py

A complete snippet of code to produce these 3 columns is (only the print lines are mine - everything else is from 3S5_l1b_ram_maps/map_SCFrame_V2.py:

import numpy as np


gf = {
    1: 7.0e-5,
    2: 7.9e-5,
    3: 9.7e-5,
    4: 11.2e-5,
    5: 14.0e-5,
    6: 17.7e-5,
    7: 22.5e-5,
    8: 6.721e-5
}

dg = {
    1: 4.9e-5,
    2: 5.5e-5,
    3: 6.8e-5,
    4: 3.0e-5,
    5: 4.5e-5,
    6: 2.0e-5,
    7: 1.4e-5,
    8: 6.721e-5
}

scale = 0.63529412

gf = {k: v * scale for k, v in gf.items()}

dg = {k: v * scale for k, v in dg.items()}

gfu = gf.copy()
gfl = gf.copy()

dgu = gf.copy()
dgl = gf.copy()

scale_u = 1.57407407

gfu = {k: v * scale_u for k, v in gfu.items()}
dgu = {k: gfu[k] - gf[k] for k in gf}
dgu = {k: np.sqrt(dgu[k]**2 + dg[k]**2) for k in dg}

scale_l = 0.36728395

gfl = {k: v * scale_l for k, v in gfl.items()}
dgl = {k: gf[k] - gfl[k] for k in gf}
dgl = {k: np.sqrt(dgl[k]**2 + dg[k]**2) for k in dg}

print("gf\n-----")
for _gf in gf.values():
    print(_gf)
print("dgu\n-----")
for _dgu in dgu.values():
    print(_dgu)
print("dgl\n-----")
for _dgl in dgl.values():
    print(_dgl)

This needs to be signed off by someone on the Lo team.

[vineetbansal]

@frahmanif - the task here is to simply ensure that the gf/dgu/dgl values used in the SDC pipeline and the dropbox pipeline are mutually consistent. I don't know where the authoritative values sit, so the fix could be in either direction. This PR tweaks the SDC's values, but if the fix is in the dropbox pipeline, we can make the changes there.