fix: correct genesis op indexing and backfill missing content_ids from snapshots
Three-part fix for the genesis provenance timeline:
1. backfill_genesis_ops — sets op='add' on the oldest history entry for every symbol where the genesis bug caused it to be recorded as 'modify'. Ran on localhost: corrected 7,365 birth entries. Deploy script at deploy/backfill_genesis_ops.py.
2. backfill_content_ids_from_snapshots — fills missing content_id values on file-level history entries by reading the canonical source (snapshot manifests). Covers move/rename ops where the indexer recorded no new_content_id. Ran on localhost: filled 2,578 entries. Deploy script at deploy/backfill_content_ids_from_snapshots.py.
3. ui_symbols.py version boundary fix — replaces prev_cid is not None guard with entry_idx > 0 so the first real body is always flagged as a version boundary even when the birth entry has no content_id.
Tests: 6 tests for backfill_genesis_ops, 8 tests for backfill_content_ids_from_snapshots — all green.
0 comments
muse hub commit comment sha256:f3fe82ccf0a82f7b44f8f25d4c0dfff00583525ce318c439557087f48c7d6e40 --body "your comment"
No comments yet. Be the first to start the discussion.