Jabran.com — Reconstruction Instructions
Version: 1
Audience: an archivist rebuilding the corpus from an export package.

1.  Read MANIFEST.json at the root of the package.
2.  Load semantic-kernel.json to obtain the controlled vocabulary.
3.  Load concepts/*.json to obtain the concept nodes.
4.  Load works/*.json (or works/*.md alongside works/*.json) to
    obtain the works.
5.  Load claims/*.json to obtain the extracted claims.
6.  Load passages/*.json to obtain the extracted passages.
7.  Load editions/*.json to obtain the immutable revisions.
8.  Verify edition hashes:
    for each edition, recompute SHA-256 over
    (title || '|' || body_markdown || '|' || revision_number || '|'
     || edition_date_utc_iso)
    and compare with edition_hash.
9.  Verify the edition chain: each edition's previous_edition_hash
    should match the prior edition's edition_hash for the same work.
10. Reconnect works to concepts via article_concepts,
    to claims via claim.article_id,
    to passages via passage.article_id,
    and to sources via bibliography_items.
11. Use Corpus IDs as primary identifiers.
12. Use slugs as secondary identifiers; consult slug_history when
    resolving old URLs.
13. Cite fixed editions for permanence.
14. Use living versions for current interpretation, understanding
    that they may have advanced beyond any given edition.