add calculation of dely for components of a composite model

[newville]

This adds calculation of dely for model components to ModelResult.eval_uncertainties, addressing #761

It replaces #826

Type of Changes

• Bug fix
• New feature
• Refactoring / maintenance
• Documentation / examples

Tested on

Python: 3.9.13 | packaged by conda-forge | (main, May 27 2022, 17:01:00)
[Clang 13.0.1 ]

lmfit: 1.0.3.post76+g44eb363.d20221120, scipy: 1.9.3, numpy: 1.23.4, asteval: 0.9.28, uncertainties: 3.1.7

Verification

Have you

• included docstrings that follow PEP 257?

• referenced existing Issue and/or provided relevant link to mailing list?

• verified that existing tests pass locally?

• verified that the documentation builds locally?

• squashed/minimized your commits and written descriptive commit messages?

• added or updated existing tests to cover the changes?
• updated the documentation and/or added an entry to the release notes (doc/whatsnew.rst)?
• added an example?

[codecov]

Codecov Report

Merging #831 (997c0bd) into master (44eb363) will decrease coverage by 0.02%.
The diff coverage is 92.30%.

@@            Coverage Diff             @@
##           master     #831      +/-   ##
==========================================
- Coverage   93.71%   93.68%   -0.03%     
==========================================
  Files           9        9              
  Lines        3483     3501      +18     
==========================================
+ Hits         3264     3280      +16     
- Misses        219      221       +2     

Impacted Files	Coverage Δ
lmfit/model.py	91.19% <92.30%> (-0.06%)	⬇️

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

[reneeotten]

arggghhh.... seuptools is breaking things again :( It looks like NumPy pins their version to 59.2.0, but even if I try that I still see failures in the CI - similar to here in this PR.

I'll try and fix that, so don't worry about / waste time with that part - hopefully soon and then I'll rebase/push to this PR and it should be good to go...

[newville]

@reneeotten I kind of think we actually have a serious problem here. When I see errors my default reaction has become "oh, something in the test setup is in error", either some value in our own tests (as with 99b13a4 or with the CI system itself.

I trust that running the tests locally detect will identify problems with the code. But it kind of seems like passing the Azure tests is more about "getting the Azure tests to pass" and rarely has anything to do with our code.

Honestly, that could all be restated as: I have lost confidence in our tests. That is not good.

I am going to merge this PR and release 1.0.4 by Nov 28. Tests that are failing before they even attempt to install lmfit will be removed.