Does the DOC object have a built-in method for extracting a line of poetry?

--No, but it might be a good new feature; but it's problematic since doc ~ sentences and not verses.

Or would I just declare one by extracting the contents between newlines in the raw string?

--I would create a doc object for each sentence in verse, otherwise POS tags may not resolve as desired.

It would seem that getting the POS from all the words in a line would do 90% of the work, and getting the cltk.morphology.universal_dependencies_features.Case from the declined POS would do the rest. Is there any leeway for coding ambiguity into the raw corpora? I mean aren't there texts where a form could be understood as being either of two different POS? Or either of two different cases? To pull in all possibilities I'd like to include ambiguous lines as well.

--No there's no built-in method to accommodate POS tag ambiguity in a corpus. With poetry one has to expect the rules of language to be broken.

You probably want to have a look at the HexameterScanner class and related ones; knowing which vowels must be accented can help narrow the cases; POS taggers are mostly trained on prose and unaccented texts.

Thank you so much! That explains a lot.
The workflow I had envisioned was to start first with the Hexameter scanner to see if the line was a hexameter or pentameter. Spot checking the corpora I saw that some prose intros were nestled among the poems.

Designing a specific Doc for poetry would be valuable for poetry study, but I don't think there is a single solution. Maybe cltk can propose a solution to users and show to advanced users how to customize it.
I remember I tried to transform the HexameterScanner into a Process but it was harder than expected.

Well, in the corpus poetry lines seem to be organized via line breaks, so can/could the DOC include the line break as a POS or other token? Then the workflow could go as follows:

1. If the lines lack macrons, first a hexameter scanner could work on the corpus from line-break to line-break and create a new corpus with macrons.
2. The new corpus can be processed into a DOC to get POS, more accurately thanks to macrons.
3. My new script could examine the DOC, checking the POS and cases of all the words between one line break and the next to evaluate as a likely candidate for a golden line.
   Again, thank you for your patience with me and my project on a bizarre bywater of Latin metrics.

I've been tasked to learn more about Bard AI, so here's what Bard just suggested.

import cltk

Create a new grammar

grammar = cltk.parse.LabeledTreebankGrammar()

Add a new rule to the grammar

grammar.add_rule('line_break', '\n')

Parse the sentence

sentence = "Imminet exitio vir coniugis, illa mariti; Lurida terribiles miscent aconita novercae"
tokens = cltk.parse.pos_tag(sentence, grammar)

Print the POS of the line break

print(tokens[tokens.index('\n')])