Fix wrong sign of the longitudinal wake in LongRangeResistiveWall, thanks to Keon Hee Kim for pointing this out.
Fix ImpedanceModel.add method to raise error before making modifications.

for spectral_density()
- Corrected Hermite mode to include normalisation by the mode number
- some authorefactoring

 modes to effective impedance computation
 -  for transverse plane modes, the frequency sampling should be
 with fractional part of the transverse tune.
 Otherwise code does not work for single bunch case (M=1)
 - added head_tail_form_factor(), a function that computes only
  the denominator of the effective impedance.
  To be used to compute tune shifts in low intensity approximation
  - added tune_shift_from_effective_impedance() to implement it
  in the future
  - some autorefactoring

Add optional output for read_IW2D, read_IW2D_folder and read_ABCI
For read_ABCI, skip the longxdip and longydip cases as they are not implemented yet

Can be used to compute the bunch standard deviations and share it between the different bunches.

Change the default binning for WakePotential to 80 (better value for most cases).

np.float and np.complex are deprecated since NumPy 1.20.

Now works correctly for up to date version of seaborn.

 with monopolar and dipolar components together.
 The longitudinal dipolar component is not saved.

Merge branch 'develop' of https://gitlab.synchrotron-soleil.fr/PA/collective-effects/mbtrack2 into develop

1) group_attributes() will group attributes in the impedance model.
A group can be formed from a list of elements.
This list can be autogenerated by a matching string
2) rename_attribute() simple wrapper to rename an attribute.
Can be used to rename impedance model components.

Correct FIRDamper docstring.

Add version number

Now same as in get_wakepotential

Sections with rho=0 are not considered in resistive_wall_effective_radius any more.
Add a plot_resistivity method.

Fix a hard coded value of  0.5 A ...

Added class representation for ComplexData, Impedance, WakeFunction, WakeField and Beam classes.
Rework the deconvolution/from_wakepotential/to_impedance methods for them to work with any data loaded in the Impedance/WakeFunction classes and not just CST text files.
Add a plot method for the Impedance/WakeFunction classes.
Improved WakeField class by adding __iter__ and __add__ methods.
Add a read_ECHO2D function.

The ImpedanceModel class can now deal with "global impedance" such a resistive wall in addition to do the beta weigthing for "localized" impedances.
The summed wake function after beta weigthing is now normalized by the tracking beta function (ring.optics.local_beta).

Fix transverse wake sign and add Wydip component which was missing.

Add the read_ABCI function to import ABCI computed wake and impedance.
Improve the read_CST function to support wake potential.
Add a plot method for both Impedance and WakeFunction class.

at is kept as a "soft" dependency

Fix ADTS implementation in TransverseMap and TransverseMapSector to use Jx/Jy instead of x/y.
Add Synchrotron.get_adts function which compute ADTS from AT lattice.
Change Synchrotron class docstring regarding ADTS.

Fix ImpedanceModel.power_loss_spectrum that needs a double sided impedance spectrum.
Fix docstrings.

Now split correctly the chromaticity in transverse_map_sector_generator function.
Clarify usage for transverse_map_sector_generator.

PhysicalModel.change_values and PhysicalModel .taper are now more pratical to use.
PhysicalModel.get_aperture provide a way to check the apperture at a given location.

Update the Optics class to be able to compute the phase advance at any location on the lattice.
Add the TransverseMapSector element to be able to split the lattice in several parts.
Add the transverse_map_sector_generator convenience function to split a lattice in several TransverseMapSector elements.

Simplify and optimize adding element into ImpedanceModel.

Save and load methods allow to save Bunch and Beam object and reload them later on.
Load method for Beam allow to choose if mpi is used or not.
Initial modification from Watanyu.

Suppress numpy warning for floating-point operations in the WakePotential class.

Change sign of the imaginary part of the impedance in the Coating class to agree with mbtrack2 convention.