Handle TEI works that have a language of “Arabic [Karshuni]”

[alexwlchan]

These works are currently failing in the transformer; there's some discussion in Slack about how to handle them: https://wellcome.slack.com/archives/CS4G8SAG3/p1651134810190059

[alexwlchan]

Useful to know here: the Language model has an id and a label field, but it behaves slightly differently to our other types. See the comment here: wellcomecollection/catalogue-pipeline#2109

[paul-butcher]

It looks like the recording of the "script: Syriac" part is already handled by existing handNote processing

https://wellcome.slack.com/archives/CS4G8SAG3/p1651143685381139?thread_ts=1651134810.190059&cid=CS4G8SAG3

As evidenced here: https://github.com/wellcomecollection/catalogue-pipeline/blob/ef4de14ab90d3a7436c79f9b040b7d62cf74fbaf/pipeline/transformer/transformer_tei/src/test/scala/weco/pipeline/transformer/tei/TeiTransformerTest.scala#L88

So this is just a question of adding an appropriate customLanguage line here https://github.com/wellcomecollection/catalogue-pipeline/blob/ef4de14ab90d3a7436c79f9b040b7d62cf74fbaf/pipeline/transformer/transformer_tei/src/main/scala/weco/pipeline/transformer/tei/data/TeiLanguageData.scala#L88

[alexwlchan]

Yes, that sounds right to me.