README.md

canon/tiers/02_Tier_2_Region_Pages/_processed/README.md

Processed — completed source staging

The done-side twin of _ready_for_processing/. When a staged SOURCE has been fully

converted/ingested by the book pipeline, its source folder moves here so the inbox only ever

contains pending work. Nothing in here is pending; nothing in here is read by the pipeline.

What lands here and where its product lives:

• nationalgeographic/discN/ — original .cng disc staging, moved here after

cng_to_jpg.py converts it with ZERO failures. Product: the JPG visual-reference library at

.rag/book_src/natgeo_jpg/discN/.

• rejected_noise/ — inbox files the triage lane classified as zero-value for the corpus

(with rejected_noise/_dispositions.jsonl recording who/what/why). Visible and reversible,

never silently deleted; move a file back to the inbox with a .meta.json to override.

• Other completed staging trees (e.g. a drained sacred-texts mirror) move here whole, once

their extracted products are ingested.

Individually ingested BOOKS do not land here — the ingest itself moves each file to

.rag/book_src/processed/<region>/SRC_NNNNN__<title>.pdf|.md (organized by region, renamed

with its registry id). This folder is for completed STAGING TREES and rejected files, so

"what is left to do" is always exactly the inbox contents.

Disk note: these are large local binaries, gitignored (only this README + .gitignore are

tracked). They stay local; keep your own backup — git is not their home.

Generated by harness/site/structure_site.py — the URL path is the repo path. review root