ruamel.yaml.representer.RepresenterError - Why ruaml.yaml can't represent an np.array?

I'm new using Pydantic and ruaml.yaml, working actually on a project using these two.

I have an error when I try to load a model config and represent it with mkdocs, It appear that ruaml.yaml is unable to dump on np.array object of my Pydantic schema.

File "/usr/local/lib/python3.10/site-packages/ruamel/yaml/representer.py", line 337, in represent_undefined
                raise RepresenterError(f'cannot represent an object: {data!s}')
            ruamel.yaml.representer.RepresenterError: cannot represent an object: [0 0 0 0]

It happen when I try to represent that :

std: Optional[np.ndarray] = np.array([1.0, 1.0, 1.0, 1.0])
mean: Optional[np.ndarray] = np.array([0, 0, 0, 0])

Is there a solution or something that I can change ?
Thanks in advance !

TEST :
I try to replace it by :

std: Optional[np.ndarray] = [1.0, 1.0, 1.0, 1.0]
mean: Optional[np.ndarray] = [0, 0, 0, 0]

But need these to be np.array object for the rest of my code..

YAML has several types defined in the Language Independent types for YAML 1.1.
numpy.array is not in that list, primarily because it is not language independent, so there is no representation for something non-standard like that.

So you will have to provide a representer yourself, and if you represent it with a tag, you can load back the resulting representation if you provide a constructor for that tag.

It doesn't suffice to make a representer for numpy.array though, as neither std nor mean is of that type, as
you already indicate yourself by the type information you supply. You have to provide a representer for numpy.ndarray,
and also for numpy.float64 (for the elements of std and numpy.int (for the elements of mean)

Once you do that you can dump std and mean.

import sys
from pathlib import Path
import numpy
import ruamel.yaml

path = Path('numpy.yaml')

np = numpy
data = dict(
  std = np.array([1.0, 1.0, 1.0, 1.0]),
  mean = np.array([0, 0, 0, 0]),
)

def represent_numpy_float64(self, value):
    return self.represent_float(value)  # alternatively dump as a tagged float

def represent_numpy_int64(self, value):
    return self.represent_int(value)  # alternatively dump as a tagged int

def represent_numpy_array(self, array, flow_style=None):
    tag = '!numpy.ndarray'
    value = []
    node = ruamel.yaml.nodes.SequenceNode(tag, value, flow_style=flow_style)
    for elem in array:
        node_elem = self.represent_data(elem)
        value.append(node_elem)
    if flow_style is None:
        node.flow_style = True
    return node


yaml = ruamel.yaml.YAML()
yaml.Representer.add_representer(numpy.ndarray, represent_numpy_array)
yaml.Representer.add_representer(numpy.float64, represent_numpy_float64)
yaml.Representer.add_representer(numpy.int64, represent_numpy_int64)
yaml.dump(data, path)
print(path.read_text(), end='')  # the file ends in a newline, don't add another one

which gives:

std: !numpy.ndarray [1.0, 1.0, 1.0, 1.0]
mean: !numpy.ndarray [0, 0, 0, 0]

ruamel.yaml's default round-trip loader, can load the resuling file:

import sys

yaml = ruamel.yaml.YAML()
rtd = yaml.load(path)
print(f'{rtd}')
print('type of std', type(rtd['std']))
yaml.dump(rtd, sys.stdout)

which gives:

{'std': [1.0, 1.0, 1.0, 1.0], 'mean': [0, 0, 0, 0]}
type of std <class 'ruamel.yaml.comments.CommentedSeq'>
std: !numpy.ndarray [1.0, 1.0, 1.0, 1.0]
mean: !numpy.ndarray [0, 0, 0, 0]

as you can see, this doesn't load the file as a numpy.ndarray. For that you need to provide a constructor,
for the tag (since we don't dump the floats/ints tagged, you don't need to provide a constructor for those):

def construct_numpy_array(self, node):
    return numpy.array(self.construct_sequence(node))

yaml.Constructor.add_constructor('!numpy.ndarray', construct_numpy_array)

rtd = yaml.load(path)
print(f'{rtd}')
print('type of std', type(rtd['std']))

which gives:

{'std': array([1., 1., 1., 1.]), 'mean': array([0, 0, 0, 0])}
type of std <class 'numpy.ndarray'>

Note that the above doesn't work for (indirectly) recursive dumps, as the construct_numpy_array
does not have a yield statement. It would be better to do something like:

def construct_numpy_array(self, node):
    data = numpy.array([])
    yield data
    numpy.append(data, numpy.array(self.construct_sequence(node)))

but that did not give the same result (because of my unfamiliarity with numpy).