tauRecTools: drop "TauEtaCalib" 4-momentum calibration

Hello,

This MR is dropping the "TauEtaCalib" 4-momentum calibration from the tau reconstruction.
This is a calo-only calibration which is almost identical to the "TauEnergyScale" calibration, except that a tiny correction is applied to the reconstructed calo-only eta to compensate for a tiny average bias (correction derived w.r.t. truth tau visible eta).
Given the poor calo-only angular resolution, an absolute eta correction of at most 0.005 on eta is irrelevant in terms of tau performance.
We could consider deploying an MVA regression at xAOD level to improve the angular resolution, so we don't need this intermediate momentum calibration in the reconstruction.
Benefit: there will be one less tool scheduled (we used to call this tool twice), and a bunch of floats are no longer saved in the xAOD.
In the few places where we were using the TauEtaCalib calibration, we are now using the TauEnergyScale calibration.

This MR affects a number of xAOD output variables, but the changes are very small and under control.
For instance, for our final MVA pt, the relative change in the average pt is ~2.5e-4.
The "TauEtaCalib" variables are now 0. Same for TauJetParameters::TESCalibConstant and TauJetParameters::TESOffset (not useful to keep).
The rest is noise.
```
Py:diff-root         INFO  [TauJetsAux.BDTJetScoreSigTrans]: 38 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNEleScore]: 2 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNEleScoreSigTrans]: 30 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNJetScore]: 3 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNJetScoreSigTrans]: 35 leaves differ
Py:diff-root         INFO  [TauJetsAux.TESCalibConstant]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAux.eta]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.etaFinalCalib]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.etaPanTauCellBased]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.etaTauEtaCalib]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAux.mPanTauCellBased]: 8 leaves differ
Py:diff-root         INFO  [TauJetsAux.phiTauEtaCalib]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAux.pt]: 63 leaves differ
Py:diff-root         INFO  [TauJetsAux.ptFinalCalib]: 63 leaves differ
Py:diff-root         INFO  [TauJetsAux.ptPanTauCellBased]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.ptTauEtaCalib]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAuxDyn.eta_combined]: 40 leaves differ
Py:diff-root         INFO  [TauJetsAuxDyn.m_combined]: 39 leaves differ
Py:diff-root         INFO  [TauJetsAuxDyn.pt_combined]: 63 leaves differ
```

I've minimally updated TrigTauAlgorithmsHolder.py. Further cleanup of deprecated TauCalibrateLC options could be done in TrigTauRecConfig.py and TrigTauRecConfigMT.py, but I would leave this to trigger experts.
Let me just mention the tool name should be unique to distinguish the "caloOnly=True" and "caloOnly=False" cases: currently, a first TauCalibrateLC tool will be configured, and then reused regardless of the caloOnly setting.
https://gitlab.cern.ch/atlas/athena/-/blob/master/Trigger/TrigAlgorithms/TrigTauRec/python/TrigTauAlgorithmsHolder.py#L113

Cheers,
Bertrand