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.

Better than the generic `convert_Matlab_v6.sh` because we do not need
all the variables of file `adt....mat`.

The bug comes from commit fb9652fb.

So we do not read the name of input `.mat` files from the command
line, we assume they follow a pattern, and we need the files to be
present in the current directory. Also, since the dates are specified
in the script, we check that each date is equal to the date stored
inside the corresponding `.mat` file.

Preparing for the loop on dates. The mat files will be open inside the
loop which will output shapes, so the output of namelist, which use
mat data will need to be after.

Separate the definition of the fields of shapefiles and the output of
the namelist from the output of the shapes. This will allow us to loop
on dates only for the output of the shapes. Create two sets of writers
so that we will be able to alternate between the two sets when we loop
on multiple dates. So we create a function `define_fields`, we replace
the argument `SHP_triplet` of function write by argument writers and
we create the dictionary factories.

This allows to run `plot_eddy_contours.py` on the results.

Useful for automated test. See also commit 92bc155d.

Replace `reader_extr`, `reader_outer`, `reader_m_s` by dictionary
readers. Clearer and more concise.

Motivation: ease plotting of several snapshots coming from different
triplets of shapefiles, on the same figure. We may want to do this if
we have anticyclones and cyclones in separated triplets of shapefiles.

In order to extract the code for this function from the main part of
the script, we have to modify the behaviour of the script a little:
when the file `ishape_last.txt` could not be read or there is a single
value in it then args.k, if not None, must be equal to k1, instead of
being ignored.

Replace arguments `matlab_data` and `eddy_orientation` by cell and
cyclone. Remove argument `date_index`: just use the global variable.

I do not really see the point any longer of importing this module
instead of running it as a script.

Cf. commit bbcfcc78.

To progressively replace the output of Matlab, we modify it by
steps. So, for now at least, we want to keep the Matlab indexing of
eddies, which is separate for cyclones and anticyclones. So we have to
create two triplets of shapefiles, one for cyclones and one for anticyclones.

The choice is made with a command line argument. As in commit 1e4d2ee1.

Since we now have separate graphs.

Group `new_unit` and open in the same procedure, that we rename from
`set_unit_edge` to `open_edge_file`. This removes the need for a
special version of `new_unit` and it is closer to the use of the
`new_unit` keyword in Fortran 2008.