Skip to content

Add evidence-backed baeocyte formation - #1803

Merged
realmarcin merged 1 commit into
mainfrom
codex/add-baeocyte-formation
Oct 8, 2026
Merged

realmarcin merged 1 commit into
mainfrom
codex/add-baeocyte-formation

Conversation

@realmarcin

@realmarcin realmarcin commented Oct 8, 2026 •

Copy link
Copy Markdown
Contributor

Summary

  • Add traitmech:000672, baeocyte formation, a PROPOSED MORPHOLOGY class under phenotype METPO:1000059.
  • Reserve v548 / METPO:1062500 (whole block 1062500-1062599) with a matching 11-column ROBOT class template and round-trip plan.
  • Add two source snippets, one non-axenic KI4 example, two OPEN scope/mechanism discussions, a guarded dry-run writer, 20 regression tests and append-only Codex history.

Evidence And Scope

  • Definition authority: https://doi.org/10.1007/s10482-021-01672-x. Direct publisher genus/species descriptions, Collection Methods, actual Figure 4 and caption inspected. Its body quote is NOT_IN_ABSTRACT, independently checked against the body; that verdict is not relabeled.
  • Division-sequence evidence: https://pubmed.ncbi.nlm.nih.gov/18365724/. Direct English scientific abstract; snippet resolver VERIFIED. Russian full text and figures were not inspected.
  • Canonical example: NCBITaxon:2874699, Waterburya agarophytonicola KI4. NCBI authority independently checked; strain-level rank is no rank. Culture conditions remain source-qualified; C-A-99685 is an herbarium specimen, not a living-culture accession. No genome-derived phenotype or protein mechanism asserted.
  • No invariant cell number/size, simultaneous cleavage, release route, motility, dormancy, sheath state, exact synonyms, xrefs or organism-level disjointness asserted. Dormant-endospore sporulation and coenobium formation are not semantic duplicates.

Ignored-and-hidden novelty/allocation checks covered complete local worktrees, the CommunityMech checkout, pinned METPO and fully paginated open-PR heads including drafts. The publication refresh also inspected #1802 at ed0caa0f68592dc4ab1ee905fa80e5bb58c1efd3: it updates existing Shedu traitmech:000220, with no conflicting allocation. Other pending heads remain #1782, #1476, #973 and draft #924.

Validation

Passed locally:

  • New writer tests: 20 passed; dry run, replay, fail-closed parent/target/template guards and prevalidation.
  • Target LinkML and strict validation; predicate/node grounding; Biolink coverage.
  • Proposal verification and ROBOT/ELK; parsed RDF confirms w3id IRIs and phenotype/quality ancestry (15 class, 12,628 merged, 12,632 reasoned triples).
  • History validation, product ID-label validation, Ruff and online canonical-example audit: 749 resolved examples, zero errors (24 existing label-drift warnings).
  • Maintained citation/discussion/QC-dashboard/page/priority generation; 552 proposed records pass the two-citation gate.
  • Preservation audit: all 1,066 prior trait YAMLs, old histories/proposals/embeddings and discussion shell unchanged. Old trait-page differences are footer totals plus the phenotype child list. Existing priority rows change only phenotype's child count, 167 to 168.
  • Playwright and actual screenshots at 1440 and 390 pixels: two quotes, one qualified example, two OPEN discussions, correct local-ID provenance, browse/hierarchy navigation, dashboard count/image, no overflow or page errors.
  • Scoped whitespace checks. Ordinary staged check reports only the inspected ROBOT directive row's three required trailing tabs; no global exception used.

Full local pytest: 3,327 passed in 960.88s. Focused artifact tests: 65 passed in 520.07s. Committed just qc, history-base audit and PR sanity passed. All eight required Actions checks passed on d3117c2e75074da4538ef7d95e324ff9d7d6a68e; CI ran 3,327 tests in 431.00s. Native combined-candidate checks remain a separate requirement.

