-
Estimator
-
for hawkes estim, not mandatory to have true values. in init, in the estim plotters.
-
relplot, in lineplot write dependencies.
-
scatter: write with ax and save
-
a function on line 345 to write, there is an example of how to do it.
-
function applied to a column (and create a new column)
-
change name of super draw to something else
- make readme tutorial about the whole pipeline training + selecting best model + prediction with the best model
-
-
APLOT. The 3D plot have a different logic, due to the fact that axis is 3D axis. How to deal with it ? 3DAPlot?
perhaps an idea, split APlot into: displayable plot -> grid_plot -> APlot.
the grid_plot would take care of all the parameters, and in particular, would handle the dict_plot_param, dict_ax... the issue with some parameters is that for dict_plot_param they are just given to the plotter. It should be filtered first. This is the reason to problems in 3D axis.
For example, the cmap is a pain. grid_plot could create a function that does this, everything a function needs to check before starting the inside: then one just needs to wrap all the function inside APlot. that is actually my main struggle: how to deal with function that can take any keyword argument and has a behavior depending on that? like colors, linewidth etc...
- change the nb_ax by index_ax.
- homogeneous input, not nb_ax then xx but xx then nb_ax. It should always be the same order. A possibility would be to put first the data, then the axis.
- change return to give an ax. such that one can continue drawing on an axis!
- what is happening with bis axis is a bit obscure. Let s clarify it. Not sure how.
- labels when both axis are on the same graph, see if there is any comment about it and change it to adapt the new behavior.
- verify that if I access a nb_ax, I also check that the number is correct!
- put the test at the bottom into a right test.
- when dict of parameters given, mention if some are unused.