ALMA test xmls are not valid MARCXML

[TobiasNx]

See here

The problem is, that the namespace is not provided in the marc xml files:
https://github.com/hbz/lobid-resources/blob/b1b8a24d6958026a2adebf1ff607a6ec9dd664aa/src/test/resources/alma

It s: <record>

Should be: <record xmlns="http://www.loc.gov/MARC21/slim">

If there is a<collection>element as in https://github.com/hbz/lobid-resources/blob/4fbc7cfa76792bc5730f19a17f2292d24ab515ac/src/test/resources/alma/almaMarcXmlTestFiles.xml.tar.bz2 then the namespace should include the namespace refrence not the <record> element.
It s: <collection><record>

Should be: <collection xmlns="http://www.loc.gov/MARC21/slim"><record>

Because of this the playground as well as other Flux-Scripts do not recognize the files without the namespace as marc-xml.

[dr0i]

But our ETL is working with it.
What looks your flux like?
In java its: FileOpener->XmlDecoder->MarcXmlHandler ...

[dr0i]

ah, uh - and use marcXmlHandler.setNamespace(null); - this prevents an ns-check. In flux, use | handle-marcxml(namespace=null).
Try that. May be it's not working because the null must be a null and not a string "null", so that this workaround doesn't work with flux.
Then you could also try | handle-marcxml(namespace="").

[TobiasNx]

Thanks. But the option behaviour seems odd and is not documented:
https://github.com/metafacture/metafacture-documentation/blob/master/flux-commands.md#handle-marcxml

This should somehow be documented since normally this would be "tru"/"false"

[blackwinter]

But the option behaviour [...] is not documented

Of course it's documented:

options: namespace (String), attributemarker (String)

You can either set it to null as @dr0i suggested (if that works in Flux) in order to disable the namespace check or set it to the required namespace value ("http://www.loc.gov/MARC21/slim") in order make the check pass (see metafacture/metafacture-core#331).

normally this would be "tru[e]"/"false"

This only applies to boolean options.

[blackwinter]

But the option behaviour [...] is not documented

Oh, did you mean the part about disabling the namespace check? Then you're right, that's currently not documented. Sorry if I misunderstood.

[dr0i]

@TobiasNx can you document it?

[TobiasNx]

metafacture/metafacture-core#504

[dr0i]

This issue is related to metafacture/metafacture-core#569.
Btw: is it solved? What is missing here?

[TobiasNx]

While we can work with the marc data from lobid with metafacture now, due to the adjustemts made in metafacture/metafacture-core#569 . The marc data is still not valid.

But since the alma publishing data is part of our records one could argue the data is no "pure" marc xml.

So we could still add the namespace to the marc xml records OR just put this on wont fix until somebody asks for it.

[dr0i]

Won't fix atm as discussed in our meeting. Closing.