Validation with lxml

Apart from the built-in DTD support in parsers, lxml currently supports three schema languages: DTD, Relax NG and XML Schema. All three provide identical APIs in lxml, represented by validator classes with the obvious names.

There is also initial support for Schematron. However, it does not currently support error reporting in the validation phase due to insufficiencies in the implementation as of libxml2 2.6.30.

Contents

The usual setup procedure:

>>> from lxml import etree

Validation at parse time

The parser in lxml can do on-the-fly validation of a document against a DTD or an XML schema. The DTD is retrieved automatically based on the DOCTYPE of the parsed document. All you have to do is use a parser that has DTD validation enabled:

>>> parser = etree.XMLParser(dtd_validation=True)

Obviously, a request for validation enables the DTD loading feature. There are two other options that enable loading the DTD, but that do not perform any validation. The first is the load_dtd keyword option, which simply loads the DTD into the parser and makes it available to the document as external subset. You can retrieve the DTD from the parsed document using the docinfo property of the result ElementTree object. The internal subset is available as internalDTD, the external subset is provided as externalDTD.

The third way way to activate DTD loading is with the attribute_defaults option, which loads the DTD and weaves attribute default values into the document. Again, no validation is performed unless explicitly requested.

XML schema is supported in a similar way, but requires an explicit schema to be provided:

>>> schema_root = etree.XML('''\
...   <xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema">
...     <xsd:element name="a" type="xsd:integer"/>
...   </xsd:schema>
... ''')
>>> schema = etree.XMLSchema(schema_root)

>>> parser = etree.XMLParser(schema = schema)
>>> root = etree.fromstring("<a>5</a>", parser)

If the validation fails (be it for a DTD or an XML schema), the parser will raise an exception:

>>> root = etree.fromstring("<a>no int</a>", parser)
Traceback (most recent call last):
lxml.etree.XMLSyntaxError: Element 'a': 'no int' is not a valid value of the atomic type 'xs:integer'.

If you want the parser to succeed regardless of the outcome of the validation, you should use a non validating parser and run the validation separately after parsing the document.

DTD

As described above, the parser support for DTDs depends on internal or external subsets of the XML file. This means that the XML file itself must either contain a DTD or must reference a DTD to make this work. If you want to validate an XML document against a DTD that is not referenced by the document itself, you can use the DTD class.

To use the DTD class, you must first pass a filename or file-like object into the constructor to parse a DTD:

>>> f = StringIO("<!ELEMENT b EMPTY>")
>>> dtd = etree.DTD(f)

Now you can use it to validate documents:

>>> root = etree.XML("<b/>")
>>> print(dtd.validate(root))
True

>>> root = etree.XML("<b><a/></b>")
>>> print(dtd.validate(root))
False

The reason for the validation failure can be found in the error log:

>>> print(dtd.error_log.filter_from_errors()[0])
<string>:1:0:ERROR:VALID:DTD_NOT_EMPTY: Element b was declared EMPTY this one has content

As an alternative to parsing from a file, you can use the external_id keyword argument to parse from a catalog. The following example reads the DocBook DTD in version 4.2, if available in the system catalog:

dtd = etree.DTD(external_id = "-//OASIS//DTD DocBook XML V4.2//EN")

RelaxNG

The RelaxNG class takes an ElementTree object to construct a Relax NG validator:

>>> f = StringIO('''\
... <element name="a" xmlns="http://relaxng.org/ns/structure/1.0">
...  <zeroOrMore>
...     <element name="b">
...       <text />
...     </element>
...  </zeroOrMore>
... </element>
... ''')
>>> relaxng_doc = etree.parse(f)
>>> relaxng = etree.RelaxNG(relaxng_doc)

Alternatively, pass a filename to the file keyword argument to parse from a file. This also enables correct handling of include files from within the RelaxNG parser.

You can then validate some ElementTree document against the schema. You'll get back True if the document is valid against the Relax NG schema, and False if not:

>>> valid = StringIO('<a><b></b></a>')
>>> doc = etree.parse(valid)
>>> relaxng.validate(doc)
True

>>> invalid = StringIO('<a><c></c></a>')
>>> doc2 = etree.parse(invalid)
>>> relaxng.validate(doc2)
False

Calling the schema object has the same effect as calling its validate method. This is sometimes used in conditional statements:

>>> invalid = StringIO('<a><c></c></a>')
>>> doc2 = etree.parse(invalid)
>>> if not relaxng(doc2):
...     print("invalid!")
invalid!

