Found by the adversarial review of the uncommitted #1050 fix, upheld 2/3. #1050 itself is otherwise met; this is the residual.
Joining data/transformed/microbedecoder/edges.tsv close_match objects against data/transformed/gtdb/nodes.tsv: 15,584 total, 15,524 declared, 60 missing.
| class |
count |
example |
| base accession absent, but gtdb declares the GCF/GCA twin |
41 |
ncbi.assembly:GCF_011008945.1 → gtdb has GCA_011008945.1 |
| gtdb declares a different version |
12 |
GCA_000803105.1 → gtdb has GCF_000803105.2 |
| absent entirely |
7 |
|
They are visible: the new invariants report them as unexpected ncbi.assembly stubs, and docs/MERGE_INTEGRITY_REPAIRS.md mentions "60 assembly accessions absent from the current GTDB transformed release". But 41 could resolve through gtdb's same_as twin pairing (#882), and 12 are version skew between MicrobeDecoder's snapshot and the pinned GTDB release.
Acceptance: decide per class — resolve the 41 through the twin mapping, record the 12 as version skew with the releases involved, and leave only genuinely absent assemblies as stubs.
Found by the adversarial review of the uncommitted #1050 fix, upheld 2/3. #1050 itself is otherwise met; this is the residual.
Joining
data/transformed/microbedecoder/edges.tsvclose_matchobjects againstdata/transformed/gtdb/nodes.tsv: 15,584 total, 15,524 declared, 60 missing.ncbi.assembly:GCF_011008945.1→ gtdb hasGCA_011008945.1GCA_000803105.1→ gtdb hasGCF_000803105.2They are visible: the new invariants report them as unexpected
ncbi.assemblystubs, anddocs/MERGE_INTEGRITY_REPAIRS.mdmentions "60 assembly accessions absent from the current GTDB transformed release". But 41 could resolve through gtdb'ssame_astwin pairing (#882), and 12 are version skew between MicrobeDecoder's snapshot and the pinned GTDB release.Acceptance: decide per class — resolve the 41 through the twin mapping, record the 12 as version skew with the releases involved, and leave only genuinely absent assemblies as stubs.