Error with lxml package

[WolverineTerminator]

Hello,

I installed xlml into my environment using pipenv. I then executed the validate.py file and got an error in the xlml package.

I'm not quiet sure where to dig in lxml is a formidable package so I'm looking for anyone who has experienced this and giving me some guidance on what to try. Thanks, T

[cwhanse]

This error originates with COLLADA, which has a dependency on the mathml2.xsd schema. The mathml2.xsd schema is not being loaded which raises the XMLSchemaParseError.

lxml is supposed to handle importing other schemas for you. I can run validate.py without error and I don't have mathxml2.xsd locally. So it appears that lxml is fetching mathml2.xsd for me from the schema location at http://www.w3.org/Math/XMLSchema/mathml2/mathml2.xsd.

Can you check your internet connection?

What do you get when you run this little python script?

from lxml import etree
e = etree.XMLSchema(file='http://www.w3.org/Math/XMLSchema/mathml2/mathml2.xsd')
print(e)

You should get something like

<lxml.etree.XMLSchema object at 0x0000025D59BF59E0>

[WolverineTerminator]

Thanks Cliff.

Larger question - should we have a pip file to establish the environment to run the schema so that we know which version of the package we need? I have lxml == Version: 6.0.2

[cwhanse]

I have the same version of lxml. Hate to sound like tech support, but I think it's a connection or proxy server issue on your end.

You could download the mathml2.xsd file and put it in the same directory as the other xsd files, that might get you around the error. Or might not, since we're not going to edit the COLLADA xsd file to change the schema location value.

https://www.rsc.org/schema/mathml2/ has a copy archived. You will need to unzip that into your pvcollada/schema folder.

[cwhanse]

I can confirm that lxml, when parsing the COLLADA xsd, needs to also access http://www.w3.org/Math/XMLSchema/mathml2 to obtain the mathml2 schema and its upstream references (about 20 of them). I can't see how to redirect lxml to local files without modifying a copy of COLLADA xsd.

Maybe knowing the URL helps.

[SeSamSeg]

Also ran into this issue with another tool. W3C seems to actively do rate limiting on these schemes.

I'm behind a CGN (carrier NAT), which seems to make this issue worse and more sporadic.

W3C advises to use a XML catalog to point the W3C schemes to a local copy when used in a tool (like CI script) or in a web service as to not stress their servers.

I guess you could also use a custom resolver if the tool allows that.