Quality

Requirements, tests, evidence

A reviewed requirement matrix per component; test titles that carry the requirement IDs; one normalized evidence stream across testthat and vitest; drift guards that fail CI when any committed artifact stops matching a fresh run.

265reviewed requirements11 matrices

441distinct test recordsjs-unit · r-unit · text-review

99%requirement coverage263 evidenced

99%tests passing0 failing · 5 pending review

Every requirement ID referenced by a test resolves to a reviewed matrix row.

2 reviewed requirements not yet referenced by any test: QC-CI-001, QC-CI-002

Components

Component	Kind	Matrix	Reqs	Tests	Result	Status
TFL Engine tfl-engine	Engine	tfl-engine.md	62	74	74 passing	evidenced
Text Library text	Text Library	text.md	85	155	5 pending review	evidenced
Report Template Library templates	Report Templates	templates.md	25	30	30 passing	evidenced
Traceability traceability	Framework	traceability.md	16	23	23 passing	evidenced
Quality Framework quality	Framework	quality.md	63	155	155 passing	evidenced
Subject Disposition t-disposition	Display	displays.md	14	4	4 passing	evidenced
Demographic and Baseline Characteristics t-demographics	Display	displays.md	14	4	4 passing	evidenced
Extent of Exposure to Study Drug t-exposure	Display	displays.md	14	4	4 passing	evidenced
Overview of Treatment-Emergent Adverse Events t-ae-overview	Display	displays.md	14	7	7 passing	evidenced
Treatment-Emergent Adverse Events by System Organ Class and Preferred Term t-ae-common	Display	displays.md	14	7	7 passing	evidenced
Listing of Serious Adverse Events l-ae-serious	Display	displays.md	14	3	3 passing	evidenced

How the loop works

1MatrixA reviewed Markdown table in quality/requirements/, one row per requirement.
2Test titleTests name the IDs they evidence: "DSP-AE-001: overview counts match ADAE (#1)". Guard tests in both suites enforce the convention.
3Evidencescripts/evidence.mjs normalizes testthat and vitest into one evidence.json per module, with the data → ARD → display traceability object attached.
4PageThis site joins the matrix extract to the evidence set and publishes it.
5Guardsrequirements:check and evidence:check fail CI when a committed artifact drifts from a fresh run.

Requirement ID scheme

All IDs match ^[A-Z]{2,4}-[A-Z]+-\d+[A-D]?$ — unchanged from safety.viz, because every downstream tool depends on it.

TFL- TFL Builder engine
DSP- per-display requirements
TXT- Text Library blocks
RPT- Report template assembly
TRC- traceability and provenance
QC- the quality framework itself

Test naming, both suites: "<REQ-ID>[, <REQ-ID>]: <description> (#1)".

Reproduce

npm ci
npx vitest run                         # the JS suite
Rscript qc/run-tests.R                 # testthat → qc/testthat-results.json
node scripts/requirements.mjs          # matrices → docs/requirements/*.json
node scripts/evidence.mjs              # both suites → docs/evidence/*/evidence.json
node scripts/requirements.mjs --check  # drift guard
node scripts/evidence.mjs --check      # drift guard
node scripts/site.mjs                  # this site → site/_build/