*************** Models Metadata *************** The code generator's default output format is python `dataclasses `_ and ships with its own modules for xml and json marshalling. The generated models don't depend on the xsdata library, except for specific xsd data types that can't be mapped directly to python's builtin types. They only include some metadata properties in order to control how the data structures are transferred during data binding. .. hint:: The code generator supports processing xml documents directly. That means even without a schema you can easily create at the very least an initial draft of your models just from samples. If you use a directory with multiple samples the transformer will merge and flatten duplicate classes, fields and field types. .. versionadded:: 20.4 .. testsetup:: * from dataclasses import dataclass from dataclasses import field from decimal import Decimal from typing import List from xsdata.formats.dataclass.parsers import XmlParser Basic Model =========== .. testcode:: @dataclass class Currency: id: int = field(metadata=dict(type="Attribute", name="ID")) name: str = field(metadata=dict(name="Name")) num_code: int = field(metadata=dict(name="NumCode")) iso_code: str = field(metadata=dict(name="CharCode")) nominal: int = field(metadata=dict(name="Nominal")) value: Decimal = field(metadata=dict(name="Value")) @dataclass class Currencies: class Meta: name = "ValCurs" date: str = field(metadata=dict(type="Attribute", name="Date")) name: str = field(metadata=dict(type="Attribute")) values: List[Currency] = field(default_factory=list, metadata=dict(name="Valute")) xml = """ 978 EUR 1 Euro 19.2743 840 USD 1 US Dollar 17.7177 """ obj = XmlParser().from_string(xml, Currencies) print(obj.values[1]) .. testoutput:: Currency(id=44, name='US Dollar', num_code=840, iso_code='USD', nominal=1, value=Decimal('17.7177')) Class Metadata ============== Through the Meta class you can control the model's behaviour during data binding procedures. .. list-table:: :widths: 20 10 300 :header-rows: 1 * - Property - Type - Description * - name - str - The real/local name of the element this class represents. * - nillable - bool - Specifies whether an explicit empty value can be assigned, default: False * - namespace - str - The element xml namespace. * - element_name_generator - Callable - Element name generator * - attribute_name_generator - Callable - Attribute name generator Field Typing ============ Simply follow the Python lib `dataclasses `_ documentation. .. warning:: Currently only List, Dict and Union annotations are supported. Everything else will raise an exception as unsupported. Field Metadata ============== Through the metadata properties you can control the field's behaviour during data binding procedures. .. list-table:: :widths: 20 10 250 :header-rows: 1 * - Property - Type - Description * - name - str - The real/local name of the element or attribute this field represents. * - type - str - The field xml type: ``Ignore | Text | Element | Elements | Attribute | Wildcard | Attributes ``, default: ``Text`` or ``Element`` * - nillable - bool - Specifies whether an explicit empty value can be assigned. * - mixed - bool - Specifies whether the field supports mixed content. ([#M1]_) * - sequential - bool - Specifies whether the field value(s) must appear in sequence with other sequential sibling fields. eg ```` * - tokens - bool - Use a list to map simple values. eg ``element: List[Union[int, bool, str]] -> 1 a true -> [1, "a", True]`` * - namespace - str - Specifies the field xml namespace. ([#M2]_) * - format - str - Format option for types like datetime, or bytes, see :ref:`Data Types` The code generator adds also the field restrictions like `minLength` or `required` flag but currently they are only used to troubleshoot the code generator. .. [#M1] Mixed content must be combined ``Wildcard`` fields with type ``List[object]``. `w3schools `_ .. [#M2] It's a common practice in schema definitions to require elements to be qualified and attributes to be unqualified. ``Element`` fields with an omitted namespace inherit the namespace from the parent class/element and ``Attribute`` fields don't. If you need to break the namespace inheritance for ``Element`` fields set the namespace to an empty string ``namespace=""``. Type: Ignore ~~~~~~~~~~~~ This type will force the internal xml context instance to ignore the