Naive XML Bindings for python#

xsData is a complete data binding library for python allowing developers to access and use XML and JSON documents as simple objects rather than using DOM.

The code generator supports XML schemas, DTD, WSDL definitions, XML & JSON documents. It produces simple dataclasses with type hints and simple binding metadata.

The included XML and JSON parser/serializer are highly optimized and adaptable, with multiple handlers and configuration properties.

xsData is constantly tested against the W3C XML Schema 1.1 test suite.

Getting started#

$ # Install all dependencies
$ pip install xsdata[cli,lxml,soap]
$ # Generate models
$ xsdata tests/fixtures/primer/order.xsd --package tests.fixtures.primer
>>> # Parse XML
>>> from pathlib import Path
>>> from tests.fixtures.primer import PurchaseOrder
>>> from xsdata.formats.dataclass.parsers import XmlParser
>>> xml_string = Path("tests/fixtures/primer/sample.xml").read_text()
>>> parser = XmlParser()
>>> order = parser.from_string(xml_string, PurchaseOrder)
>>> order.bill_to
Usaddress(name='Robert Smith', street='8 Oak Avenue', city='Old Town', state='PA', zip=Decimal('95819'), country='US')

Check the documentation for more ✨✨✨


  • Generate code from:

    • XML Schemas 1.0 & 1.1

    • WSDL 1.1 definitions with SOAP 1.1 bindings

    • DTD external definitions

    • Directly from XML and JSON Documents

    • Extensive configuration to customize output

    • Pluggable code writer for custom output formats

  • Default Output:

    • Pure python dataclasses with metadata

    • Type hints with support for forward references and unions

    • Enumerations and inner classes

    • Support namespace qualified elements and attributes

  • Data Binding:

    • XML and JSON parser, serializer

    • PyCode serializer

    • Handlers and Writers based on lxml and native xml python

    • Support wildcard elements and attributes

    • Support xinclude statements and unknown properties

    • Customize behaviour through config

Changelog: 24.1 (2024-01-04)#

  • Fixed XmlParser to ignore xsi attrs when fail on unknown attributes is enabled (#846)

  • Fixed parsing mandatory byte elements with no value (#873)

  • Fixed issue in json binding with union fields (#864)

  • Fixed PycodeSerializer to escape unicode characters in string values (#877)

  • Fixed compound field choices with forward references (#886)

  • Fixed google style docstrings to add missing colon (#884)

  • Fixed deprecation warnings for datetime.datetime.utcnow()

  • Fixed XmlSerializer to ignore empty non-nillable/required tokens (#902)

  • Fixed issue with invalid variable names when using originalCase naming convention (#881)

  • Added type hints for compound fields (#858, #885)

  • Added http header when loading remote resources (#867)

  • Added warning when converting parent field to a list (#871)

  • Added ruff to format generated code (#892)

  • Added option to use substitution group names for compound field name (#905)

  • Updated base64 decoding to enable validation (#875)

  • Updated generator to render prohibited parent fields with restriction extensions (#908)

  • Updated generator so plugins can easily override templates

Why naive?

The W3C XML Schema is too complicated but with good reason. It needs to support any api design. On the other hand when you consume xml you don’t necessarily care about any of that. This is where xsData comes in, to simplify things by making a lot of assumptions like the following one that started everything:

All xs:schema elements are classes everything else is either noise or class properties