Fix plot_lm and _z_scale for scipy 1.10

[OriolAbril]

Description

plot_lm used np.tile on the plotter lists which contain 4 elements, a string, two dicts and an
array, somehow coercing all this into a numpy array while doing some repetition of elements
to match x and y values. This always rose many warnings, it now breaks the code and therefore the
tests.

This PR updates this to use lists and itertools all the way to perform the same type of repetitions,
no array features were really used so I think it makes more sense to use lists.

I have also updated the code in _z_scale to work with scipy 1.10 and higher as the nan policy in rankdata has changed.

Checklist

• Does the PR follow official PR format?
• Does the PR include new or updated tests to prevent issue recurrence (using pytest fixture pattern)?
  It does not, tests are already broken due to this so it can't go back.
• Is the code style correct (follows pylint and black guidelines)?
• Is the fix listed in the Maintenance and fixes
  section of the changelog?

---

📚 Documentation preview 📚: https://arviz--2186.org.readthedocs.build/en/2186/

[codecov]

Codecov Report

Merging #2186 (681bad3) into main (3c82bdd) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #2186   +/-   ##
=======================================
  Coverage   89.96%   89.96%           
=======================================
  Files         119      119           
  Lines       12401    12408    +7     
=======================================
+ Hits        11156    11163    +7     
  Misses       1245     1245           

Impacted Files	Coverage Δ
arviz/plots/lmplot.py	96.84% <100.00%> (+0.10%)	⬆️
arviz/stats/diagnostics.py	98.97% <100.00%> (+0.01%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.