How to handle chunking when relevant text are not near each other

AI Discussions
,
1

Sentence chunking is easy to code up. Just split a corpus of text on newlines and periods. This works great if each chunk as subject-verb-object type structures like:

Llamas are vegetarians and have very efficient digestive systems.
Llamas live to be about 20 years old, though some only live for 15 years and others live to be 30 years old.
Llamas will sometimes spit on people.

But what is the strategy if the text is like this:

In general, llamas:

  • will sometimes spit on people
  • can live to be 20 years old
  • are vegetarians and have very efficient digestive systems

Where the subject, “llamas” is on a different line from the details (bullet points). What should the strategy be here?

2

Perhaps turn the bulleted list into a single sentence, connected by the word “and” as needed.