Required since commit ed60f104.

From directory Analysis. Extract corresponding tests from
`Analysis/tests.json` into new file `Inst_eddies/Analysis/tests.json`.

Because we want to use the module parser from dateutil.

This is more robust: the default value of `max_radius_deg` remains
reasonable when we use input files with different resolutions.

I do not see a better way than to ask the user to provide the template
of basename as an additional argument. We cannot do as in
`Convert_Matlab/inst_eddies_v6.py`, that is try a list of templates,
because here the file names could be anything.

So we no longer have to create a symbolic link when we run
`inst_eddies.py`.

Preparing for configuration of location of binary executable.

Much simpler and more robust. This should also ease submitting the
script to a batch scheduler. And we can now check the return code of
Matlab.

Take advantage of version 2019 of Matlab now available on Ciclad,
which has option `-batch`. This clears the output and opens the way to
subprocess instead of `pty.spawn`.

I have found that that, with `LC_NUMERIC=fr_FR.UTF-8`, sort on edglist
does not work on Ciclad, which has sort version `8.4`. This would be
because of blank as thousands separator.

So we do not risk loading `adt_v6.mat` from a previous iteration if
Matlab fails in the current iteration.

This is useful when we run the script through batch scheduling.

It is not enough to parse the input file name and extract the date
because, when we process other dates, we have to construct file names
in the same format. So do not use `parser.parse` for the input file
name, try various formats and remember the one that works.

Method poly of shapefile writer requires a list of lists, not a list
of numpy arrays. This bug goes unnoticed in pyshp `2.1.0` but raises an
exception with pyshp `2.1.3`.

This is an almost-duplicate file of `report_graph.int`: the difference
is that it processes an edgelist with nodes identified by node index,
instead of date and eddy index. We could make `report_graph.py` more
general instead, but that would degrade its clarity. Let us just have this
almost duplicate file, pending a decision on the format of edgelist.

Move functions `node_to_date_eddy` and `date_eddy_to_node` from
`convert_id.py` to `report_graph_int.py`, where they are used.

Revert part of commit dfb78dbc. Motivation: being able to use
`plot_eddy_contours.py` on results from Matlab.

Change of terminology: date subscript to date or days since 1950.

Do not print arrays.

`cropped.nc` does not exist if not `args.bbox`.

We need to update this input directory because of the change of DBF
field `date_index` to `days_1950`.

Multiple bug fixes. New version of `TAGS.cmake` imported from the
repository LMDZE.

Update required following commit 05dd4fc1.

Following the update of `Region_6_SHP_triplet` in commit ecfaaa91.

Better to check this before the creation of all the temporary files
and links, so the user does not have to clean them up before running
again.

Following update of libraries: NetCDF-Fortran is private in NetCDF95
and MPI-Fortran is private in EZMPI.