If you prefer getting an exception when validating, you can use the assert_ or assertValid methods:

>>> relaxng.assertValid(doc2)
Traceback (most recent call last):
  ...
lxml.etree.DocumentInvalid: Did not expect element c there, line 1

>>> relaxng.assert_(doc2)
Traceback (most recent call last):
  ...
AssertionError: Did not expect element c there, line 1

If you want to find out why the validation failed in the second case, you can look up the error log of the validation process and check it for relevant messages:

>>> log = relaxng.error_log
>>> print(log.last_error)
<string>:1:0:ERROR:RELAXNGV:RELAXNG_ERR_ELEMWRONG: Did not expect element c there

You can see that the error (ERROR) happened during RelaxNG validation (RELAXNGV). The message then tells you what went wrong. You can also look at the error domain and its type directly:

>>> error = log.last_error
>>> print(error.domain_name)
RELAXNGV
>>> print(error.type_name)
RELAXNG_ERR_ELEMWRONG

Note that this error log is local to the RelaxNG object. It will only contain log entries that appeared during the validation.

Similar to XSLT, there's also a less efficient but easier shortcut method to do one-shot RelaxNG validation:

>>> doc.relaxng(relaxng_doc)
True
>>> doc2.relaxng(relaxng_doc)
False

libxml2 does not currently support the RelaxNG Compact Syntax. However, the trang translator can convert the compact syntax to the XML syntax, which can then be used with lxml.

XMLSchema

lxml.etree also has XML Schema (XSD) support, using the class lxml.etree.XMLSchema. The API is very similar to the Relax NG and DTD classes. Pass an ElementTree object to construct a XMLSchema validator:

>>> f = StringIO('''\
... <xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema">
... <xsd:element name="a" type="AType"/>
... <xsd:complexType name="AType">
...   <xsd:sequence>
...     <xsd:element name="b" type="xsd:string" />
...   </xsd:sequence>
... </xsd:complexType>
... </xsd:schema>
... ''')
>>> xmlschema_doc = etree.parse(f)
>>> xmlschema = etree.XMLSchema(xmlschema_doc)

You can then validate some ElementTree document with this. Like with RelaxNG, you'll get back true if the document is valid against the XML schema, and false if not:

>>> valid = StringIO('<a><b></b></a>')
>>> doc = etree.parse(valid)
>>> xmlschema.validate(doc)
True

>>> invalid = StringIO('<a><c></c></a>')
>>> doc2 = etree.parse(invalid)
>>> xmlschema.validate(doc2)
False

Calling the schema object has the same effect as calling its validate method. This is sometimes used in conditional statements:

>>> invalid = StringIO('<a><c></c></a>')
>>> doc2 = etree.parse(invalid)
>>> if not xmlschema(doc2):
...     print("invalid!")
invalid!

If you prefer getting an exception when validating, you can use the assert_ or assertValid methods:

>>> xmlschema.assertValid(doc2)
Traceback (most recent call last):
  ...
lxml.etree.DocumentInvalid: Element 'c': This element is not expected. Expected is ( b )., line 1

>>> xmlschema.assert_(doc2)
Traceback (most recent call last):
  ...
AssertionError: Element 'c': This element is not expected. Expected is ( b )., line 1

Error reporting works as for the RelaxNG class:

>>> log = xmlschema.error_log
>>> error = log.last_error
>>> print(error.domain_name)
SCHEMASV
>>> print(error.type_name)
SCHEMAV_ELEMENT_CONTENT

If you were to print this log entry, you would get something like the following. Note that the error message depends on the libxml2 version in use:

<string>:1:ERROR::SCHEMAV_ELEMENT_CONTENT: Element 'c': This element is not expected. Expected is ( b ).

Similar to XSLT and RelaxNG, there's also a less efficient but easier shortcut method to do XML Schema validation:

>>> doc.xmlschema(xmlschema_doc)
True
>>> doc2.xmlschema(xmlschema_doc)
False

Schematron

Since version 2.0, lxml.etree features Schematron support, using the class lxml.etree.Schematron. It requires at least libxml2 2.6.21 to work. The API is the same as for the other validators. Pass an ElementTree object to construct a Schematron validator:

>>> f = StringIO('''\
... <schema xmlns="http://www.ascc.net/xml/schematron" >
...   <pattern name="Sum equals 100%.">
...     <rule context="Total">
...       <assert test="sum(//Percent)=100">Sum is not 100%.</assert>
...     </rule>
...   </pattern>
... </schema>
... ''')

>>> sct_doc = etree.parse(f)
>>> schematron = etree.Schematron(sct_doc)

You can then validate some ElementTree document with this. Like with RelaxNG, you'll get back true if the document is valid against the schema, and false if not:

>>> valid = StringIO('''\
... <Total>
...   <Percent>20</Percent>
...   <Percent>30</Percent>
...   <Percent>50</Percent>
... </Total>
... ''')

>>> doc = etree.parse(valid)
>>> schematron.validate(doc)
True

>>> etree.SubElement(doc.getroot(), "Percent").text = "10"

>>> schematron.validate(doc)
False

Calling the schema object has the same effect as calling its validate method. This is sometimes used in conditional statements:

>>> is_valid = etree.Schematron(sct_doc)

>>> if not is_valid(doc):
...     print("invalid!")
invalid!

Note that libxml2 restricts error reporting to the parsing step (when creating the Schematron instance). There is not currently any support for error reporting during validation.