Skip to content

Add evidence-backed fungal auxiliary cell formation - #1801

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

realmarcin merged 1 commit into
mainfrom
codex/add-fungal-auxiliary-cell-formation

Conversation

@realmarcin

@realmarcin realmarcin commented Oct 8, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Add traitmech:000671 fungal auxiliary cell formation as a PROPOSED
MORPHOLOGY class, with METPO proposal v547 / METPO:1062400 and reserved
block 1062400-1062499. The phenotype is external auxiliary-cell formation,
not an internal vesicle, subcellular organelle, storage-function claim or
sequence-profile assignment.

Evidence: Brundrett terminology,
Kuga 2008, and
Declerck 2004. All three have
contiguous snippets. One source-qualified MAFF520054 / NCBITaxon:4874 example;
NARO and Ohtomo 2025 provide provenance context only. Two OPEN discussions
retain hierarchy and functional uncertainties. No causal graph is asserted.

Includes a guarded dry-run writer, 20 writer tests, append-only CREATE history,
README counts and maintained citation/discussion/QC/page/priority products.
All 1,065 existing trait YAML files are byte-identical to the base, including
the protected spore-germination record. Existing pages differ only in corpus
footers and the phenotype parent's expected new child. Old histories, proposal
templates/narratives, embeddings and discussion shell assets are unchanged.

Validation

  • Passed: LinkML, strict record validation, predicate/node grounding, Biolink
    coverage, proposal verifier and ROBOT. RDF graphs contain 15 / 12,628 /
    12,632 triples, with correct w3id METPO parent and quality hierarchy.
  • Passed: online taxonomy (748 examples resolved, zero errors; 24 existing
    unrelated label warnings), history, products, Ruff and writer tests.
  • Snippet resolver: Declerck VERIFIED against the scientific abstract; Kuga
    NOT_IN_ABSTRACT, with the stored clause independently matched to Results.
    URL terminology checked directly; no resolver verdict was relabeled.
  • Desktop/mobile Playwright checks and inspected screenshots: 1440 / 390 px,
    three quotes, qualified example, two OPEN discussions, local identifier,
    browse/hierarchy navigation, dashboard 1,066, loaded image and no overflow
    or page errors.
  • Ordinary whitespace check flags only the ROBOT directive row's three
    required trailing empty cells; parsed 3-by-11 shape and exact-path scoped
    whitespace checks pass.
  • Embeddings not regenerated: both exact configured input paths are absent.
  • Passed: full committed just qc, committed-diff history audit against
    refreshed main, PR sanity, 65 focused tests (975.69s), and all 3,307 local
    tests (901.34s). CI also passed all 3,307 tests (213.41s) and all eight
    required checks on fb415fb4565ceaae9c2cbc82f2771f59161fb72d.

Novelty/allocation searches included ignored and hidden files, current main,
all local worktrees, CommunityMech and fully paginated immutable pending PR
heads, including drafts. Publication refresh includes PRs 1800, 1782, 1476,
973 and 924; no competing reservation. Proposal records exact heads.

Review and Landing

Request Copilot and perform an exact-head author adversarial review. Previously
verified Claude/Shepherd quota exhaustion is availability evidence only; no
independent review or approval is claimed from those failures. File and fix
genuine review defects, inspect late feedback, require all eight head and
combined-candidate checks, use the native queue, verify the actual landed
parent/tree, then delete the local and remote feature branch.

Upstream proposal and the two explicitly OPEN curation discussions remain for
future evidence-backed work. This addition does not exhaust microbial traits.

Author adversarial review found no actionable defect, so no defect issue was
fabricated: #1801 (comment).
Independent-review availability is recorded separately on
#742 (comment).
All 1,093 published changed blobs match the local commit. The subsequently
landed main commit 51f5eb2 changes only the
shared tooling pin; it has no trait/artifact overlap and is left intact.

@realmarcin

Copy link
Copy Markdown
Contributor Author

Author adversarial review of exact head
fb415fb4565ceaae9c2cbc82f2771f59161fb72d (not independent approval).

