For `Test_compare` version > 0.4.

Rather than from current series of tests.

Downgrade `sort_ssh` from argument of procedure `complete_ssh` to
local variable. `sort_ssh` is not used any longer in procedure
`set_max_speed` since commit 4a64f87d.

This variable was already used in other tests.

Because some tests require results from other tests in this series of
tests.

Because it can be created by analysis scripts.

It must be the same than the one used for other input files of the test.

Remove input file `outside_points_get_1_outerm.csv`. Use result from
test `Region_1_20454_cyclone_3` instead. There is one less point in
the file `outside_points.csv` from the test `Region_1_20454_cyclone_3`
but this does not change the results of the two impacted tests:
`Get_1_outerm` and `Extr_proj`.

The only difference in the test specification of
`Get_1_outerm_noise_2` with test `Get_1_outerm` is the input file
`outside_points.csv`. This difference in specification leads to no
difference in results of the two tests. The input file
`outside_points.csv` for `Get_1_outerm` is problematic because it has
an extra point at (9.375, -35.857) which corresponds to a cyclone for
an old version of SSH data. So we are going to change the input file
of `Get_1_outerm` to take
`$tests_old_dir/Region_1_20454_cyclone_3/outside_points.csv`, the same
one than for `Get_1_outerm_noise_2`, so there is really no use for
`Get_1_outerm_noise_2`.

Remove input file `outside_points_get_1_outerm_noise_2.csv`. Use
result from test instead.

Motivation: provide file `outside_points.csv` instead of keeping it in
`Inst_eddies/Tests/Input/Region_1`.

To new version provided by Aviso.

There is no significant difference with
`Extraction_eddies_region_1_noise`: no impact of `min_amp` on this
region. For an impact of `min_amp`, compare tests
`Extraction_eddies_region_3` and `Extraction_eddies_region_3_min`.

It is `min_area` that we need in the code, in `get_1_outerm`. Instead
of computing it in the main program and trickling it down to
`get_1_outerm`, define it in `config_m`.

Read path to the pre-existing `cont_list` shapefile on the command line.

Because we are going to use it for another path.

It is a little simpler to use `cont_list_m` and we then have the speed
in `cont_list.dbf`. The pre-existing `cont_list` shapefile is now open
for read access only.

This is not very useful, now that there is the test on
`write_cont_list`. The test on `write_cont_list` should not be a
performance penalty.

Move output of `cont_list` from `set_contours` to `set_max_speed`. The
advantage is that we will be able to add the speed to the shapefile.

We rewrite `inst_eddies_Aviso.py` so it is more convenient to use. The
script is only useful for several input files. Otherwise we can
directly use `inst_eddies.py`. So instead of optional arguments
template and last date, we use positional arguments. Then there is no
point in distinguishing dirname and basename. Also, it does not make
sense any longer to read the date from the content of the first
file. The script becomes more general: we do not assume that all input
NetCDF files are in the same -directory and that the time coordinate
is in days since 1950-1-1. The interface is also clearer.

We need ssh of extremum in `set_contours`, to compute `innermost_level_2`.

Because we are going to use it for another field.

Instead of standard output.

Since it is not in the shapefiles.

Add the possibility to write `cont_list` from `set_contours`. We want
to do that without impacting the main program unit `inst_eddies`. We
do not want to centralize variables at hight level in
`inst_eddies`. So we do not want to promote `cont_list` and `n_cont`
to dummy arguments of `set_contours`.

Split dummy argument `e` of `set_max_speed` into relevant components.
Motivation: make appear clearly input and output.