Describe the issue:

On x86_64, NumPy ints cannot be serialized with json, unlike Python ints. You should implement a default method to fix this.

Reproduce the code example:

import json
import numpy as np

A = np.zeros(10, dtype=int)
with open("myfile.txt", "w") as f:
    json.dump({'number': A[0]}, f)

Error message:

Traceback (most recent call last):
  File "/home/chillenb/src/fcdmft/serialexample.py", line 6, in <module>
    json.dump({'number': A[0]}, f)
  File "/home/chillenb/miniforge3/envs/fcdmftenv/lib/python3.11/json/__init__.py", line 179, in dump
    for chunk in iterable:
  File "/home/chillenb/miniforge3/envs/fcdmftenv/lib/python3.11/json/encoder.py", line 432, in _iterencode
    yield from _iterencode_dict(o, _current_indent_level)
  File "/home/chillenb/miniforge3/envs/fcdmftenv/lib/python3.11/json/encoder.py", line 406, in _iterencode_dict
    yield from chunks
  File "/home/chillenb/miniforge3/envs/fcdmftenv/lib/python3.11/json/encoder.py", line 439, in _iterencode
    o = _default(o)
        ^^^^^^^^^^^
  File "/home/chillenb/miniforge3/envs/fcdmftenv/lib/python3.11/json/encoder.py", line 180, in default
    raise TypeError(f'Object of type {o.__class__.__name__} '
TypeError: Object of type int64 is not JSON serializable

Python and NumPy Versions:

2.0.2
3.11.9 | packaged by conda-forge | (main, Apr 19 2024, 18:36:13) [GCC 12.3.0]

Runtime Environment:

[{'numpy_version': '2.0.2',
'python': '3.11.9 | packaged by conda-forge | (main, Apr 19 2024, 18:36:13) '
'[GCC 12.3.0]',
'uname': uname_result(system='Linux', node='pizza', release='6.10.10-arch1-1-custom', version='#1 SMP Tue, 22 Oct 2024 13:11:16 +0000', machine='x86_64')},
{'simd_extensions': {'baseline': ['SSE', 'SSE2', 'SSE3'],
'found': ['SSSE3',
'SSE41',
'POPCNT',
'SSE42',
'AVX',
'F16C',
'FMA3',
'AVX2',
'AVX512F',
'AVX512CD',
'AVX512_SKX',
'AVX512_CLX'],
'not_found': ['AVX512_KNL',
'AVX512_KNM',
'AVX512_CNL',
'AVX512_ICL',
'AVX512_SPR']}},
{'filepath': '/home/chillenb/miniforge3/envs/fcdmftenv/lib/libmkl_rt.so.2',
'internal_api': 'mkl',
'num_threads': 48,
'prefix': 'libmkl_rt',
'threading_layer': 'intel',
'user_api': 'blas',
'version': '2023.2-Product'},
{'filepath': '/home/chillenb/miniforge3/envs/fcdmftenv/lib/libomp.so',
'internal_api': 'openmp',
'num_threads': 48,
'prefix': 'libomp',
'user_api': 'openmp',
'version': None}]

Context for the issue:

This is bad because it will require lots of casting to fix, to the detriment of code readability.

It'd be nice if NumPy had a type that behaved exactly like Python ints!