saxonpy 0.0.3

A python package for the Saxon/C 1.2.1, an XML processor by Saxonica.

Welcome to Saxonpy!

This's a python wheel package for Saxon, an XML processor from the Saxonica company. Saxon includes a range of tools for XML transformations, XML queries, and schema validations. For this release, we only include the support for the home edition or the open-source version.

When to choose Saxonpy

If you need to use XSLT versions greater than 1.0 like 2.0 and 3.0, then you may use Saxonpy.

• As of 2021, there are two great python packages already available: xml and lxml . The Python xml seems to be great if you just need to parse the xml documents while lxml seems to offer more options for processing xml files. Because lxml uses libxslt, a C library for transforming xmls, it supports xslt version 1.0 only.

Installation

pip install saxonpy

Usage

I will make references to the Saxonica official documentation a lot from here because they list all the Python APIs with examples for some.
Import all the modules

from saxonpy import *

Import specific modules

from saxonpy import PySaxonProcessor

Now, let's check what 6 other processors are available from Saxon by visiting the Saxonica's site. In the API section, you will find more info about what each processor does.

Next, we will use PySaxonProcessor to check the Saxon version.

from saxonpy import PySaxonProcessor

with PySaxonProcessor(license=False) as proc:
	print(proc.version)

It will print the version like below if your installation is successful.

Saxon/C 1.2.1 running with Saxon-HE 9.9.1.5C from Saxonica

Please note that license=False indicates the open-source version of Saxon.

Example #1

Let's parse a toy XML that was available from Saxonica source code.

from  saxonpy  import PySaxonProcessor

with PySaxonProcessor(license=False) as  proc:
	xml = """\
		<out>
			<person att1='value1' att2='value2'>text1</person>
			<person>text2</person>
			<person>text3</person>
		</out>
		"""
	node = proc.parse_xml(xml_text=xml)
	print("node.node_kind="+ node.node_kind_str)
	print("node.size="+ str(node.size))
	outNode = node.children
	print("len of children="+str(len(node.children)))
	print('element name='+outNode[0].name)
	children = outNode[0].children
	print(*children, sep= ', ')
	attrs = children[1].attributes
	if  len(attrs) == 2:
		print(attrs[1].string_value)

In the output, we will get this.

node.node_kind=document
node.size=1
len of children=1
element name=out
        , <person att1="value1" att2="value2">text1</person>, 
        , <person>text2</person>, 
        , <person>text3</person>, 
value2

As we can see, we can explore the XML node structure, attributes, and many other things if you check more on the APIs.

Example #2

Let's use the XML path processor from Saxonica.

from  saxonpy  import PySaxonProcessor

with PySaxonProcessor(license=False) as  proc:
	xml = """\
		<out>
			<person>text1</person>
			<person>text2</person>
			<person>text3</person>
		</out>"""

	xp = proc.new_xpath_processor()
	node = proc.parse_xml(xml_text=xml)
	xp.set_context(xdm_item=node)

	item = xp.evaluate_single('//person[1]')
	if  isinstance(item,PyXdmNode):
		print(item.string_value)
	# pay attention, Saxon's xdm data type
	value = proc.make_double_value(3.5)
	print(value.primitive_type_name)

The output shows here.

text1
Q{http://www.w3.org/2001/XMLSchema}double

Saxon shows the result given the path.

Example #3

The XSLT processor, #1

from  saxonpy  import PySaxonProcessor

with PySaxonProcessor(license=False) as  proc:
	xsltproc = proc.new_xslt_processor()

	document = proc.parse_xml(xml_text="<out><person>text1</person><person>text2</person><person>text3</person></out>")

	xsltproc.set_source(xdm_node=document)
	xsltproc.compile_stylesheet(stylesheet_text="<xsl:stylesheet xmlns:xsl='http://www.w3.org/1999/XSL/Transform' version='2.0'> <xsl:param name='values' select='(2,3,4)' /><xsl:output method='xml' indent='yes' /><xsl:template match='*'><output><xsl:value-of select='//person[1]'/><xsl:for-each select='$values' ><out><xsl:value-of select='. * 3'/></out></xsl:for-each></output></xsl:template></xsl:stylesheet>")

	output2 = xsltproc.transform_to_string()
	print(output2)

Here is the output that shows the result of trasformation.

<?xml version="1.0" encoding="UTF-8"?>
<output>text1<out>6</out>
   <out>9</out>
   <out>12</out>
</output>

Example #4

Watch out for the not-pythonic way!

from  saxonpy  import PySaxonProcessor

with PySaxonProcessor(license=False) as  proc:
	xsltproc = proc.new_xslt_processor()

	xml = '<a><b>Text</b></a>'
	xslt = '''\
		<xsl:stylesheet version="1.0"
		xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
		<xsl:param name="a" />
		<xsl:output method='xml' indent='yes' />
		<xsl:template match="/">
		<foo><xsl:value-of select="$a" /></foo>
		</xsl:template>
		</xsl:stylesheet>'''

	document = proc.parse_xml(xml_text=xml)

	#please note the not Python way in the next two lines.
	xdm_a = proc.make_string_value('a was given in the parameter')
	xsltproc.set_parameter('a', xdm_a)

	xsltproc.set_source(xdm_node=document)
	xsltproc.compile_stylesheet(stylesheet_text=xslt)
	output2 = xsltproc.transform_to_string()
	print(output2)

Here we use an XSLT/stylesheet parameter. I just want to highlight that Python String (object) is not the same string that Saxon uses, and the same goes for other types. We need to convert to it by make_string_value. This may be a little exotic and not so pythonic because Saxon is written in Java and cross-compiled for C and then Python. Just be aware of it but you don't need to know anything more about Java or C to use Saxonpy. Please check out the Saxonica's documentation for xquery, schema validation, and others.

Why the with keyword

with is good to clear out the underlying processes when the program exits.

Source code & Development

• We downloaded the Saxonc HE from the Saxonica's site
• We have the source code on a github repo to make the wheel packages for Linux and macOS. We have a separate repo for Windows.
• We use Github actions runners for CI and releases.

Our use cases

Here at the projects of the University of Virginia, we use Saxon for Tibetan cataloging and SolrDb indexing.