It would be nice if the repositories could understand the relationship between models.

Such a neat project that reduces the mental burden on developers.
This reminds me a similar project walrus.
I'm curious how Redbird will handle Redis.

The available pull requests validate pydantic v2 and increment features on current library version. May someone review them? This will be very useful on my API.

Best regards.

from redbird.ext import MemoryRepo
repo = MemoryRepo()

change by:

from redbird.repos import MemoryRepo
repo = MemoryRepo()

from scheduler import app as app_rocketry

raise error:

raise PydanticUserError('`const` is removed, use `Literal` instead', code='removed-kwargs')
pydantic.errors.PydanticUserError: `const` is removed, use `Literal` instead

and it's in file: redbird/base.py, line 153

ordered: bool = Field(default=False, const=True)

need to comment the json["_id"] out for it to work as self.id_field is none and likely the same for all the operations

def item_to_data(self, item:BaseModel):
json = self.item_to_dict(item, exclude_unset=False)
# Rename whatever is as id_field to _id
#json["_id"] = json.pop(self.id_field)
return json

Awesome work!

I tried redbird with mongorepo and noticed that the id_field is not optional.
Is there any reason behind that?

The id_field is set to _id field on mongo which turns of some features of mongo (mongo generates faster the _id and makes it easier to sort).

I assume that this is because ObjectId from mongo is not json parsable?

Also, I didn't found any options to use limit and skip together for pagination (there is just limit).

First of all, kudos for making this library! I'm evaluating the usages in one of my project, it can use a better separation between the data layer and the logic!

I think that one thing that is missing is the add_many method to leverage the (often more performant) backend methods.

The base implementation can be simply a series of calls to the add method.

Copying here an issue posted originally on Rocketry, just for the integrity and the tracking - since the problem stems from red-bird.
Miksus/rocketry#210

Is your feature request related to a problem? Please describe.
red-bird is not compatible with Pydantic V2

Describe the solution you'd like
Update Red-Bird (if it is still maintained) to use Pydantic V2

Describe alternatives you've considered
None

Additional context
Pydantic V2 was released on June 30th, and ever since red-bird and Rocketry are not usable.
Traceback for example:

Traceback (most recent call last):
  File "/app/main.py", line 5, in <module>
    from rocketry import Rocketry
  File "/usr/local/lib/python3.9/site-packages/rocketry/__init__.py", line 1, in <module>
    from .session import Session
  File "/usr/local/lib/python3.9/site-packages/rocketry/session.py", line 18, in <module>
    from rocketry.log.defaults import create_default_handler
  File "/usr/local/lib/python3.9/site-packages/rocketry/log/defaults.py", line 1, in <module>
    from redbird.logging import RepoHandler
  File "/usr/local/lib/python3.9/site-packages/redbird/__init__.py", line 2, in <module>
    from .base import BaseRepo, BaseResult
  File "/usr/local/lib/python3.9/site-packages/redbird/base.py", line 116, in <module>
    class BaseRepo(ABC, BaseModel):
  File "/usr/local/lib/python3.9/site-packages/redbird/base.py", line 153, in BaseRepo
    ordered: bool = Field(default=False, const=True)
  File "/usr/local/lib/python3.9/site-packages/pydantic/fields.py", line 675, in Field
    raise PydanticUserError('`const` is removed, use `Literal` instead', code='removed-kwargs')
pydantic.errors.PydanticUserError: `const` is removed, use `Literal` instead

Building on the JSONDirectoryRepo framework, I implemented a model with a Pydantic field_serializer to convert a Decimal value to string. That worked well when calling the BaseRepo methods to add or replace files in the repository, but I turned up an unusual error when attempting to call update. I haven't tried to track it down more because I am using replace as a workaround.

From what I can tell, it appears that the update call attempts to serialize the model twice, causing a failure when the field_serializer is handed a string instead of the expected Decimal value. That points to a likely unnecessary call to item_to_dict somewhere along the way, probably in the JSONDirectoryRepo override of update.

I'm new to the codebase, and using @ManiMozaffar's fork to support Pydantic v2 - but the error appears to be in the mainline codebase. I'll attach the traceback output as a comment for reference.