Make Prior.__repr__ round-trippable for numpy scalar parameters

[lorenzopompili00]

With numpy 2.0+, repr() of numpy scalars carry an np. prefix (e.g. repr(np.float64(0.0)) is "np.float64(0.0)"). This goes into Prior.__repr__, and bilby's parser Prior._parse_argument_string then fails to re-parse it: it treatsnp as a module and calls import_module("np") which fails.

Can be reproduced by running

import numpy as np
from bilby.core.prior import Uniform, PriorDict

p = Uniform(minimum=np.float64(0.0), maximum=np.float64(1.0), name="x")
print(repr(p))          # Uniform(minimum=np.float64(0.0), maximum=np.float64(1.0), ...)
PriorDict({"x": repr(p)})   # ModuleNotFoundError: No module named 'np'

This fixes it by converting numpy scalars to native Python types in Prior.__repr__

[Copilot]

Pull request overview

This PR restores Prior.__repr__ round-trippability when prior parameters are NumPy scalar types (notably under NumPy >= 2.0 where scalar repr() includes an np. prefix), preventing PriorDict parsing from mistakenly attempting import_module("np").

Changes:

• Convert NumPy scalar values (np.generic) to native Python scalars via .item() when building Prior.__repr__.
• Add a regression test ensuring repr() output for Uniform with NumPy scalar bounds contains no np. and can be parsed back via PriorDict.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
bilby/core/prior/base.py	Ensures Prior.__repr__ emits native Python scalar literals for NumPy scalar parameters to keep from_repr / _parse_argument_string round-trippable.
test/core/prior/base_test.py	Adds coverage to prevent regressions in repr() round-tripping for priors constructed with NumPy scalar parameters.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.