Skip to content

p-core-034-evaluation-bundle.md

Latest commit

 

History

History
90 lines (64 loc) · 4.08 KB

p-core-034-evaluation-bundle.md

File metadata and controls

90 lines (64 loc) · 4.08 KB
Raw

P-CORE-034 Evaluation Bundle Review Packet

This packet makes one public Permea artifact system reviewable from GitHub. It is intended for human review and structured assisted review.

It should be read together with the linked source files, tests, report, and validation command output.

Packet Metadata

Field Value
Packet ID p-core-034-evaluation-bundle
Artifact path docs/examples/generated/EVALUATION_PACKET.md
Artifact type public evaluation template and reference packet surface

Purpose

Help a reviewer inspect the evaluation bundle through its generated packet, template inputs, source module, validation handoff, reproducibility handoff, evidence record, tests, boundaries, and limitations.

Related Evidence And Report Links

Review surface Link
scripts/permea_evaluate.py scripts/permea_evaluate.py
scripts/permea_reproduce.py scripts/permea_reproduce.py
scripts/permea_validate.py scripts/permea_validate.py
src/permea_core/evaluation/bundle.py src/permea_core/evaluation/bundle.py
docs/examples/generated/EVALUATION_PACKET.md docs/examples/generated/EVALUATION_PACKET.md
docs/examples/generated/EVALUATION_PACKET.json docs/examples/generated/EVALUATION_PACKET.json
docs/evidence/EVIDENCE-034-evaluation-bundle.md docs/evidence/EVIDENCE-034-evaluation-bundle.md
EVALUATION.md EVALUATION.md
tests/test_evaluation_bundle.py tests/test_evaluation_bundle.py

Validation Commands

python3 scripts/permea_evaluate.py
python3 scripts/permea_reproduce.py --report-only
python3 scripts/permea_validate.py
python3 -m pytest tests/test_evaluation_bundle.py

Raw Readability Notes

  • The evaluation packet points to generated Markdown and JSON packet outputs.
  • This markdown packet is intentionally written as physical newline-separated text.
  • This JSON packet is intentionally written with indent=2, sort_keys=True, and a trailing newline.
  • Use commit-SHA raw URLs for external review when branch raw views may be stale or transformed.

Claim Boundary Notes

  • This packet reviews a template/reference evaluation workflow only.
  • It does not load datasets, execute acquisition, call external services, run ML, or score candidates.
  • It does not create scientific evidence, benchmark results, or biological validation.
  • It does not claim wet-lab validation, clinical efficacy, model performance, or solved delivery.
  • A passing packet means the listed evaluation surface is easier to inspect; it does not prove biological outcomes.

Reviewer Checklist

  • Open the evaluation packet Markdown and JSON outputs and confirm the input-family links are public and existing.
  • Run the listed evaluation, reproduction, validation, and focused test commands from the repository root.
  • Confirm validation and reproducibility handoffs are visible in the generated evaluation packet.
  • Confirm limitations and non-claims remain explicit in the packet and evidence record.
  • Record any missing input family, stale command, unclear boundary, or validation failure before approval.

Limitations

  • The packet covers a reusable template/reference workflow, not a completed external evaluation result.
  • The packet checks reviewability and local generation coverage, not scientific correctness.
  • The evaluation bundle does not establish access rights, redistribution status, biological performance, or clinical utility.

Next Review Step

Regenerate this packet:

python3 scripts/permea_review_packet.py

Then inspect this packet together with the linked report, generated artifacts, and command output.