More exotic types

Hi,

At my company we would like to migrate to your package, but we need more types out of the box. Below you can find list of additional types and validators and I like to discuss with you what do you think:

1. Min and max length validators for list type, maybe something like conlist(type_=..., min_length=..., max_length=...). Of course I can implement length validation using validator decorator, but it’s too complex to share as common type for company needs.
2. (Done #175) Decimal type not supported at all. Also we would like to have validators for minimum, maximum, maximum digits and decimal places like in Django (https://github.com/django/django/blob/master/django/forms/fields.py#L322, https://github.com/django/django/blob/master/django/core/validators.py#L388).
3. (Done #166) Float type: minimum and maximum validators.
4. String type:
   1. (Done #163) We need additional strip_whitespace argument in constr to strip whitespaces, just value.strip(). In my opinion it’s common practice for validation libraries to have this option.
   2. Also I would like to discuss with you coerce logic (https://github.com/samuelcolvin/pydantic/blob/master/pydantic/validators.py#L30). Why it supports only for float, int and decimal types? Can we use logic which already implemented for dict and list types (https://github.com/samuelcolvin/pydantic/blob/master/pydantic/validators.py#L83)?
5. (Done #167) UUID type version validation, just value.version == ....

If you are okay with my proposals, then I will create a few PRs for each my proposal.

[samuelcolvin]

This is not what conlist is for, conlist is for controlling the size of the outer list.

You can achieve what you’re looking for here either with custom validators, a Point type that implements it’s own validation, or with a simple tuple:

from typing import List, Tuple, Union
from pydantic.dataclasses import dataclass

@dataclass
class Geometry:
    type: str
    coordinates: List[Union[Tuple[float, float], Tuple[float, float, float]]]

print(Geometry(type='Polygon', coordinates=[[0, 0], [10, 10], [10, 0, 0.5]]))

If you have anymore questions, please create a new issue.

[alex-lechner]

Oh I’m sorry I forgot to elaborate on my issue a bit more. In my case, I need to validate geometry objects based on the GeoJSON Specification. In my example above I validate a Polygon object. The structure of the Polygon in a GeoJSON has to look like this:

{
  "type": "Polygon",
  "coordinates": [
    [
      [0, 0], [10, 10], [10, 0], [0, 0]
    ]
  ]
}

As you can see the coordinates are a nested list with an exterior Part (second list) and Points (lists within the second list). The order of a Point described by the Geojson is [longitude, latitude, elevation]. So a Point can only have either 2 or 3 values in a list but not more or less. That’s why conlist might come in handy because the conlist would set a constraint on the list/array object.

Would you accept a PR for GeoJSON support? Also for the conlist?