feat: add plot_recipe method to FitRecipe

[cadenmyers13]

closes #142

[cadenmyers13]

@sbillinge ready for review, new plot_recipe method for srfit. I tested this on the Ni refinement and one where there are two fit contributions and it worked as expected

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 69.36%. Comparing base (ff4d668) to head (65a4c85).
⚠️ Report is 17 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #143      +/-   ##
==========================================
+ Coverage   66.60%   69.36%   +2.76%     
==========================================
  Files          25       25              
  Lines        3171     3457     +286     
==========================================
+ Hits         2112     2398     +286     
  Misses       1059     1059              

Files with missing lines	Coverage Δ
tests/conftest.py	91.48% <100.00%> (+4.82%)	⬆️
tests/test_fitrecipe.py	99.76% <100.00%> (+0.27%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[cadenmyers13]

An example plot

[sbillinge]

this looks great. Actually, more configurable than I kind of intended, but ok. A few questions:

1. make sure that we are using the group style-sheet.
2. Start all the docstring descriptions with The (so Label for y-axis becomes The label for the y-axis. Never don't do that (i.e. ALWAYS start descriptions with The, it always works.)
3. I like the examples. Very nice.
4. It would be quite a bit of typing, but do we want to hold the parameters as attributes of the class? That way the user could just update the attribute to customize their plot and it would stay persistent for all their plots until they reset it or reinstantiated?
5. Tests look good. Very comprehensive.
6. Do we want to move the "build_recipe" functions to conftest.py as fixtures? In that way simple test recipes will be available to all tests.

All in all, this is excellent.

[sbillinge]

I think we shouldn't need six, what is it being used for?

[cadenmyers13]

@sbillinge for instance checks like isinstance(x, six.string_types). This can be replaced with just str

[sbillinge]

in the plot there are no figure axis labels. Is there a ready way to guess these? I guess the challenge is to know what is being fit and what is on teh x-axis so it may be more trouble than it is worth.

[cadenmyers13]

@sbillinge ready for review

make sure that we are using the group style-sheet.

Done

Start all the docstring descriptions with The

Done

It would be quite a bit of typing, but do we want to hold the parameters as attributes of the class? That way the user could just update the attribute to customize their plot and it would stay persistent for all their plots until they reset it or reinstantiated?

I added a plot_options attribute dictionary and a function that can be used to update it. Users can set global defaults and still modify individual plots if they want. I wrote a few more tests for this too.

Do we want to move the "build_recipe" functions to conftest.py as fixtures? In that way simple test recipes will be available to all tests.

Done

[sbillinge]

I think this docstring needs to be moved somewhere, at least, deleted from here. But maybe method that sets these

[cadenmyers13]

@sbillinge I moved it to set_plot_defaults

[cadenmyers13]

in the plot there are no figure axis labels. Is there a ready way to guess these? I guess the challenge is to know what is being fit and what is on teh x-axis so it may be more trouble than it is worth.

@sbillinge I added functionality to do this just for PDF files. The PDFParser loads metadata and one of the metadata is filename. so if gr is in the filename it defaults the labels to PDF labels. The user-defined labels will override that default though

[sbillinge]

I merged. Please can we update the docs with, more or less what is in the docstrings which are very good and informational.

the code was a bit verbose, I am guessing we could have structured it better, but let's leave that for the future......don't let perfect be the enemy of good.