field during binding. Make sure your field is declared with `init=False` or with a default value otherwise data binding will fail. .. code-block:: python index: int = field( default_factory=int, init=False, metadata={"type": "Ignore"}, ) Type: Element ~~~~~~~~~~~~~ This type represents a traditional xml element and can be the building block and container for other elements, attributes, text or any combination of them. .. code-block:: python annotation: List[Annotation] = field( default_factory=list, metadata={ "name": "annotation", "type": "Element", "namespace": "http://www.w3.org/XML/2004/xml-schema-test-suite/", } ) .. code-block:: xml ... ... ... ... Type: Elements ~~~~~~~~~~~~~~ This type represents repeating xs:choice elements. It's a compound list field for elements and wildcards that can be used to preserve elements ordering between data marshalling. .. code-block:: python node_or_id_or_idref: List[object] = field( default_factory=list, metadata={ "type": "Elements", "choices": ( { "name": "node", "type": Type["Node"], }, { "name": "e1", "type": str, "nillable": True, }, { "name": "e2", "type": int, "namespace": "xsdata", }, ), } ) .. code-block:: xml a b 1 c 2 ... **Choice Metadata** .. list-table:: :widths: 20 10 250 :header-rows: 1 * - Property - Type - Description * - name - str - The real name of the element this choice represents. * - type - str - The field type hint. * - nillable - bool - Specifies whether an explicit empty value can be assigned. * - wildcard - bool - Specifies whether this is a ``Wildcard`` that can match any tag. * - tokens - bool - Use a list to map simple values. eg ``element: List[Union[int, bool, str]] -> 1 a true -> [1, "a", True]`` * - namespace - str - Specifies the field xml namespace. * - format - str - Format option for types like datetime, or bytes, see :ref:`Data Types` * - default - Any - Default value * - default_factory - Any - Default value factory .. warning:: Compound fields preserve elements ordering but instead the direct element name association is lost during marshalling. If the choices include multiple elements with the same type then it's actually impossible to map correctly values to elements. For that reason the xml parser will use the generic class :class:`~xsdata.formats.dataclass.models.generics.DerivedElement` to wrap values in order to maintain the original qualified name as well. If your compound field includes only unique types and you are working with a dataclass instance manually you can skip the usage of the wrapper as the xml serializer will try to match a type to a choice as well. ``obj.node_or_id_or_idref.extend(("a", "b", 1, "c", "2"))`` Type: Attribute ~~~~~~~~~~~~~~~ This type represents a traditional xml attribute. .. code-block:: python language: Optional[str] = field( default=None, metadata={ "name": "lang", "type": "Attribute", "namespace": "http://www.w3.org/XML/1998/namespace" } ) .. code-block:: xml Type: Wildcard ~~~~~~~~~~~~~~ This type represents ``xs:any`` elements or elements with type ``xs:AnyType``. Wildcards can have a normal uri namespace or use one of xml schema generics. .. list-table:: :widths: 25 220 :header-rows: 1 * - Namespace - Description * - ##any - element from any namespace is allowed * - ##other - element from any namespace other than the parent's namespace * - ##local - element must come from no namespace * - ##targetNamespace - element from the namespace of the parent can be present .. code-block:: python any_element: List[object] = field( default_factory=list, metadata={ "type": "Wildcard", "namespace": "##any", } ) This type of field accepts any primitive value or an another dataclass instance or a generic :class:`~xsdata.formats.dataclass.models.generics.AnyElement` instance. Type: Attributes ~~~~~~~~~~~~~~~~ This type represents ``xs:anyAttribute`` elements. It needs to be defined as a dictionary of. The wildcard namespace features also apply. .. code-block:: python any_attributes: Dict = field( default_factory=dict, metadata={ "type": "Attributes", "namespace": "##other" } ) Type: Text ~~~~~~~~~~ This is the default field type and represents any atomic value. The value of this field is directly assigned as text to elements. .. code-block:: python @dataclass class Root: class Meta: name = "root" value: Optional[int] = field(default=None) .. code-block:: xml 2020 .. hint:: Check the :ref:`examples ` for more advanced topics.