check if language recognition works for Latin

[dietervu]

Eg the following fragment does not get recognized automatically:

Quod ad me pertinet, desidiosus meo more, vix titulum
libris, quos te monente scripturus sum, imposui. Neque me
tamen ita desidiosum existimes, ut plane tempus inutiliter
conteram; immò nunquam utilius, sed in rebus quas ingenium tuum, altioribus occupatum, haud dubie contemnet,
et ex edito Scientiarum caelo despiciet: nempe in Picturâ,
Architecturâ militari, et praecipue sermone Belgico. In quo
quid profecerim, brevi visurus es: petam enim Middelburgum, si Deus sinat, quadragesimâ ineunte.

[proycon]

I don't know what you use for language detection in the switchboard, but I wrote a small wrapper around lingua-rs which works quite decently (also works on this latin text): https://git.sr.ht/~proycon/lingua-cli , might be useful.

[andmor-]

We use Apache Tika. Given the backend in Java, running Rust of Shell commands, is not the most elegant solution. But can be done and we can take a look in case SB language detection underperforms.

Please add references to this ticket in case you find other cases where the SB language detection is failing