No actionable curation defect found. I attempted to falsify:

  • Novelty and allocation: ignored/hidden worktree and pending-head searches,
    historical auxiliary-body/external-vesicle aliases, neighbor definitions,
    obsolete mycorrhization and the entire reserved block. The additional
    extramatrical/extraradical alias scan was empty; current own matches are
    the new record and supporting artifacts, not competing reservations.
  • Biological scope: a germ-tube observation defeats a current-host-contact
    requirement; external auxiliary swellings are distinct from root-internal
    vesicles and BAS. Typical clustering does not require a universal cell count,
    ornamentation, storage role or propagule function. No unsupported exact
    structure synonyms, xrefs, SSSOM equivalence or causal graph were added.
  • Evidence: directly reread the terminology page and Kuga Methods/Results;
    retained NOT_IN_ABSTRACT for the body quote. Declerck's directly retrieved
    scientific abstract is the VERIFIED quote, not an editorial summary.
    Source accessibility, uninspected paper images and temperature-glyph
    uncertainty are stated rather than converted into verified facts.
  • Canonical example: species authority and MAFF520054 provenance are separate
    from phenotype evidence. Carrot transformation is not fungal engineering;
    ASV assignments are not culture counts, contamination proof or phenotype.
  • Implementation and outputs: parent/target/template drift guards,
    prevalidation and replay tests; correct 11-column ROBOT template and w3id
    parent hierarchy; append-only history with the actual Codex actor; all 1,065
    prior YAML records and old histories/proposals/embeddings preserved.
    Existing pages have only expected footer and parent-list deltas. Playwright
    checks and actual desktop/mobile screenshots passed.

The complete paginated published-file audit matched all 1,093 changed blobs
to this local commit, with the head rechecked before and after. Review,
inline-comment and discussion endpoints were empty at this pass.

No issue was fabricated for an access limitation or review-provider quota.
Local full validation remains running and this PR stays draft; this comment
does not authorize admission before all required head/candidate gates pass.

@realmarcin
realmarcin marked this pull request as ready for review October 8, 2026 02:39
@realmarcin

Copy link
Copy Markdown
Contributor Author

Final author review / admission receipt for exact head
fb415fb4565ceaae9c2cbc82f2771f59161fb72d.

All local validation is complete: 3,307 tests passed in 901.34s, 65 focused
tests passed in 975.69s, committed just qc, history, products, strict,
LinkML, taxonomy, snippets with their stated source limitations, proposal/
ROBOT/RDF, preservation and desktop/mobile checks passed. All eight required
Actions/App 15368 checks passed on this head; CI ran 3,307 tests in 213.41s.
No tracked working-tree drift remains. The original checkout is clean on main.

The exact-head author adversarial review and 1,093-file published blob audit
remain valid; no actionable defect was found. No new curation issue was filed.
The earlier review comment records the then-draft state; the PR was subsequently
marked ready on the same head. Copilot's successful empty assignment receipt
is not an independent review. Prior quota-blocked Claude/Shepherd routes and
the lack of independent approval are documented on #742, comment 6051074009.

Refreshed main has one intervening change, tooling pin commit
51f5eb2035505495ce817b0b420e2aeb81500df6; no trait/artifact conflict or
regeneration is required. The history audit against that refreshed base passed.
Proceed only through the configured native queue, inspect its actual candidate
checks, then verify the actual merged parent/tree before either branch deletion.

@realmarcin
realmarcin added this pull request to the merge queue Oct 8, 2026
@realmarcin

Copy link
Copy Markdown
Contributor Author

Native queue admission confirmed for exact head fb415fb: isInMergeQueue=true, position 1, state QUEUED. The normal gh pr merge --auto --match-head-commit path admitted this head; no fallback or bypass was needed. 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.

Merged via the queue into main with commit d59d8c1 Oct 8, 2026
8 checks passed
@realmarcin
realmarcin deleted the codex/add-fungal-auxiliary-cell-formation branch October 8, 2026 03:09
@realmarcin

Copy link
Copy Markdown
Contributor Author

Landing and cleanup verified for #1801.

  • MERGED at 2026-10-08T03:07:18Z; reviewed head
    fb415fb4565ceaae9c2cbc82f2771f59161fb72d.
  • Actual merge d59d8c1b16836c9fe457f7c1a0d479b6bfe083cc, actual parent
    51f5eb2035505495ce817b0b420e2aeb81500df6. All eight required head gates
    and all eight combined-candidate gates passed on their exact SHAs.
  • Maintained verify_merge_integrity.py completed with one verified landing,
    no errors and no mismatch. Expected and actual tree are both
    9c51904e57dc68d12f099f8bb8be1356f5a83ab5.
  • 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 --heads are empty
    for codex/add-fungal-auxiliary-cell-formation.
  • Original main fast-forwarded cleanly to the actual merge and equals
    origin/main. Other worktrees and their changes were preserved.

Validation: 3,307 local tests, 65 focused tests, full QC and domain/browser
checks passed. Author adversarial review found no actionable defect, so no
defect issue was fabricated. Independent reviewer availability remains recorded
on #742 (comment 6051074009), not represented as approval.

The new record retains its PROPOSED status, v547 upstream proposal and two
OPEN evidence/hierarchy discussions. Further microbial-trait leads 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