missing language attribute in CTS metadata

[lcerrato]

Hi, after running a test ingest of this repository for Greg Crane, I noticed there are missing language attributes in the work metadata.
These should be included for each edition.

<ti:edition urn="urn:cts:pta:pta9999.pta001.pta-syc1" workUrn="urn:cts:pta:pta9999.pta001">
should be
<ti:edition urn="urn:cts:pta:pta9999.pta001.pta-syc1" workUrn="urn:cts:pta:pta9999.pta001" xml:lang="syc"> etc.

For example here:
https://scaife-dev.perseus.org/library/urn:cts:pta:pta9999.pta001/

The first edition is listed as Greek, not Syriac

Also when viewing a file, if a user wants an alternate edition or translation, the language will also be incorrect.
Here it appears as if there are two Greek editions only.

[pharos-alexandria]

@lcerrato Thanks for the report! I'll correct it with the next update. This will take some time, as I need to correct some more files to fit to the latest schema.

[pharos-alexandria]

@lcerrato I've checked with http://capitains.org/pages/guidelines and see now the possible reason for this: In the guidelines there's expected to have a @xml:lang for ti:work, but not for ti:edition, but then again for ti:translation. Obviously Scaife Viewer also expects it for ti:edition.
In pta_data it's missing throughout. It's now only visible as missing where there are editions of different language in a work, besides the bible also for example here: https://scaife-dev.perseus.org/library/urn:cts:pta:pta0003.pta019/.

[lcerrato]

@pharos-alexandria
Yes, those guidelines may be deprecated versus how things were put into practice in subsequent years.
The language attribute was something we needed for software that predated Scaife as well, so we've been using it at Perseus for a while.

[pharos-alexandria]

@lcerrato It does not "hurt" our setup and pipeline for PTA to have this xml:lang info. I've added it to my script and with the next release of pta_data they will be available.

[mgbilby]

@pharos-alexandria would it be possible to add the xml:lang info to the PTA schema as a requirement so that it is enforced during file validation? That would be helpful for my workflow, since I'm now defaulting to the PTA schema, but would like the files to be cross-compatible with Perseus/First1KGreek as much as possible.

[lcerrato]

@pharos-alexandria @mgbilby
In terms of the metadata language attribute, it is only going to be an issue where you have multilingual editions/translations for a particular work ID. Otherwise, the language will be inherited from the top level default.
We have this issue as well: it is a very easy thing to miss because it will still pass the Capitains Hook Test.

I did not look comprehensively at the ingest as you have a large collection, so I would encourage you to take a look while things are not yet published: https://scaife-dev.perseus.org/library/ to see if there are other things.

[pharos-alexandria]

@mgbilby It's actually already in the Schema for the TEI edition files proper (as a schematron constraint), but has not been in the __cts__.xml files within the file structure of PTA. But with my change in the script generating these files, that problem does not exist anymore.

[mgbilby]

@lcerrato @pharos-alexandria I see now. Thank you for clarifying. I'd already been adding xml:lang to the edition metadata file in the work subdirectories. But this was a good opportunity to practice situations with multilingual editions, so I've now added the one Latin edition of Marcion's Evangelion to the edition metadata.