Modify API fitting tutorial

[Astro-Kirsty]

This PR is to address https://github.com/gammapy/gammapy/issues/5291#issuecomment-2163313687

We want to avoid accessing datasets.parameters[name] but rather take directly from the model

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 94.48%. Comparing base (64cdfa2) to head (969565f). Report is 1 commits behind head on main.

:exclamation: Current head 969565f differs from pull request most recent head bba6072

Please upload reports for the commit bba6072 to get more accurate results.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #5347   +/-   ##
=======================================
  Coverage   94.48%   94.48%           
=======================================
  Files         232      232           
  Lines       35742    35742           
=======================================
  Hits        33772    33772           
  Misses       1970     1970           

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[Astro-Kirsty]

I am also going to adjust the plotting of the stat_contour because these lines can be confusing for the user

[Astro-Kirsty]

https://astro-kirsty.github.io/gammapy-docs-preview/tutorials/api/fitting.html#sphx-glr-tutorials-api-fitting-py

[bkhelifi]

I have a side question: the discusssion outcomes was not to use the FitResult object within the tutorials?

[Astro-Kirsty]

I thought the end decision was to use datasets.models or fit_result.models

[AtreyeeS]

I have a side question: the discusssion outcomes was not to use the FitResult object within the tutorials?

@bkhelifi : Thanks to @QRemy recent efforts, datasets.models is now complete. We can, of course, expose better fit_results as a matter of habit, but it is not urgent