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.

Call the Matlab script from the Python script. We cannot do this with
subprocess because Matlab requires a terminal. With subprocess, the
Matlab script complains: "inappropriate ioctl for device". Note that
we cannot test that the Matlab script worked normally because it exit
with code 0 regardless: the exit instruction returns code 0.

The idea here is that we do not want to store the converted v6 mat files.

We keep the Matlab script as simple as possible: we do not process
filenames with embedded dates, we assume general names, implying that
we create a symbolic link before calling the Matlab script and rename
the output file.

I found taking advantage of pkg-config useful on Ciclad, because we
already have to play with pkg-config to find NetCDF-Fortran.

Include a test so shapefile center is skipped if not present. The
triplet of shapefiles has possibly become a collection of shapefiles
since commit 4e2ab0a6.

`-init=minus_huge` does not exist for ifort 15 and ifort 15 is the
most recent version available on Ciclad.

We already have it: it is `sys.argv[1]`.

n is for node number, e for eddy number at a given date.

It is not really meaningful to keep function `print_edgelist` since
commit 1e4d2ee1.

e1 is directly i1 + 1 and we just need the new function `date_index`
to process k1. Additional benefit: we process k1 out of the loop on
`id_child_1`.

Which is separate for cyclones and anticyclones. For now. Then we do
not need the argument cyclone of functions `convert_id` and
`print_edgelist`, and we do not need to read `Nanti.mat` for
processing cyclones.

Make them equal to basenames of shapefiles. Clearer and allows a
little simplification in programming.

There may be a fourth shapefile for centers of eddies.

A record is not a dictionary. The bug comes from commit 1431d67b.

Preparing for a fourth shapefile containing centers.

This occurs when we read results from the Matlab program.

The bug occurred when: `args.anim` is None; `shp_tr_dir` contains a
file `ishape_last.txt`; `ishape_last.txt` contains a single
value (there is a single date in `shp_tr_dir`). Then loadtxt without
the argument ndmin return a numpy scalar. Then we cannot call
`len(ishape_last)`, at line 286.

Remove fields interpolat, cyclone and valid from `extremum.dbf` and
field radius4 from `outermost_contour.dbf`. For now, we just want to
change the format of Matlab results, not add to the contents.

Read the path to first file and the last date from the command line.

Take into account renaming of files in f7075db7 and moving to new
directory in 125569d8.

`inst_eddies`: instantaneous eddies. `overlap`: overlapping.