Configuration

The command line tool includes most but not all available options. The most advanced settings can be enabled through a project configuration file.

Create or Update

You can create a default one with the command:

$ xsdata init-config --help
========= xsdata v24.4 / Python 3.11.6 / Platform linux =========

Usage: xsdata init-config [OPTIONS] [OUTPUT]

  Create or update a configuration file.

Options:
  -pp, --print  Print output
  --help        Show this message and exit.

Output: .xsdata.xml

<?xml version="1.0" encoding="UTF-8"?>
<Config xmlns="http://pypi.org/project/xsdata" version="24.4">
  <Output maxLineLength="79" subscriptableTypes="false" unionType="false">
    <Package>generated</Package>
    <Format repr="true" eq="true" order="false" unsafeHash="false" frozen="false" slots="false" kwOnly="false">dataclasses</Format>
    <Structure>filenames</Structure>
    <DocstringStyle>reStructuredText</DocstringStyle>
    <RelativeImports>false</RelativeImports>
    <CompoundFields defaultName="choice" useSubstitutionGroups="false" forceDefaultName="false" maxNameParts="3">false</CompoundFields>
    <WrapperFields>false</WrapperFields>
    <PostponedAnnotations>false</PostponedAnnotations>
    <UnnestClasses>false</UnnestClasses>
    <IgnorePatterns>false</IgnorePatterns>
    <IncludeHeader>false</IncludeHeader>
  </Output>
  <Conventions>
    <ClassName case="pascalCase" safePrefix="type"/>
    <FieldName case="snakeCase" safePrefix="value"/>
    <ConstantName case="screamingSnakeCase" safePrefix="value"/>
    <ModuleName case="snakeCase" safePrefix="mod"/>
    <PackageName case="snakeCase" safePrefix="pkg"/>
  </Conventions>
  <Substitutions>
    <Substitution type="package" search="http://www.w3.org/2001/XMLSchema" replace="xs"/>
    <Substitution type="package" search="http://www.w3.org/XML/1998/namespace" replace="xml"/>
    <Substitution type="package" search="http://www.w3.org/2001/XMLSchema-instance" replace="xsi"/>
    <Substitution type="package" search="http://www.w3.org/1998/Math/MathML" replace="mathml3"/>
    <Substitution type="package" search="http://www.w3.org/1999/xlink" replace="xlink"/>
    <Substitution type="package" search="http://www.w3.org/1999/xhtml" replace="xhtml"/>
    <Substitution type="package" search="http://schemas.xmlsoap.org/wsdl/soap/" replace="soap"/>
    <Substitution type="package" search="http://schemas.xmlsoap.org/wsdl/soap12/" replace="soap12"/>
    <Substitution type="package" search="http://schemas.xmlsoap.org/soap/envelope/" replace="soapenv"/>
    <Substitution type="class" search="(.*)Class$" replace="\1Type"/>
  </Substitutions>
  <Extensions/>
</Config>

Usage

$ xsdata --config project/.xsdata.xml

CLI options override the project configuration settings.

Output Settings

Configuration settings related to output generation.

`maxLineLength`

The maximum line length of the generated code.

Default Value: 79

CLI Option: -mll, --max-line-length INTEGER

`subscriptableTypes`

Use PEP-585 type hinting generics in standard collections.

tuple # typing.Tuple
list # typing.List
dict # typing.Dict
set # typing.Set
type # typing.Type

Default Value: False

CLI Option: --subscriptable-types / --no-subscriptable-types

Requirements: python>=3.9

`unionType`

Use PEP-604, allow writing union types as X | Y.

str | None # Optional[str]
str | int # Union[str, int]

Default Value: False

CLI Option: --union-type / --no-union-type

Requirements: python>=3.10

Package

The output package for the generated code, e.g. code.models

Default Value: generated

CLI Option: -p, --package TEXT

Format

The output format for the generated code, e.g. code.models

Default Value: dataclasses

CLI Option: -o, --output TEXT

Attributes

The dataclass parameters.

repr: Generate the repr method.
eq: Generate the eq method.
order: Generates the lt, le, gt, ge methods.
frozen: This emulates read-only frozen instances.
unsafeHash: Generates a hash method according to how eq and frozen are set.
slots: Generates the class slots attribute. python >= 3.10
kwOnly: All fields will be marked as keyword-only. python >= 3.10

Warning

A TypeError is raised if a field without a default value follows a field with a default value. This is true whether this occurs in a single class, or as a result of class inheritance. If this option is not enabled, the generator will mark all required fields without default values as optional with default value None.

