TopolearnAlmost every web API, webhook and config file speaks JSON.
JSON (JavaScript Object Notation) is a text format for structured data: objects ({"key": value}), arrays ([1, 2]), strings, numbers, true, false and null.
The json module has two core functions for strings:
json.dumps(obj)("dump string") serialises a Python object to a JSONstr.json.loads(text)("load string") parses JSON text into Python objects. It also acceptsbytesin UTF-8, UTF-16 or UTF-32, so you can pass an HTTP response body directly.
The result of loads is made of plain Python types: dicts, lists, str, int, float, bool and None. There are no special JSON objects to learn; you index into the result like any other dict or list.
A round trip json.loads(json.dumps(x)) gives back an equal value for data made of these basic types, with a few exceptions covered in the type-mapping lesson (tuples become lists, non-string keys become strings).
JSON syntax is stricter than Python's: strings use double quotes, there are no trailing commas, no comments, and the literals are lowercase true/false/null. So str(my_dict) is not JSON; "{'a': True}" fails to parse. Always produce JSON with json.dumps.
Example
import json
user = {"name": "Ann", "age": 31, "admin": False, "tags": ["a", "b"], "boss": None}
text = json.dumps(user)
print(text)
# {"name": "Ann", "age": 31, "admin": false, "tags": ["a", "b"], "boss": null}
print(type(text).__name__) # str
data = json.loads('{"id": 7, "items": [1.5, 2], "ok": true}')
print(data["items"][0], data["ok"]) # 1.5 True
print(json.loads(b'[1, 2, 3]')) # [1, 2, 3]
print(str({"a": True})) # {'a': True} -- not JSON!Remember
dumpsmakes a JSON string;loadsparses one into dicts, lists and scalars- JSON uses double quotes and lowercase true/false/null
str(dict)is not JSON; always usejson.dumps
Watch out for
- Building JSON with f-strings or
str(); quoting and escaping break as soon as a value contains a quote.