BetterCRUD Changelog¶

Introduction¶

The Changelog documents all notable changes made to BetterCRUD. This includes new features, bug fixes, and improvements. It's organized by version and date, providing a clear history of the library's development.

[0.2.6] - Oct 11, 2026¶

improvements

Fix the license field: _version.py declared "Apache License 2.0" while the LICENSE file, the classifiers and the repository were MIT, so PyPI published an incorrect license
Improve the PyPI summary to lead with "FastAPI CRUD" and expand keywords with crudrouter, fastapi-crud, sqlmodel and rest so the package is discoverable from the terms people actually search for
Add classifiers for Application Frameworks, WWW/HTTP and AsyncIO, plus Changelog / Issue Tracker project URLs
Add the missing py.typed marker: the Typing :: Typed classifier was declared but inline types were not visible to mypy or pyright
Configure site_url and site_description for the docs: the published sitemap.xml was empty and pages had no canonical URL
Expand the test matrix from Python 3.9 only to 3.9-3.12 so CI covers the Python versions the package declares

[0.2.5] - Aug 18, 2026¶

feature

Add query.filters for declaring typed, OpenAPI-visible query parameters on get_many that automatically participate in WHERE (supports base fields and relationship fields such as company.name)
Validate query.filters at route registration: reject names colliding with reserved get_many parameters, and require relationship fields to be declared in query.joins with join=True

[0.2.4] - Aug 18, 2026¶

fixed issues

Use an explicit crud_action response_model as the final response shape as-is (passing ResponseModel[CurrentUser] no longer gets wrapped into ResponseModel[ResponseModel[CurrentUser]])

[0.2.3] - Aug 17, 2026¶

feature

Infer crud_action response model from the return type annotation when response_schema is configured (instead of always falling back to serialize.base)
Skip response schema wrapping when the endpoint already returns a response schema instance (no more double-wrapped payloads)
Preserve Optional[X] / X | None return annotations so None responses stay valid
Fix route registration crashing with NameError/TypeError when a crud_action return annotation cannot be resolved at import time (e.g. from __future__ import annotations referencing a non-global model); falls back to the bare response_schema (data unconstrained) and emits a warning

[0.2.2] - Aug 09, 2026¶

feature

Add crud_action decorator for custom endpoints
Make bulk create/update operations atomic (single transaction with rollback)
Add soft-delete recover route (PATCH /{id}/recover, opt-in via allow_recover)

[0.2.1] - Aug 09, 2026¶

feature

Support optional pagination mode for get_many (pagination_mode: always/optional/disabled)

[0.2.0] - Nov 10, 2025¶

fixed issues

$notany operator support empty

[0.1.8] - Jun 16, 2025¶

fixed issues

Add parms parameters only for crud routes

[0.1.7] - Jun 12, 2025¶

feature

Expose registered crud routes

[0.1.6] - Jun 8, 2025¶

fixed issues

BetterCrudGlobalConfig DEFAULT_ACTION_MAP value should be str,should not be an enumeration value

[0.1.5] - May 31, 2025¶

Support many-to-many relationships in object mode

class Zone(SQLModel, table=True):
    id: Optional[int] = Field(default=None, primary_key=True)
    name: Optional[str] = None
    description: Optional[str] = Field(default=None)
    is_active: Optional[bool] = Field(default=True)


class UserZoneLink(SQLModel, table=True):
    __tablename__ = "user_zone"
    user_id: Optional[int] = Field(
        default=None, foreign_key="user.id", primary_key=True
    )
    zone_id: Optional[int] = Field(
        default=None, foreign_key="zone.id", primary_key=True
    )



class User(UserBase, table=True):
    ...
    zone: Optional[Zone] = Relationship(
        sa_relationship_kwargs={
            "lazy": "noload", "uselist": False},
        link_model=UserZoneLink
    )
    ...

class UserCreate(UserBase):
    ...
    zone:Optional[int] = None

You can post zone data by passing zone_id client.post("/user", json={**user_data,"zone":zone_id})

[0.1.4] - Apr 21, 2025¶

RouteOptions add custom action

[0.1.3] - Apr 6, 2025¶

Add more filter operators
$startsL (starts with,not case sensitive )
$endsL (ends with,not case sensitive )
$contL (contains,not case sensitive )
$exclL (not contains,not case sensitive )
$eqL (=,not case sensitive )
$neL (!=,not case sensitive )
$inL (IN, in range, accepts multiple values separated by commas,not case sensitive )
$notinL (NOT IN, not in range, accepts multiple values separated by commas,not case sensitive )

[0.1.2] - Mar 30, 2025¶

Some design refactoring
Remove the crud_get_many parameters criteria, options, joins

[0.1.1] - Mar 26, 2025¶

Features¶

Add some parameters to crud_get_many method
criterions embedded query criteria
options load options

joins relationship joins

joins = [(UserProfile, UserProfile.id == User.profile_id)]
options = [joinedload(User.profile), joinedload(User.roles)]
criterions = [UserProfile.gender == "female"]
fetched_records = await user_service.crud_get_many(
    test_request,
    joins=joins,
    criterions=criterions,
    options=options,
    db_session=async_session
)

[0.1.0] - Mar 25, 2025¶

Added¶

Keep the s and filter query conditions not mutually exclusive and can be run in parallel

[0.0.9] - Mar 25, 2025¶

Added¶

query filter support lambda function,it can realize the purpose of query condition transformation

@crud(
    user_router,
    feature="user",
    query={
        "filter": lambda x: {"$and": [*x["$and"], {
            "is_active": True,
        }]}
    },
    serialize={
        "base": UserPublic,
    }
)

[0.0.8] - Mar 24, 2025¶

Fixed¶

routes options only set to empty list does not take effect

@crud(
    user_router,
    routes={
        "only": []
    }
)

[0.0.7] - Mar 11, 2025¶

Fixed¶

Support model alias in nested relationship

from sqlalchemy.orm import aliased

ModifierUser = aliased(User)

query={
    "joins": {
        "user": {
            "join": True,
            "select": True
        },
        "user.modifier": {
            "join": True,
            "alias": ModifierUser
        }
    }
}

[0.0.6] - Feb 16, 2025¶

Fixed¶

Sorting can't use the Related Table alias field
Under certain conditions, queries using alias tables are invalid

[0.0.5] - Feb 12, 2025¶

Added¶

Add model alias to query,avoid Not unique table/alias errors

Usage Example¶

from sqlalchemy.orm import aliased

ModifierUser = aliased(User)

class Post(PostBase, table=True):
    __tablename__ = "post"
    id: Optional[int] = Field(default=None, primary_key=True)
    creater_user_id: Optional[int] = Field(
        default=None, foreign_key="user.id"
    )
    modifier_user_id: Optional[int] = Field(
        default=None, foreign_key="user.id"
    )

    creater_user: User = Relationship(
        sa_relationship_kwargs={"uselist": False, "lazy": "noload", "foreign_keys": "[Post.creater_user_id]"})

    modifier_user: User = Relationship(
        sa_relationship_kwargs={"uselist": False, "lazy": "noload", "foreign_keys": "[Post.modifier_user_id]"})

@crud(
    router,
    dto={"create": PostCreate, "update": PostUpdate},
    serialize={"base": PostPublic},
    query={
        "joins": {
            "creater_user": {
                "select": True,
                "join": True
            },
            "modifier_user": {
                "select": True,
                "join": True,
                "alias": ModifierUser
            }
        },
    }
)
class PostController():
    service: PostService = Depends(PostService)