The proposal of this package is to provide XPath 1.0, 2.0, 3.0 and 3.1 selectors for ElementTree XML data structures, both for the standard ElementTree library and for the lxml.etree library.

For lxml.etree this package can be useful for providing XPath 2.0/3.0/3.1 selectors, because lxml.etree already has it's own implementation of XPath 1.0.

You can install the package with pip in a Python 3.8+ environment:

pip install elementpath

For using it import the package and apply the selectors on ElementTree nodes:

>>> import elementpath
>>> from xml.etree import ElementTree
>>> root = ElementTree.XML('<A><B1/><B2><C1/><C2/><C3/></B2></A>')
>>> elementpath.select(root, '/A/B2/*')
[<Element 'C1' at ...>, <Element 'C2' at ...>, <Element 'C3' at ...>]

The select API provides the standard XPath result format that is a list or an elementary datatype's value. If you want only to iterate over results you can use the generator function iter_select that accepts the same arguments of select.

The selectors API works also using XML data trees based on the lxml.etree library:

>>> import elementpath
>>> import lxml.etree as etree
>>> root = etree.XML('<A><B1/><B2><C1/><C2/><C3/></B2></A>')
>>> elementpath.select(root, '/A/B2/*')
[<Element C1 at ...>, <Element C2 at ...>, <Element C3 at ...>]

When you need to apply the same XPath expression to several XML data you can also use the Selector class, creating an instance and then using it to apply the path on distinct XML data:

>>> import elementpath
>>> import lxml.etree as etree
>>> selector = elementpath.Selector('/A/*/*')
>>> root = etree.XML('<A><B1/><B2><C1/><C2/><C3/></B2></A>')
>>> selector.select(root)
[<Element C1 at ...>, <Element C2 at ...>, <Element C3 at ...>]
>>> root = etree.XML('<A><B1><C0/></B1><B2><C1/><C2/><C3/></B2></A>')
>>> selector.select(root)
[<Element C0 at ...>, <Element C1 at ...>, <Element C2 at ...>, <Element C3 at ...>]

Public API classes and functions are described into the elementpath manual on the "Read the Docs" site.

For default the XPath 2.0 is used. If you need XPath 1.0 parser provide the parser argument:

>>> from elementpath import select, XPath1Parser
>>> from xml.etree import ElementTree
>>> root = ElementTree.XML('<A><B1/><B2><C1/><C2/><C3/></B2></A>')
>>> select(root, '/A/B2/*', parser=XPath1Parser)
[<Element 'C1' at ...>, <Element 'C2' at ...>, <Element 'C3' at ...>]

For XPath 3.0/3.1 import the parser from elementpath.xpath3 subpackage, that is not loaded for default:

>>> from elementpath.xpath3 import XPath3Parser
>>> select(root, 'math:atan(1.0e0)', parser=XPath3Parser)
0.7853981633974483

Note: XPath3Parser is an alias of XPath31Parser.

If you need only XPath 3.0 you can also use a more specific subpackage, avoiding the loading of XPath 3.1 implementation:

>>> from elementpath.xpath30 import XPath30Parser
>>> select(root, 'math:atan(1.0e0)', parser=XPath30Parser)
0.7853981633974483

You can contribute to this package reporting bugs, using the issue tracker or by a pull request. In case you open an issue please try to provide a test or test data for reproducing the wrong behaviour. The provided testing code shall be added to the tests of the package.

The XPath parsers are based on an implementation of the Pratt's Top Down Operator Precedence parser. The implemented parser includes some lookup-ahead features, helpers for registering tokens and for extending language implementations. Also the token class has been generalized using a MutableSequence as base class. See tdop.py for the basic internal classes and xpath1_parser.py for extensions and for a basic usage of the parser.

If you like you can use the basic parser and tokens provided by the tdop.py module to implement other types of parsers (I think it could be also a funny exercise!).

This software is distributed under the terms of the MIT License. See the file 'LICENSE' in the root directory of the present distribution, or http://opensource.org/licenses/MIT.