Time float, increase of amplitude precision

[AntSimi]

No description provided.

[CoriPegliasco]

English modifs :
In CHANGELOG.rst

• Now time allows second precision (instead of daily precision) in storage on uint32 from 01/01/1950 to 01/01/2086
  New identifications are produced with this type, old files could still be loaded.
  If you use old identifications for tracking use the --unraw option to unpack old times and store data with the new format.
• Now amplitude is stored with .1 mm of precision (instead of 1 mm), same advice as for time.

Changes in doc/run_tracking.rst
L8 : Before tracking, ...
L10 : Before tracking, display some identification files. You will learn a lot.
L27: Number of consecutive timesteps with missing detection allowed
L29 : Minimal number of timesteps to be considered as a long trajectory
L70 : same as L27
L72 : same as L 29

In share/tracking.yaml
L2 : Files produced ...
L4 : Path to save outputs
L7 : Minimal number of timesteps to be considered as a long trajectory
L9 : Number of consecutive timesteps with missing detection allowed

Questions :
In src/py_eddy_tracker/appli/eddies.py
L284 : why the exception is "several days steps"?

I think you know better than me what needed change to take into account the precision changes :)

didn't take into account further commits

[ludwigVonKoopa]

some comments on the code,

but as it is an important change, and affect a lot of module / class / functions, we should implements some tests ?

[AntSimi]

but as it is an important change, and affect a lot of module / class / functions, we should implements some tests ?

I run tests from detection to final tracking with grid file with 3h step. For now i didn't imagine other tests.