Configured embedding inputs were checked at their exact paths and are absent; existing embedding artifacts were preserved, not regenerated. No protected record was edited.

Review And Landing

Exact-head author adversarial review found no actionable defect; all 1,094 published changed-file blobs match the local commit. The earlier comment preserves its then-draft/pending status. Independent approval is not presumed. Previously exhausted Claude/Shepherd routes are tracked on #742; quota failures are not clean reviews or curation defects. Final feedback inspection and the Copilot request are recorded separately.

Merge only through the configured native queue after all gates pass. Verify actual queue membership and exact combined-candidate checks, then actual MERGED state and maintained actual-parent/tree integrity before deleting either feature branch.

@realmarcin

Copy link
Copy Markdown
Contributor Author

Author adversarial review of exact head
d3117c2e75074da4538ef7d95e324ff9d7d6a68e (not independent approval).

No actionable curation defect found. I tried to falsify:

  • Novelty and allocation: ignored/hidden whole-worktree, ontology, historical-proposal and complete pending-head searches. The old morphology frontier is not an allocated record. Coenobial progeny and another bacterium's multiple fission do not carry this cyanobacterial reproductive-cell scope. The new Add literature-backed SduA protein example to Shedu system #1802 edits existing Shedu and reserves none of the new IDs.
  • Hierarchy and semantics: active phenotype/quality ancestry, obsolete reproductive/fission terms, the local dormant-endospore meaning of sporulation, and the colony-development criterion of coenobium formation. Historical nanocyte/endospore labels are not promoted to exact synonyms or xrefs.
  • Evidence: the first snippet is an exact publisher genus-Description sentence and remains NOT_IN_ABSTRACT; the second is the complete first sentence of the directly retrieved English scientific abstract, resolver VERIFIED. The later abstract sentence links multiple fission to baeocytes. No search-result quote, digest, title or reconstructed fragment is used.
  • Morphology versus mechanism: actual Figure 4 and caption were checked, including panel b's adjacent-cell warning. Neither the images nor predicted genes establish a universal protein mechanism, fixed count, release route or dormancy. Russian full paper and supplements remain explicitly uninspected.
  • Canonical example: NCBI resolves the exact KI4 strain label at a no-rank taxon, not a species ID. Culture provenance, non-axenic status, distinct culture descriptions and the herbarium-only accession are retained. No CALU 1126 canonical assignment is made without provenance verification.
  • Implementation and artifacts: dry-run/default, prevalidation, replay and drift-refusal tests; 11-column template and parsed w3id RDF hierarchy; correct append-only Codex actor/history; all 1,066 old YAMLs and old history/proposals/embeddings byte-preserved. Existing page changes are footer totals plus phenotype's child list; desktop/mobile screenshots and navigation checks passed.

The full paginated published-file audit matched all 1,094 changed blobs to this commit, with the PR head checked before and after. Formal reviews, inline comments and discussion endpoints were empty at this pass. No defect issue was fabricated for unavailable reviewers.

This is the then-draft review receipt. Full local tests/QC and CI remain in progress; this comment does not authorize queue admission before every required gate and final feedback check passes.

@realmarcin
realmarcin marked this pull request as ready for review October 8, 2026 03:43
@realmarcin

Copy link
Copy Markdown
Contributor Author

Final author review / admission receipt for exact head
d3117c2e75074da4538ef7d95e324ff9d7d6a68e (not independent approval).

All local work is validated: 3,327 tests passed in 960.88s, 65 focused tests
passed in 520.07s, and committed QC/history/products/PR-sanity, strict,
LinkML, proposal/ROBOT/RDF, taxonomy, snippets with stated limitations,
preservation and desktop/mobile checks passed. All eight required Actions
checks (App 15368) passed on this head; CI ran 3,327 tests in 431.00s.

The author adversarial review and 1,094-blob published-file audit remain valid.
No actionable defect was found and no curation issue was fabricated. The
earlier comment records the then-draft state; this head is now ready.
Copilot was requested once, with no observable assigned reviewer or review.
The prior exhausted Claude/Shepherd routes are documented on #742; none of
these availability outcomes is an independent approval.

