From the adversarial review of #436. Pre-existing and repo-wide, not introduced by that PR — filed so the divergence is recorded rather than assumed fixed.
#436 corrects NCBITaxon:169215 (the plant genus Bosea) to NCBITaxon:85413 in the YAML, but two tracked, published artifacts still carry the old id:
1. docs/communities/MSC1_Dominant_Core.html:612-614,1851 and docs/communities/EcoFAB_Ring_Trial_SynCom17.html:684-686,2195 render NCBITaxon:169215 as a live link to ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=169215 — the plant — for a bacterium. generate-pages.yaml serves committed docs/ verbatim, so the published site now contradicts the KB on this id.
This is not specific to Bosea: docs/communities/ has not been regenerated since #268, and other data PRs have drifted the same way — e.g. docs/communities/KBase_ORT_Workflow_Community_Model.html still shows NCBITaxon:2157/1236 where the YAML has had 1033997/2689614 since 2a3b691. The real issue is that nothing regenerates or gates docs/ against the KB.
2. reports/gtdb_denominators.tsv:152 still carries NCBITaxon:169215 Bosea GTDB:g__Bosea ..., handing the plant id four bacterial GTDB outcomes. scripts/gtdb_denominator_compare.py enumerates KB term.ids, so the row is now orphaned. The script documents the report as a non-gated snapshot, so staleness is expected — but this particular row is the exact defect class the gate now blocks.
Worth deciding whether docs/ regeneration belongs in CI, or whether these artifacts should stop being committed.
From the adversarial review of #436. Pre-existing and repo-wide, not introduced by that PR — filed so the divergence is recorded rather than assumed fixed.
#436 corrects
NCBITaxon:169215(the plant genus Bosea) toNCBITaxon:85413in the YAML, but two tracked, published artifacts still carry the old id:1.
docs/communities/MSC1_Dominant_Core.html:612-614,1851anddocs/communities/EcoFAB_Ring_Trial_SynCom17.html:684-686,2195renderNCBITaxon:169215as a live link toncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=169215— the plant — for a bacterium.generate-pages.yamlserves committeddocs/verbatim, so the published site now contradicts the KB on this id.This is not specific to Bosea:
docs/communities/has not been regenerated since #268, and other data PRs have drifted the same way — e.g.docs/communities/KBase_ORT_Workflow_Community_Model.htmlstill showsNCBITaxon:2157/1236where the YAML has had1033997/2689614since 2a3b691. The real issue is that nothing regenerates or gatesdocs/against the KB.2.
reports/gtdb_denominators.tsv:152still carriesNCBITaxon:169215 Bosea GTDB:g__Bosea ..., handing the plant id four bacterial GTDB outcomes.scripts/gtdb_denominator_compare.pyenumerates KBterm.ids, so the row is now orphaned. The script documents the report as a non-gated snapshot, so staleness is expected — but this particular row is the exact defect class the gate now blocks.Worth deciding whether
docs/regeneration belongs in CI, or whether these artifacts should stop being committed.