add predictive_finder method (experimental)

[aloctavodia]

This adds an experimental method for predictive elicitation, the user must provide a model with an initial guess for the prior parameter and a guess for the observed distribution. The method iteratively will try to get the prior predictive distribution closer to the target distribution. The user can visually "supervise" the process and decide to stop and get the new priors or keep iterating. It has been only tested on a few simple models and there are more tests and work to be done. But it looks like a good start.

[codecov-commenter]

Codecov Report

Merging #293 (f88ba8f) into main (d0f0cfb) will decrease coverage by 1.57%.
The diff coverage is 30.72%.

@@            Coverage Diff             @@
##             main     #293      +/-   ##
==========================================
- Coverage   82.49%   80.93%   -1.57%     
==========================================
  Files          32       34       +2     
  Lines        3937     4038     +101     
==========================================
+ Hits         3248     3268      +20     
- Misses        689      770      +81     

Files	Coverage Δ
preliz/predictive/__init__.py	100.00% <100.00%> (ø)
preliz/unidimensional/quartile_int.py	67.85% <100.00%> (+0.71%)	⬆️
preliz/predictive/ppa.py	10.20% <50.00%> (-0.52%)	⬇️
preliz/internal/parser.py	61.33% <25.00%> (-1.69%)	⬇️
preliz/internal/plot_helper.py	72.13% <75.00%> (-0.24%)	⬇️
preliz/internal/predictive_helper.py	22.58% <22.58%> (ø)
preliz/predictive/predictive_finder.py	15.95% <15.95%> (ø)