No support for pydantic-2

[stabacco]

It has been released, with huge performance boosts:

https://docs.pydantic.dev/2.0/migration/

[cvarner-shippo]

Any movement here? Lack of pydantic 2 support currently makes this library incompatible with FastAPI

[necaris]

Unfortunately neither of the current maintainers is using FastAPI on a regular basis currently, so we don't have a lot of time to port things. We'd be glad to accept and help with any potential PRs, though!

[xalteropsx]

@necaris isnt django also uses ninja api which use pydantic v2 as well

[necaris]

@xalteropsx Yes, I believe django-ninja also uses Pydantic v2 at its most recent version. Neither of us is using that, either, so it doesn't change how much time we have to work on this. We're still happy to accept any potential PRs!

[xalteropsx]

@necaris i able to solved it do i share u the code ? it just 7line will u update it on ur repo ?

[necaris]

@xalteropsx yes, please share the code! We'd be very happy to accept a PR.

[xalteropsx]

give me 1-2day just gonna complete some office work then i will do it by taking pr else i will share u it

[xalteropsx]

@necaris i read ur library is not supported properly pydantic v2 as u can see type is now rename to annoatation and alias to fieldname here how i manage to make it work

> 
@necaris i take a look alot but i found it really hard cause alot of thing not yet docs on pydantic migration 


but here some example 

```py

def Pydantic_Graphene(model, exclude_fields=None):
    if exclude_fields is None:
        exclude_fields = []

    fields = {}
    for field_name, field in model.model_fields.items():
        if field_name not in exclude_fields:
            if field.annotation == str or field.annotation == Optional[str]:
                field_type = String
            elif field.annotation == int or field.annotation == Optional[int]:
                field_type = Int
            elif field.annotation == float or field.annotation == Optional[float]:
                field_type = Float
            elif field.annotation == bool or field.annotation == Optional[bool]:
                field_type = Boolean
            elif field.annotation == List[str] or field.annotation == Optional[List[str]]:
                field_type = List(String)
            elif field.annotation == List[int] or field.annotation == Optional[List[int]]:
                field_type = List(Int)
            elif field.annotation == List[float] or field.annotation == Optional[List[float]]:
                field_type = List(Float)
            elif field.annotation == List[bool] or field.annotation == Optional[List[bool]]:
                field_type = List(Boolean)
            
            fields[field_name] = field_type(required=field.is_required)

    return type(model.__name__ + 'ObjectType', (ObjectType,), fields)

UserObjectType = Pydantic_Graphene(User_Pydantic)

[xalteropsx]

[abhinand-c]

@necaris We're working on bringing Pydantic V2 to graphene-pydantic, and do have a working implementation but had to drop support for Pydantic V1 (Reducing backward compatibility).

@mak626 can throw more light on our progress.

[mak626]

@necaris We're working on bringing Pydantic V2 to graphene-pydantic, and do have a working implementation but had to drop support for Pydantic V1 (Reducing backward compatibility).

@mak626 can throw more light on our progress.

Here is our test release supporting pydantic v2
v0.5.0-alpha.2

Changelog

• Added support for pydantic v2
  • Change removed ModelField -> FieldInfo
  • pydantic.parse_obj_as(FooBar, data["listFooBars"]) was changed to new syntax TypeAdapter(FooBar).validate_python(data["createFooBar"])
• Dropped support for pydantic v1 completely.
• Support for python11 Union using pipe syntax
• Added support for bson ObjectId -> graphene.ID
• Added support for dict -> JSONString
• Added default is_type_of resolvers. This simplifies using UnionType as is_type_of need not be defined individually in the base types of Union

Pydantic V2 also changed the way we define fields. Please refer their doc Migration Guide. Some of the breaking changes include Optional[type] no longer has a default value None unless given explicitly.

[necaris]

@mak626 @abhinand-c sounds like we will need a new major version & branch to support both. Would you be willing to open a PR from your test release?

[abhinand-c]

@necaris Sure, will open a PR soon.