`implement_array_function method already has a docstring` Minimal Reproducible Example

numfoo.pyx

cimport cython

import numpy as np
cimport numpy as np

np.import_array()

ctypedef cython.floating floating_t

ctypedef fused numeric_collection:
    np.ndarray
    list
    tuple
    object


cdef object double_dtype = np.float64


cdef np.ndarray[double] coerce_data(numeric_collection data):
    cdef np.ndarray npdata
    if numeric_collection is object:
        return np.array(list(data), dtype=np.float64)
    elif numeric_collection is list or numeric_collection is tuple:
        return np.array(data, dtype=np.float64)
    elif numeric_collection is np.ndarray:
        npdata = data
        if npdata.dtype != double_dtype:
            return npdata.astype(double_dtype)
        else:
            return npdata

def make_array(numeric_collection data):
    return coerce_data(data)

setup.py

from setuptools import setup, Extension

import numpy as np

from Cython.Build import cythonize

ext_modules = [
    Extension(
        "numfoo", ["numfoo.pyx"],
        language='c++',
        include_dirs=[np.get_include()]
    )
]

ext_modules = cythonize(ext_modules)

setup(name="numfoo", ext_modules=ext_modules)

test.py

import numfoo
import numpy as np

data = [1.2, 2.4]

# The data are a plain list
print(data)

# Invoking the fused function explicitly with a list works
print(numfoo.make_array[list](data))

# convert it into a NumPy array
data = np.array(data)

# Invoking the fused function explicitly with an np.ndarray works
print(numfoo.make_array[np.ndarray](data))

# This fails
print(numfoo.make_array(data))