The Ingestion Pipeline

1. What is the primary purpose of the ingestion pipeline in LyraLearn?

To serve user requests directly To turn raw documents into searchable evidence To analyze users' questions To manage external API calls

2. What does the chunking process in LyraLearn utilize to create chunks from documents?

Fixed character count splits Random paragraph selection Heading-aware splits Total word count consideration

3. Why is local-first embedding important in the ingestion pipeline?

It allows for faster retrieval It enables use of external APIs It ensures document content stays within the network It increases the number of vectors processed

4. How does the ingestion pipeline handle unchanged chunks to optimize performance?

It re-embeds all chunks every time It skips unchanged chunks by checking their content hash It deletes all old chunks before new ones are added It stores unchanged chunks in a separate database

Continue: Retrieval and Evidence Grading → Review this lesson Retake quiz