Structure

The file structure style to create.

Style	Description
`filenames`	Group classes by the resource file location.
`namespaces`	Group classes by the target namespace.
`clusters`	Group classes by strongly connected dependencies. The closest thing to one-class-per-package, safe from circular imports.
`single-package`	Group classes in a single package. Safe from circular imports, but can create some huge files.
`namespace-clusters`	Group classes by strongly connected dependencies and target namespaces.

Default Value: filenames

CLI Option: -ss, --structure-style [filenames|namespaces|clusters|single-package|namespace-clusters]

DocstringStyle

The style of docstrings to create.

Default Value: reStructuredText

CLI Option: -ds, --docstring-style [reStructuredText|NumPy|Google|Accessible|Blank]

RelativeImports

Generate relative instead of absolute imports.

Default Value: False

CLI Option: --relative-imports / --no-relative-imports

CompoundFields

xsdata relies on the field ordering for serialization. This process fails for repeating choice or complex sequence elements. When you enable compound fields, these elements are grouped into a single field.

    <xsd:element name="root">
        <xsd:complexType>
            <xsd:choice maxOccurs="unbounded">
                <xsd:element ref="alpha"/>
                <xsd:element ref="bravo"/>
                <xsd:element name="charlie" type="charlie"/>
            </xsd:choice>
        </xsd:complexType>
    </xsd:element>

Default Value: False

CLI Option: --compound-fields / --no-compound-fields

Sub Settings

defaultName: The default compound field name, default: choice
forceDefaultName: Force the default name in all compound fields, default: False
useSubstitutionGroups: When all elements are part of a substitution, use the group name as the field name, default: False
maxNameParts: The maximum length of elements names allowed before using the default name, default: 3

Examples:

# Force default name or max name parts > 3
choice: list[str | int | float | bool] = field(...)

# max name parts <= 3
hat_or_bat_cat: list[str | int | float] = field(...)

# All types belong to the same substitution group `product`
product: list[Shoe | Shirt | Hat] = field(...)

WrapperFields

Generate wrapper fields whenever possible for single or collections of simple and complex elements.

The wrapper and wrapped elements can't be optional. If the wrapped value is a list it must have minimum occurs >= 1.

    <xsd:element name="wrapper">
        <xsd:complexType>
            <xsd:sequence>
                <xsd:element ref="xd:alphas"/>
                <xsd:element ref="xd:bravos"/>
                <xsd:element ref="xd:charlies"/>
            </xsd:sequence>
        </xsd:complexType>
    </xsd:element>
    <xsd:element name="alphas">
        <xsd:complexType>
            <xsd:sequence>
                <xsd:element name="alpha" type="xsd:string"/>
            </xsd:sequence>
        </xsd:complexType>
    </xsd:element>

Default Value: False

CLI Option: --wrapper-fields / --no-wrapper-fields

Examples:

alpha: str = field(
    metadata={
        "wrapper": "alphas",
        "type": "Element",
    },
)
bravo: List[int] = field(
    default_factory=list,
    metadata={
        "wrapper": "bravos",
        "type": "Element",
    },
)
charlie: List[Charlie] = field(
    default_factory=list,
    metadata={
        "wrapper": "charlies",
        "type": "Element",
    },
)

PostponedAnnotations

Use PEP-563, postponed evaluation of annotations.

This will add this import from __future__ import annotations in all generated files, in order to avoid forward references.

Default Value: False

CLI Option: --postponed-annotations / --no-postponed-annotations

UnnestClasses

The generator creates inner classes for xs:complexContent. This option allows to unnest all inner classes.

