This avoids copying `cont_list(n_cont)` to `out_cont` if it has
insufficient area.

Add dummy arguments `cont_list` and `n_cont` to `get_1_outerm`.

Because we are going to add another intent out argument.

Keep the good contours found when looking for the outermost contour,
subject to a maximum of `n_max_cont` contours. The idea is to use them
later when looking for the maximum-speed contour. Inside function
`get_1_outerm`, we temporarily store the good contours in a local
explicit-shape array, `cont_list`.

Rename dummy argument `innermost_level` to `innermost_level_2`.

Since `NR_util` has been integrated into Jumble.

Also split the tests, Fortran source files, Python files and JSon
files. Keep at the top level the files used in both `Inst_eddies` and
Overlap.

Motivations for the split:

- The top directory contained many files.

- It may be useful to compile only one of the two sub-projects. For
example, only `Inst_eddies` on Ciclad, where MPI 3 is not available.

Take into account new policy in `filter.py`: the names of the three
shapefiles are always the same and form a unit in a snaphot directory.

Rename `compil_prod_dir`to `build_dir` in JSON test files. Add a test
of `successive_overlap` with global grid, normal resolution.

Take into account new policy in `filter.py`: the names of the three
shapefiles are always the same and form a unit in a snaphot directory.

Rename `compil_prod_dir`to `build_dir` in JSON test files. Add a test
of `successive_overlap` with global grid, normal resolution.

spherical_polygon_area to spher_polygon_area.

In procedure max_speed_contour_ssh, instead of aborting when the speed
is not defined, continue with missing_ssh. Found such a case in
November 29th, 2015 data. Added corresponding test. In procedure
set_max_speed, if max_speed_contour_ssh returns missing_ssh then we
just take the outermost contour as max-speed contour.

values at longitude boundaries, in case of periodicity.

Add argument "periodic" to procedures get_snapshot and
set_all_outerm. Duplicate values at longitude boundaries in ssh, u, v
if periodic. set_all_outerm can now call local_extrema with periodic
true. In procedure set_all_outerm, we shift the longitudes of outside
points if necessary.

In program test_get_snapshot, define periodic from step and nlon.

internal procedure.

In procedure local_extrema, if periodic, pass arrays "field" and
"extr_map" with a duplicate column at each bound. Then there is no
need to copy to local 3x3 arrays. Add argument my_lbound.

reading and spares some disk space: we can reduce the size of the
field from 20 to 14 characters. We do not loose precision with this
reduction because the order of magnitude of the area is 1e3 km2 or
more so the decimal part of the field was 0. Accordingly, change the
are for a null contour from - 1 to - 1e6 m2.

Bug fix in program test_set_max_speed. No need for width 20 for field "speed".

amplitude min_amp if the difference between innermost level from
local_extrema and ssh extremum is smaller. This is not only a
performance improvement but also an improvement of results: the
algorithm sometimes finds a contour at just min_amp when before it did
not find any: the bisection algorithm converged just below
min_amp. Also, as a result, if an outermost contour is found, it is
now necessarily of sufficient amplitude.

Stdin to Input. Two input dir provided for in tests.json: $input_dir
and $large_input_dir.

Rename everywhere variable nearby_extr to outside_points to avoid
confusion with function nearby_extr. Rename file "nearby_extr.csv" to
"outside_points.csv".

In procedure nearby_extr, include in the selection extrema with
different sign, even if they do not have sufficient amplitude.

Bug fix in read_eddy: use the value of read attribute to define
e%interpolated.

procedure set_all_outerm. The idea is that, in this branch,
nearby_extr will eventually include extrema depending not only on them
having sufficient amplitude but also on their cyclonicity. Both pieces
of information cannot be simply encoded in extr_map.

src_test_get_snapshot.

Move field speed from DBF file max_speed_contour to DBF file extremum
because it can be associated to either the contour in
outermost_contour or in max_speed_contour.

instead of defining the component ssh to the ssh of the extremum or to
the ssh of the outermost contour, use a missing value flag
(1e4). Also, when radius4 >= 2 and the max speed contour is the
outermost contour, use missing value flag in the ssh component of max
speed contour instead of ssh on outermost contour. This makes clearer
output.

Add function null_ssh_contour.

In procedure get_1_outerm, simplify the comparison of level_try and
level_good. So we do not need any longer the dummy argument
ssh_extremum.

For ssh, missing value is huge(0.). We have to take a large value
because Contour_531 has an argument zmax for missing values. For
velocity, missing value is a quiet NaN. We have to take this because
we need to know if the interpolation procedure in mean_speed has not
used any missing value and we do not want to alter the interpolation
procedure itself.

In local_extrema, we accept as local extrema only points with 8 valid
values around them.

When we look for the maximum speed contour, we are hoping that good
contours do not englobe missing values. If this happens, the program
aborts.

When computing the mean speed on a contour, we are hoping that the
interpolation of speed on contour points does not use missing
values. If this happens, the program aborts.

Draw coastlines in plot_snaphot.py. At the same time, this allows the
use of any projection.

with some components intent(in) and one component intent(out) was less
clear. Also, replace argument ind_targ_extr by argument i. This is
simpler both in get_1_outerm and in set_all_outerm.

There is now no need to use an eddy in test_get_1_outerm.

set_all_outerm, for clarity.

acceptable. So define flat_extr by a strict inequality in procedure
get_snapshot.

Procedure set_outermost_contour no longer has a dummy argument
noise_around. We no longer call outermost_possible_level so we do not
have an artificial discontinuity of results with and without maximum
amplitude. We abandon the idea of reducing the amplitude of initial
level_good compared to innermost_level: problem of consistence with
get_snapshot, and it does not seem worth the trouble, just abandon
those problematic extrema. Allow for null outermost_contour instead of
aborting.

In procedure get_snapshot, s%list_vis(i)%suff_amp is first defined
only if flat_extr(i). Also we are no longer sure that
set_outermost_contour finds an outermost contour so we have to test
this to define s%list_vis(i)%suff_amp. noise_around is now defined
only for flat_extr(i) and s%list_vis(i)%suff_amp. Also, since we may
not find an outermost contour even if not flat_extr(i), we update
s%extr_map after the second call to set_outermost_contour.

In procedure local_extrema, we use the mask diff_center instead of
mask_center. So we no longer need procedure construct_mask_center. In
procedure local_extrema, we ignore one in two adjacent degenerate
extrema instead of aborting.

In plot_snapshot.py, color extrema and allow for null outermost
contour.

Synthesize output in test_local_extrema.py.