json-numpy

2.1.1 verified Mon Apr 27 auth: no python

A library for encoding and decoding NumPy arrays and scalars to/from JSON-compatible formats. Current version 2.1.1, compatible with Python >=3.8. Maintained actively.

pip install json-numpy

Common errors

error AttributeError: module 'json_numpy' has no attribute 'dumps' ↓

cause Incorrect import assuming dumps/loads are top-level; they are inside json_numpy module.

fix

Use json_numpy.dumps() and json_numpy.loads() after importing json_numpy.

error TypeError: Object of type ndarray is not JSON serializable ↓

cause Using standard json.dumps directly on numpy arrays.

fix

Use json_numpy.dumps(arr) instead of json.dumps(arr).

Warnings

gotcha json_numpy.dumps does not encode Infinity or NaN by default; they become null. To preserve them, pass allow_nan=True. ↓

fix Use json_numpy.dumps(obj, allow_nan=True).

gotcha Deserialization returns numpy arrays by default. To get Python lists, convert manually. ↓

fix Use json_numpy.loads(s, as_list=True) or explicit list() after decode.

Imports

json_numpy
```
import json_numpy
```
Direct import of the module for conversion functions.

Quickstart

Basic usage: encode NumPy array to JSON string and decode back.

import json_numpy
import numpy as np

arr = np.array([1, 2, 3])
encoded = json_numpy.dumps(arr)
print(encoded)
decoded = json_numpy.loads(encoded)
print(decoded)