Rather than starting with the fully-coherent Vsize, this now starts with
the input Nseg0 (i.e. from the SC search). and works back to Nseg=1.
The user is then left to check the V0 value and whether the search is
appropriate or not. Applies the idea to the follow-up example.

Also adds data from directed study

1) Adds transient section text
2) Some spell-checking and general changes

1) Allows true 'silent' running by switching off tqdm and SFThist output
2) Generalise the check old run setup to take an arbitrary set of
parameters to check (provided) they are in the saved dict. Then adds
DeltaOmega and DeltaFs check
3) Pass through the log_table and gen_tex_tables options so that one can
switch off these outputs
4) Adds create_plots to FollowUp run
5) Simple reorder to Writer parameters to group amplitude parameters

Previously changes in the number of steps required the entire setup to
be unneccerserily recalculated. This fixes this issue and attemps to
simply the logic throughout.

- Applies method to weak_signal method
- Adds methods to read/write setup files - this saves recalculating
  expensive optimial setups
- Simplifies notation and laoyout for generating setup and writing
  tables

- Add option to just update the search object in the follow-up
- Abstract loading in SFT data and check it is not already loaded before
  creating SFTCatalog
- Fix bug in plot_walkers in which the number of steps is not displaying
- Removal of unused code
- Fixes to docs

1) Moves the actual estimation to external function since it is not
contingent on using the MCMC methods etc
2) Change notation to V, Vsky and Vpe
3) Fix some table outputs (Tcoh was forcing integer)
4) Rename of names -> detector_names
5) Adds tests
6) Writes a simpler table if the sky is not searched over

tstart tend refers to signal properties, so is only used when relevant.
Otherwise use the LAL parameter min/maxStartTime

This now estimates the appropriate Fstar0 if nsegs != 1. As such, the
follow-up procudure now requires the search object ot be reinitated at
each step - this is important only if using BSGL

The motivating idea here is to give the user an idea of how well the
MCMC is likely to perform: if the typical number of templates is of
order millions or more then the MCMC will work like a random template
back, jumping back and forth without finding a peak (for a long time).

1) Adds function to produce a .loudest file
2) Improves log output of data, including a histogram of SFT timestamps
3) Generalised automated estimation of the F0s when many SFTs are loaded
4) Adds tref to the par file

Note that the dynamic injection is currently broken - it appears to only
be injecting the signal in one set of data

1) Adds Tcoh to plots
2) prints max det stat
3) Fixes total number of stages