Freshly fetched main is still d59d8c1b16836c9fe457f7c1a0d479b6bfe083cc;
no base overlap or regeneration is needed. The tracked working tree is clean.
Use only the configured native queue, inspect its combined candidate, then
verify actual MERGED state and actual-parent/tree integrity before deleting
either branch. The two OPEN scope/mechanism discussions and v548 upstream
proposal remain explicitly unresolved.

@realmarcin

Copy link
Copy Markdown
Contributor Author

Native queue admission confirmed for exact head d3117c2: isInMergeQueue=true, position 1, state QUEUED. The normal gh pr merge --auto --match-head-commit path admitted this head; no fallback or bypass was used. Initial combined headCommit is null, so candidate checks are not yet claimed. Both feature branches remain intact pending actual MERGED state and actual-parent tree verification. Reviewer availability receipt: #742 (comment).

Merged via the queue into main with commit f16cadb Oct 8, 2026
8 checks passed
@realmarcin
realmarcin deleted the codex/add-baeocyte-formation branch October 8, 2026 03:57
@realmarcin

Copy link
Copy Markdown
Contributor Author

Landing and cleanup verified for #1803.

  • MERGED at 2026-10-08T03:55:10Z; reviewed head
    d3117c2e75074da4538ef7d95e324ff9d7d6a68e.
  • Actual merge f16cadbfc9d8d11f95dcf751e456b0fc2ebef068, actual parent
    d59d8c1b16836c9fe457f7c1a0d479b6bfe083cc. All eight required head
    checks and all eight combined-candidate checks passed on their exact SHAs.
  • Maintained merge-integrity reconstruction completed: one verified landing,
    no errors or mismatch. Expected and actual tree both
    7f92f0718e8eb70f56a7e77d66a02c80281839b6.
  • Remote feature branch deleted with the exact reviewed-head lease; local
    branch deleted after detaching the retained worktree at the reviewed head.
    Fetch/prune completed; local refs and remote ls-remote show no
    codex/add-baeocyte-formation branch.
  • Original main fast-forwarded cleanly and equals origin/main at the merge.
    Other worktrees and their changes were preserved.

Validation: 3,327 local tests, 65 focused tests, complete committed QC and
domain/browser checks passed. Author adversarial review found no actionable
defect; no curation issue was filed. Independent reviewer availability is
documented on #742, comment 6051765070, and is not represented as approval.

The record retains PROPOSED status, its v548 upstream proposal and two OPEN
scope/mechanism discussions. Further microbial-trait candidates remain;
discovery is not declared exhausted.

realmarcin added a commit that referenced this pull request Oct 8, 2026
Adds tests/test_add_shedu_sdua_protein_example.py in the style of the
writer tests #1801 and #1803 shipped: the writer is pointed at a temp
copy via monkeypatched TARGET/REPO_ROOT, driven through main() with and
without --apply, and judged by before/after snapshots of the temp tree.

The writer edits an existing record, so the base it is applied to is
derived from the committed record by removing exactly the node, edge
and curation event it adds, each asserted to occur once. That derived
base was checked to equal the origin/main record.

Covered:
- dry run is the default and writes nothing, here or in the repo;
- applying to the base emits exactly the committed node (B7HFR2,
  NCBITaxon:405532, REVIEWED, entry v58 / seq v1, label-only grounding),
  the committed edge, and one curation event with the writer's
  timestamp, and changes nothing else in the record;
- refusal when sdua_immune_nuclease already exists, both after a first
  apply and on the committed record itself;
- refusal for a different identifier and unless exactly one
  shedu_dna_end_nicking graph exists, and when the edge's object node
  is missing;
- a closed-schema validation failure writes nothing.

An autouse fixture makes any socket connection fail the test, so the
suite cannot reach UniProt or Europe PMC. Mutating the writer's entry
version or timestamp makes the round-trip test fail.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant