Implemented test in respond to #15

[sami6mz]

Added a test to check whether the estimator is robut to (n,1) and (n,) inputs for t and y.

As for now :

• IPW returns aberrant indirect effects if not given .ravel inputs
• g_computation, simulation_based, G_estimator, DML_ returns ValueError
• multiply_robust is returning an error, but this will be rectified in Multiply robust #35
  This is issue need to correct IPW code for y and t input format #14 .

DML_huber also return an error but it's because dim_x=1, this is already declared in #16.

[bthirion]

The PR is good as such, but it may point to other limitations of the codebase.