Core data model and parser for Lepiter knowledge bases stored as page JSON files.

Scope

• Scans a Lepiter directory and builds a metadata index keyed by page id.
• Loads and parses individual pages lazily by id.
• Converts page snippet trees into a stable block-oriented node model.
• Preserves unknown node types as [Node::Unknown] to keep consumers resilient.

Example

use lepiter_core::KnowledgeBase;

# fn main() -> anyhow::Result<()> {
let index = KnowledgeBase::open("./lepiter")?;
for page in index.sorted_pages_by_title() {
    println!("{} - {}", page.id, page.title);
}
# Ok(())
# }