Can't use type(numpy.float64) for MagneticMultipoleP parameter

[ruisardkj]

Creating a lattice file, if the type for the Bn1 parameter in the MagneticMultipoleP class is numpy.float64, the yaml output is garbled. For other classes this type is fine. If Bn1 is type=float, everything is fine. json output is fine either way. An example is attached. SNS-BTF-xml-to-PALS.zip

example of yaml output:

- MEBT:QH01:
              kind: Quadrupole
              ApertureP:
                x_limits:
                - null
                - null
                y_limits:
                - null
                - null
                shape: ELLIPTICAL
                location: ENTRANCE_END
                material: ''
                thickness: 0.0
                aperture_shifts_with_body: false
                aperture_active: true
              length: 0.061
              MagneticMultipoleP:
                Bn1: !!python/object/apply:numpy._core.multiarray.scalar
                - &id001 !!python/object/apply:numpy.dtype
                  args:
                  - f8
                  - false
                  - true
                  state: !!python/tuple
                  - 3
                  - <
                  - null
                  - null
                  - null
                  - -1
                  - -1
                  - 0
                - !!binary |
                  AAAAAAAAOsA=

Also ApertureP limits are not set properly, but still have not ruled out user error.

[adb-xkc85723]

This is a common issue with dumping numpy objects to YAML, probably coming from this line. It can be fixed in two simple ways (that I know of):

1. Add a SafeDumper to yaml:

import yaml
import numpy as np

yaml.SafeDumper.add_representer(
    np.float64, lambda d, v: d.represent_float(float(v))
)

But then you need to do this for all conceivable such objects.

2. Add a field_serializer for specific fields or a model_serializer for the full element model:

    @field_serializer("*")
    def serialize(self, v):
        if isinstance(v, np.generic):
            return v.item()
        if isinstance(v, np.ndarray):
            return v.tolist()
        return v