Why doesn't FieldGenerator have a null constructor????!

[The9Cat]

Main change

Added a null constructor to FieldGenerator so that one can instantiate for use with non-grid features, e.g. using the points() member, without having to supply dummy grid values.

[Some comment and docstring fixes came along for the ride.]

Details

Add the null constructor to FieldGenerator.H and exposed by pybind11 in FieldWrappers.cc.

Tested in an existing work flow that I can do:

fields = pyEXP.field.FieldGenerator()
radii, array = fields.histo1dlog(reader, rmin, rmax, nbin)

Should work for other cases, such as fields.points().

[michael-petersen]

Looks good to me -- a straightforward fix with the null constructor, and always helpful to have the docs changes. Again bumping version since this is good thing to add to main (although I guess we could have made the case for adding to level, as it is sort of feature-like).

[Copilot]

Pull Request Overview

Adds a no-argument constructor for FieldGenerator so it can be used without supplying grid parameters, updates Python bindings, adjusts a density function’s normalization, and bumps the project version.

• Introduce FieldGenerator() default constructor in C++ and expose via pybind11
• Update DiskDensityFunc normalization factor and exponent scaling
• Bump project version to 7.8.6

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
pyEXP/FieldWrappers.cc	Added pybind11 null constructor and docstring
include/DiskDensityFunc.H	Adjusted exponent factor and normalization
expui/FieldGenerator.H	Introduced FieldGenerator() default constructor
CMakeLists.txt	Updated project version to 7.8.6

Comments suppressed due to low confidence (2)

pyEXP/FieldWrappers.cc:377

• Unresolved merge conflict markers are present here; please remove <<<<<<<, =======, and >>>>>>> lines and reconcile the conflicting blocks.

<<<<<<< HEAD

expui/FieldGenerator.H:40

• [nitpick] Consider adding unit tests for the new default constructor to verify that methods like points() and histo1dlog() behave correctly without grid inputs.

FieldGenerator() = default;

[The9Cat]

I agree that this could be considered a new feature. But I'd argue that it's a bug of omission in the sense that member-function semantics demand a null constructor.

[michael-petersen]

Yes, this does seem more like bug than feature. And certainly doesn't break anything. So I'm bringing it in now with the updated version.