FitManager: make uncertainty trigger the use of weighted fit

[kif]

related to #4590

• The PR title is formatted as: <Module or Topic>: <Action> <Summary> (see contributing guidelines)

PR summary

More intuitive handling of uncertainties

AI Disclosure

• No AI used
• AI tool(s) ... used for ...

[t20100]

From what I understand, your fix doe not correspond to theFitManager docstring:

    :param weight_flag: If this parameter is ``True`` and ``sigmay``
        uncertainties are not specified, the square root of ``y`` is used
        as weights in the least-squares problem. If ``False``, the
        uncertainties are set to 1.
    :type weight_flag: boolean

https://github.com/kif/silx/blob/afdef395d150e0052ec6c484793ea472ef9b8f66/src/silx/math/fit/fitmanager.py#L78-L82

-> If weight flag is False, uncertainties should be ignored...

[kif]

False is the default value, which is a mistake to me: it should be None and switch to False if no sigmay is provided.
When a user provides uncertainties, it should not crash, especially that I am following the example of the documentation ...

[t20100]

Then maybe good to add to the setdata docstring that setting sigmay enables WeightFlag. I agree it is expected but since there looks to be a few ways to set this best to make it explicit.

Did you check how this works from the fit widgets?

[kif]

No, I did not check it ... BTW, I doubt this code is used anywhere since the "split Pseudo-Voigt 2" mode could not work at all, since array are not initialized at the proper size.

The code-style is very PyMca-like, fit results would deserve to be dataclass or namedtuple instances, but this is not the use-case for this PR.

[vasole]

The code-style is very PyMca-like, fit results would deserve to be dataclass or namedtuple instances, but this is not the use-case for this PR.

I would prefer the expression "pre python2.6" (released in October 2008). PyMca could not use something that was not existing when it was written.

[kif]

I just checked, the GUI still works, and the bug about "Split pseudo-Voigt 2" is gone with this PR.

[t20100]

Thanks for checking the GUI and for looking into this!

Two minor comments, otherwise LGTM

BTW, I doubt this code is used anywhere

Provided the issues you fixed, probably...

[t20100]

Let's keep it private:

Suggested change

	class VersionInfo(NamedTuple):
	major: int
	minor: int
	micro: int
	releaselevel: str
	serial: int
	version_info = _version_info(MAJOR, MINOR, MICRO, RELEV, SERIAL)
	version_info = VersionInfo(MAJOR, MINOR, MICRO, RELEV, SERIAL)
	class _VersionInfo(NamedTuple):
	major: int
	minor: int
	micro: int
	releaselevel: str
	serial: int
	version_info = _VersionInfo(MAJOR, MINOR, MICRO, RELEV, SERIAL)