XML Classes

Simple, fast, lightweight xml parser and validator defined by classes, and useful error messages.

Reason to pick this over other xml parsers:

• Non alias. (Harder to match tag/attribute name with the class variable name)
• Need advanced features. (Like default factory, custom decoder, custom encoder etc.)

TODO: Speed test against pydantic-xml and xmltodict with pydantic.

Basic examples:

For the following XML:

<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<root value="data" />

The following python class can be used to parse it:

class RootAttribute(XmlClass):
    value: str

It automatically assumes it is an xml tag attribute.

And then parse it:

    root = RootAttribute.from_string(xml_string)

For the following XML:

<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<root>
    <value value="data"/>
</root>

The following python class can be used to parse it:

class Value(XmlClass):
    value: str

class RootElement(XmlClass):
    value: Value

When the type is set to be of XmlClass, it will be parsed as a sub element.

For xml text field, it is a bit more complex. For the following XML: """

data1 """

The following python class can be used to parse it:

class Value(XmlClass):
    data: XmlTextField[str]

When the type is set to be of XmlTextField, it will be parsed as a text field, where the text-field is of the type given in the square brackets.

Supported Annotations type:

• int
• float
• str
• bool ¹
• None
• datetime.datetime ²
• uuid.UUID
• pathlib.Path
• typing.Any ³
• Literal
• Enum
• list ⁴
• tuple ⁴
• union
• XmlClass
• XmlTextField

Other Info:

Trailing underscore

Trailing underscore is ignored when mapping from XML to python class. It is used in the casses where a child element tag and a attribute have the same name. Through not a requirement, the trailing underscore should be used in the attribute name.

Dash in tag or Attribute names

Since dash in python name is not allowed, dash in tag or attribute name is replaced with underscore.

NOTE:

• This is a work in progress, ATM it only parses the XML data, to a python class.
• UNSURE: Do None type hint as the only one even make sense?
• UNSURE: Should null value be None type hint, and something else for optional?

Tests TODO:

• Test XmlClass | None case, where there is parsing error in XmlClass.
• Test OneTwo: typing.Literal["firstPasser", "secondPasser"] | None case, where it do not Exist.
• Test xmlClass | AnotherXmlClass case, with different Names.

TODO:

• Make the error messages better for when there is a error deep inside a xmlClass.
• Make a test for typing.Any.
• Added defusedxml as dependency. (Security reasons)
• Fix ALL type hints...
• Add support for decoder and encoder.
• Add support for custom parsers. (XmlDataParser[MyParserOfData])
• Make good and useful error messages.

Footnotes

1. Where ["true", "1", "yes", "on"] is true and ["false", "0", "no", "off"] is false. ↩

2. Is parsed by the datetime.datetime.fromisoformat method. ↩

3. Is parsed as is (as string). Only check if there is data. ↩

4. Assume that the children can only be of type: XmlClass. Since multiple attributes are not allowed, and a data separator is not defined. ↩ ↩²