Ingest a Directory or Repository

Bulk ingestion is a provenance operation, not just a file copy. Pin the exact source state and produce a manifest that another operator can verify.

Repository campaign

python tools/entity_btdu_cli.py --entity-state STATE --btdu-state BTDU --adam-root ADAM --git-bin git ingest-repo REPOSITORY --provenance git://owner/repository@COMMIT

The resulting repository manifest can record the Git commit SHA-1, Git tree SHA-1, tracked-file count, aggregate SHA-256 and resulting atomic root.

Directory campaign

For a non-Git directory, create a deterministic inventory before ingest: normalized relative path, byte size, SHA-256, media type and provenance/source. Exclude transient caches and secrets explicitly rather than relying on accidental ignore behavior.

Consistency checks

• Tracked/inventory count matches expectation.
• Every ingested object has an evidence hash.
• Failures/rejections are accounted for.
• Aggregate manifest is retained in the qualification archive.
• Re-running against the same immutable source does not invent a conflicting history.

Release repositories

When ingesting ENTITY itself, use the immutable tag/commit as provenance. The v3.4.2 release commit is 6dfa3d6cc738d9369cf092d2782676bf4f2a46e4.