Infinite loop when using schematron from Python lxml

[alexis-via]

I'm trying to use the CII or UBL schematron from Python lxml's lib, but I always get an infinite loop.

Here is a scenario to reproduce the bug:

cd /tmp
git clone https://github.com/CenPC434/validation.git

Then I run the following python script:

#! /usr/bin/python2.7

from lxml import etree
from lxml import isoschematron

f = open('/tmp/validation/cii/schematron/EN16931-CII-validation.sch', 'r')
sct_root = etree.parse(f)
print("sct_root=%s" % sct_root)
# print(etree.tostring(sct_root, pretty_print=True))
schematron = isoschematron.Schematron(sct_root)
print("schematron=%s" % schematron)

I get an infinite loop when reaching isoschematron.Schematron(). I also have an infinite loop with the UBL schematron.

I don't know if the bug is caused by the schematron or if it's in python's lxml lib. So I also reported a bug to lxml: https://bugs.launchpad.net/lxml/+bug/1783890

And you get an infinite loop when you reach the line isoschematron.Schematron().

[phax]

Hi. Are you sure its an infinite loop and that it simply doesnt take forever? It sometimes takes up to 30 minutes to generate the XSLT from the SCH files!!!

[alexis-via]

30 minutes !!! No, I didn't wait that long... I'm running it again now, let's see if it works if I wait long enough.

[alexis-via]

You are right, I had to wait longer ! After 20 minutes, it fails:

Traceback (most recent call last):
  File "./fx-schematron.py", line 10, in <module>                                                   
    schematron = isoschematron.Schematron(sct_root, store_report=True)                              
  File "/usr/local/lib/python2.7/dist-packages/lxml/isoschematron/__init__.py", line 285, in __init__                                                                                                   
    validator_xslt = self._compile(schematron, **compile_params)                                    
  File "src/lxml/xslt.pxi", line 600, in lxml.etree.XSLT.__call__                                   
lxml.etree.XSLTApplyError: Fail: This implementation of ISO Schematron does not work with           
        schemas using the "xslt2" query language.

[phax]

Does your library have a way to use precompiled XSLTs? That would simplify your life - right?

[alexis-via]

According to this post https://stackoverflow.com/questions/46767903/schematronparseerror-invalid-schematron-schema-for-isosts-schema there is no implementation of schematron in Python that support queryBinding="xslt2". So I guess we can close this bug report... as it seems there is no solution to this problem in Python :-(

[phax]

Sorry :(

[alexis-via]

@phax Do you mean that this xslt file here https://github.com/CenPC434/validation/blob/master/cii/xslt/EN16931-CII-validation.xslt is a pre-compiled version of the CII schematron, so that I could use this XSLT file directly ?

[alexis-via]

When I try to use the XSLT file provided in this git report "cii/xslt/EN16931-CII-validation.xslt", it fails on line 2129:

Traceback (most recent call last):
  File "./fx-xslt.py", line 14, in <module>
    transform = etree.XSLT(xslt_root)
  File "src/lxml/xslt.pxi", line 410, in lxml.etree.XSLT.__init__
lxml.etree.XSLTParseError: xsl:when : could not compile test expression '/rsm:CrossIndustryInvoice/rsm:SupplyChainTradeTransaction/ram:ApplicableHeaderTradeAgreement/ram:SellerTradeParty/ram:SpecifiedTaxRegistration/ram:ID[@schemeID = ('VAT', 'FC')] or /rsm:CrossIndustryInvoice/rsm:SupplyChainTradeTransaction/ram:ApplicableHeaderTradeAgreement/ram:SellerTaxRepresentativeTradeParty/ram:SpecifiedTaxRegistration/ram:ID[@schemeID = 'VAT']'

The small Python script I used for that is the following:

#! /usr/bin/python

from lxml import etree

xslt_file = open('/home/alexis/new_boite/dev/schematron-fx/cii/xslt/EN16931-CII-validation.xslt', 'r')
xslt_root = etree.parse(xslt_file)
print "xslt_root=", xslt_root

f = open('/home/alexis/tmp/factur-x-demov7.xml', 'r')
fx_xml_root = etree.parse(f)
print "fx_xml_root=", fx_xml_root

transform = etree.XSLT(xslt_root)
print "transform=", transform
newdom = transform(fx_xml_root)
print "newdom=", newdom
print etree.tostring(newdom, pretty_print=True)

[kosek]

Python doesn't support XSLT2 by default. However you can invoke external XSLT2 engine like Saxon. Or perhaps someone already created Python wrapper around Saxon/C. See http://www.saxonica.com/saxon-c/index.xml

[phax]

@alexis-via Yes, that was the idea. The XSLT you mention is a pre-compiled version of the Schematron. Because what usually happens is the following:

• SCH is converted via a predefined XSLT to the SCH-XSLT (in reality its 3 conversions but who cares)
• That SCH-XSLT can than be applied as any other XSLT onto the XML to be validated, creating an SVRL output
• Parse that SVRL to determine the success/failure status.

[corymosiman12]

Any updates on a means to validate xslt2, either through lxml or other?