Default Value: False

CLI Option: --unnest-classes / --no-unnest-classes

IgnorePatterns

The generator will create a field metadata property for xs:pattern elements. This property is not used during parsing, it's only informative for the developer, if you want to reduce the noise in the generated code you can enable this option.

Default Value: False

CLI Option: --ignore-patterns / --no-ignore-patterns

IncludeHeader

The generator will add a module docstring in all the output files.

Example

"""This file was generated by xsdata, v24.1, on 2024-01-22 10:20:25

Generator: DataclassGenerator
See: https://xsdata.readthedocs.io/
"""

Default Value: False

CLI Option: --include-header / --no-include-header

Convention Settings

Apply different naming convention per identifier.

Element	Default Case	Default Safe Prefix
`ClassName`	`pascalCase`	`type`
`FieldName`	`snakeCase`	`value`
`ConstantName`	`screamingSnakeCase`	`value`
`ModuleName`	`snakeCase`	`mod`
`PackageName`	`snakeCase`	`pkg`

Attributes

case: The naming class to apply
safePrefix: A prefix to add when the output name is reserved

Cases

Case	Input	Output
`originalCase`	`aBBc`	`aBBc`
`pascalCase`	`my_type`	`MyType`
`camelCase`	`my_type`	`myType`
`snakeCase`	`MyType`	`my_type`
`screamingSnakeCase`	`MyType`	`My_Type`
`mixedCase`	`MY_TyPE`	`MYTyPE`
`mixedSnakeCase`	`MyType`	`My_Type`
`mixedPascalCase`	`my_TYPE`	`MyTYPE`

Substitution Settings

A list of search and replace patterns for identifier names, the substitutions run before and after the naming conventions.

Attributes

type: The identifier type [class|field|module|package]
search: Search Pattern
replace: Replace Pattern

Defaults

<Substitutions>
    <Substitution type="package" search="http://www.w3.org/2001/XMLSchema" replace="xs"/>
    <Substitution type="package" search="http://www.w3.org/XML/1998/namespace" replace="xml"/>
    <Substitution type="package" search="http://www.w3.org/2001/XMLSchema-instance" replace="xsi"/>
    <Substitution type="package" search="http://www.w3.org/1998/Math/MathML" replace="mathml3"/>
    <Substitution type="package" search="http://www.w3.org/1999/xlink" replace="xlink"/>
    <Substitution type="package" search="http://www.w3.org/1999/xhtml" replace="xhtml"/>
    <Substitution type="package" search="http://schemas.xmlsoap.org/wsdl/soap/" replace="soap"/>
    <Substitution type="package" search="http://schemas.xmlsoap.org/wsdl/soap12/" replace="soap12"/>
    <Substitution type="package" search="http://schemas.xmlsoap.org/soap/envelope/" replace="soapenv"/>
    <Substitution type="class" search="(.*)Class$" replace="\1Type"/>
</Substitutions>

Extension Settings

Though extensions you can add base classes, mixins or decorators to the generated classes. This way you can enhance the models functionality and add any custom business logic.

The following configuration will add a base class and a decorator to all the generated classes.

Attributes

type: The extension type [class|decorator]
class: The class name search pattern
import: The absolute import of the base class or decorator object
prepend Specify if you want the base class or decorator to added before all other
apply_if_derived Specify if you want to add the extension if the class already extends another class.

Warning

If there are two extensions of the same type for the same class with the prepend==True, the base classes or decorators are added in the reverse order they are defined in the configuration.

Example:

<Extensions>
    <Extension type="class" class=".*" import="dataclasses_jsonschema.JsonSchemaMixin" prepend="false" applyIfDerived="false"/>
    <Extension type="decorator" class=".*" import="typed_dataclass.typed_dataclass" prepend="false" applyIfDerived="false"/>
</Extensions>

from dataclasses import dataclass
from dataclasses_jsonschema import JsonSchemaMixin
from typed_dataclass import typed_dataclass

@dataclass
@typed_dataclass
class Cores(JsonSchemaMixin):
    ...