The solution is to use FunctionWrappers.jl which provides a high
interface for C function pointers.

All tests passed.

implemetend in the package.
A lot of files has been modified because some of the internal syntax has
changed. But the top level methods should still work.

Two tests about the plots of synchronized oscillatory simulations have
been added.

All tests passed.

allocation per next_state!. Now update_state! functions returns a Symbol
as location instead of modifying a vector.

check_constraints and update_state! functions of an Edge in order to
improve performance.
Acutally the gain is low, but it improved the readabilty of the code.
All tests passed.

All automata were readapted and are in accordance with the last description of the thesis. Tests passed

check_constraint and update_state functions are created inside
create_automaton_*(), by the evaluation of meta_elementary_functions <:
Expr. check_constraint and update_state don't depend on constants::Dict
anymore: constants are interpolated inside the expression.

Add of tests for euclidean distance automaton.
Add a new function euclidean_distance that computes the euclidean
distance of a Trajectory compared to observations along a discrete
timeline.

automata and proba moels are differents. This bug was introduced after
we change the structure of automata in order to better distributed
programming.
It leads to a restructuration of automata. In fine, the structure is
more stable and consistent. It is possible it is more computationally
efficient, but for now it is not well tested.

2) Creation of functions that collects data along with a timeline
over a trajectory + unit test vectorize.

3) Very small fix in plots and abc smc.

1) LHA now have names. 2) Period automaton executes well and was runned on the doping 3way oscillator model. Not stattistically tested yet

Simulation of synchronized models is easily distributable.
Change of the signature of check_constraints and update_state in order
to access the model parameters.
Test about distributed simulations was renamed and completed.

Change of signatures of unit functions update_state!, check_constraints
of the automata edges. Use getfield instead of getproperty.

Now the execution of statistical estimation of the distance G is of the
order of Cosmos !

models. It's a better semantic and improve performance as well as
readability of the code. All the tests passes (except the remark in the
last commit).

variableautomaton. It improves performance.

Tests of Cosmos doesn't pass on the bounds (with a parameter equal to
zero). It's a problem of design (there are issues for cosmos too).

of automata: it wasn't an issue about the simulation code of LHA but an
issue of LHA designs.

Improvement of a Cosmos utils method.

Now A.constants is a NamedTuple.

distributed_mean_value_lha now takes several variables.

Fix about behavior of automata when the initial state is absorbing or
the parameters are equal to zero.

Add of supplementary unit tests.

The two Cosmos tests checks the statistical correctness of estimated
distance value for R1,R2,R3,R5 experiments. Also, they test that all
trajectories are accepted.

A bunch of errors were fixed in automata files (Cosmos .lha and julia
./automata files) and simulation function that introduced statistical
bias, i.e. approximated values that were closed but not in the
confidence interval. The general structure of simulation and was
reworked in order to suits well te behavior of Cosmos.

copyto! is implemented for LHA states.

Automaton G works almost well, a bias in the distance is identified
compared to Cosmos when one of the model parameters is zero.

Implementation of volatile_simulate, which simulates SynchronizedModel
without saving values: it only returns the last LHA state of the
simulation.

Implementation of distributed computations of variable automata with
volatile_simulate.

Tests passed.

Automaton G is implemented and runs OK. Not tested with Cosmos call.

The next step of development is the restruction of data collection for
trajectories.

From now on: Trajectory.values::Matrix{Float64}, it will be
Trajectory.values::Vector{Vector{Float64}}.

According to small performance tests I've made, it will increase the
performance of the package.