SqlAlchemy support for Ex-DRF

exdrf-al is the SQLAlchemy 2.x bridge for exdrf: declarative metadata is turned into ExDataset / ExResource / ExField trees used by generators (al2pd, al2qt, al2r, …), Qt (exdrf-qt), and Alembic helpers (Click commands in this package). It depends on exdrf and exdrf-pd.

The library provides guidance and support code for using SQLAlchemy with Ex-DRF.

Usage

Start by basing your models off Base from exdrf_al.base module. It is a rather simple class that allows for model and fields iteration. If you don't want to use it, take a look at the implementation of that class for how to extract the information you need from your models.

Models

The library reads the description of models from the sqlAlchemy metadata. On top of that you can add additional information by adding members to the info key in the __table_args__ model attribute.

The keys that are used inside the info dictionary are documented in exdrf.resource.ResExtraInfo, which is the pydantic class that parses the information. Here is brief description of the most important keys:

• label: takes a string that is interpreted by the DSL in exdrf. label_dsl; it is used to generate code that computes the label of a record given the record. The exdrf.label_dsl module includes code to create python and typescript code from a parsed label definition.

Example:

class Example(Base):
    __tablename__ = "the_table"
    __table_args__ = {
        "info": {
            "label": """(if name name (concat "ID:" " id))""",
        }
    }
    id: int
    name: Optional[str] = None

Fields

Just like with the models, the model field metadata is used to extract the information. Additionally, we use values in the dictionary assigned to the info attribute to add additional information.

The keys that are used inside the info consist of a common subset for all field types and specific keys for each field type. The common keys parsed by the exdrf.field.FieldInfo class are documented with that class, and each field type is described in its own module under exdrf.field_types.

Example:

class Example(Base):
    __tablename__ = "the_table"
    id: Mapped[int] = mapped_column(
        primary_key=True,
        doc="The unique identifier for a record within its table",
        info={
            "sortable": False
        }
    )
    name: Mapped[Optional[str]] = mapped_column(
        doc="The name of the record",
        info={
            "filterable": False,
            "multiline": True,
            "min_length": 3,
        }
    )

Relationships

relationship(..., info={...}) is validated with exdrf.field_types.ref_base.RelExtraInfo, which extends exdrf.field.FieldInfo. Keys such as direction and doc are required or conventional; optional keys from FieldInfo apply as well.

When use_rel is true on a non-bridge OneToMany relationship, Qt code generation (exdrf-gen-al2qt) emits a dedicated editor tab with a DrfRelated-style transfer list (Qt{Parent}Rel{Ref}) instead of an inline Qt{Ref}MuSe control on category tabs.

Example:

items: Mapped[Set["Item"]] = relationship(
    "Item",
    back_populates="owner",
    collection_class=set,
    info={
        "doc": "Owned items",
        "direction": "OneToMany",
        "use_rel": True,
    },
)