File support S3/localfilesystem

[Matthijz98]

Hi,

I build a small proof of concept for a project too small for adding all of Django.
But one thing that is a deal breaker for me is the support for file support with s3 support.
That would be a great feature to add.

[mr-fatalyst]

Hi @Matthijz98, thanks for the suggestion!

File storage isn't really an ORM concern. Oxyde maps Python objects to relational tables, and S3 is a separate layer. Not planned for the near future.

For now you can just store the object key as a regular field and handle uploads in your app code:

import aioboto3
from oxyde import Model, Field

class Document(Model):
    id: int | None = Field(default=None, db_pk=True)
    title: str
    s3_key: str  # e.g. "uploads/2026/abc.pdf"

    class Meta:
        is_table = True

async def upload(file_bytes: bytes, key: str) -> Document:
    session = aioboto3.Session()
    async with session.client("s3") as s3:
        await s3.put_object(Bucket="my-bucket", Key=key, Body=file_bytes)
    return await Document.objects.create(title="...", s3_key=key)

Or something more custom, like this. Leveraging Pydantic's computed_field to expose URLs without storing them:

from pydantic import computed_field
from oxyde import Model, Field

class Document(Model):
    id: int | None = Field(default=None, db_pk=True)
    title: str
    s3_key: str

    class Meta:
        is_table = True

    @computed_field
    @property
    def url(self) -> str:
        return f"https://my-bucket.s3.amazonaws.com/{self.s3_key}"

    async def read_content(self) -> bytes:
        session = aioboto3.Session()
        async with session.client("s3") as s3:
            obj = await s3.get_object(Bucket="my-bucket", Key=self.s3_key)
            return await obj["Body"].read()

    async def presigned_url(self, expires: int = 3600) -> str:
        session = aioboto3.Session()
        async with session.client("s3") as s3:
            return await s3.generate_presigned_url(
                "get_object",
                Params={"Bucket": "my-bucket", "Key": self.s3_key},
                ExpiresIn=expires,
            )

Keeps the ORM focused and gives you full control over the storage backend.