pydantic
Pydantic is the most widely used data validation library for Python. Fast and extensible, Pydantic plays nicely with your linters/IDE/brain. Define how data should be in pure, canonical Python 3.7+; validate it with Pydantic.
Why pydantic?
- Powered by type hints
- Speed
- JSON Schema
- Strict and Lax mode
- Dataclasses, TypedDicts and more
- Customisation
- Ecosystem
- Battle tested
Usage
from datetime import datetime
from typing import Tuple
from pydantic import BaseModel
class Delivery(BaseModel):
timestamp: datetime
dimensions: Tuple[int, int]
m = Delivery(timestamp='2020-01-02T03:04:05Z', dimensions=['10', '20'])
print(repr(m.timestamp))
#> datetime.datetime(2020, 1, 2, 3, 4, 5, tzinfo=TzInfo(UTC))
print(m.dimensions)
#> (10, 20)
Complex data structure approaches
Comparison of three approaches on YouTube
- dataclasses - Part of standard Python.
- attrs - Third party, but a superset of dataclasses.
- Pydantic - Also third party, but integrated into FastAPI, so an easy choice if you’re already using FastAPI
Reasoning
One takeaway is that Pydantic stands out for validation, and its integration with FastAPI catches many potential data errors. Another is that Pydantic relies on inheritance (from the BaseModel class), and the other two use Python decorators to define their objects. This is more a matter of style.
In another comparison, Pydantic outperformed older validation packages like marshmallow and the intriguingly named Voluptuous. Another big plus for Pydantic is that it uses standard Python type hint syntax; older libraries pre-dated type hints and rolled their own.
Pydantic provides ways to specify any combination of these checks:
- Required versus optional
- Default value if unspecified but required
- The data type or types expected
- Value range restrictions
- Other function-based checks if needed
- Serialization and deserialization