pumpingStation/test/_output-manifest.md

pumpingStation output manifest

Single source of truth for what this node emits and where it is tested, per .claude/rules/output-coverage.md. Generated against code-ref a83a85e. Regenerate the wiki contract with npm run wiki:all and re-check this table whenever getOutput(), src/commands/index.js, or an examples/*.json fan-out changes.

Null convention for this node: a Port-0 key whose source is not yet available is emitted as explicit null (e.g. timeleft, flowSource, manualDemand outside manual mode), never silently absent. Delta-compression on Port 0 then drops keys whose value is unchanged since the previous tick.

Port 0 (process data) — specificClass.getOutput() → outputUtils.formatMsg(..., 'process')

msg.topic = config.general.name. Keys below are the full pre-delta-compression set.

Key	Source	Type	States tested	Test file
mode	getOutput ← this.mode	string (levelbased/manual/flowbased/none)	populated (manual)	test/basic/specificClass.test.js
manualDemand	getOutput ← _manualDemand	number m³/h, null outside manual	populated, null	test/basic/specificClass.test.js
direction	getOutput ← state.direction	string (filling/draining/steady)	present	test/basic/specificClass.test.js
flowSource	getOutput ← state.flowSource	string, null when no source	null (pre-child)	test/basic/specificClass.test.js
timeleft	getOutput ← state.seconds	number s, null when steady	present, null	test/basic/specificClass.test.js
percControl	getOutput ← controlState.percControl	number % 0..100	0, 25, 50, 75, 85, 100	test/basic/specificClass.test.js
dryRunLevel	_computeSafetyPoints	number m	populated	test/basic/specificClass.test.js
dryRunSafetyVol	_computeSafetyPoints	number m³	populated	test/basic/specificClass.test.js
highVolumeSafetyLevel	_computeSafetyPoints	number m	populated	test/basic/specificClass.test.js
highVolumeSafetyVol	_computeSafetyPoints	number m³	populated	test/basic/specificClass.test.js
predictedOverflowVolume	measurements overflowVolume	number m³	populated, 0	test/basic/specificClass.test.js
predictedOverflowRate	measurements flow.overflow	number m³/s	populated, 0	test/basic/specificClass.test.js
predictedUnderflowVolume	measurements underflowVolume	number m³	0	test/basic/specificClass.test.js
volume.predicted.atequipment.<childId>	measurements.getFlattenedOutput	number m³	populated	test/basic/specificClass.test.js
basin geometry: heightBasin, surfaceArea, maxVol, minVol, maxVolAtOverflow, minVolAtInflow, minVolAtOutflow, volEmptyBasin, inflowLevel, outflowLevel, overflowLevel, inletPipeDiameter, outletPipeDiameter, minHeightBasedOn	basin.snapshot()	number (m/m²/m³) / string	populated	test/basic/specificClass.test.js, test/basic/BasinGeometry.basic.test.js

Port 1 (InfluxDB telemetry) — formatMsg(..., 'influxdb')

Same key set as Port 0 (formatted via the influxdb formatter rather than process). Field names == Port-0 keys; config.general.name is the measurement tag. No Port-1-only fields. Covered transitively by the Port-0 tests above; a dedicated Port-1 line-protocol assertion is a gap (see below).

Port 2 (registration / control plumbing) — BaseNodeAdapter._scheduleRegistration

Topic	Source	Payload shape	States tested	Test file
child.register	BaseNodeAdapter.js:122	{ topic:'child.register', payload:<node.id>, positionVsParent, distance }	—	(gap — see below)

Note: the canonical outgoing topic is child.register (matching the input registry). Earlier docs said registerChild; that is the deprecated input alias, not what this node emits.

Child-facing events — measurements.emitter

Fired as <type>.<variant>.<position> when a series receives a value. Parents subscribe by event name (data-driven, not a fixed catalogue):

Event	When	Test file
volume.predicted.atequipment	each integrator tick	test/basic/flowAggregator.basic.test.js
level.predicted.atequipment	recomputed from volume	test/basic/specificClass.test.js
flow.predicted.in (child manual-qin)	set.inflow handler	test/basic/measurementRouter.basic.test.js
overflowVolume/underflowVolume/flow.predicted.overflow	integrator hits a physical bound	test/basic/flowAggregator.basic.test.js

Example-flow function-node fan-out

examples/02-Dashboard.json :: fn_status_split (outputs: 15)

#	Target widget	Payload	Populated	Degraded/null
0	ui-text "Mode"	string	✔ structure	gap
1	ui-text "Direction"	string	✔	gap
2	ui-text "Level"	number m	✔	gap
3	ui-text "Volume"	number m³	✔	gap
4	ui-text "Volume %"	number %	✔	gap
5	ui-text "percControl"	number %	✔	gap
6	ui-text "Manual demand"	number m³/h or —	gap	gap
7	ui-chart "Level (m)"	{topic,payload:number} or no-msg	✔	gap
8	ui-chart "Volume (m³)"	″	✔	gap
9	ui-chart "Volume %"	″	✔	gap
10	ui-chart "Flow (m³/h)" — Inflow	″	✔	gap
11	ui-chart "Flow (m³/h)" — Outflow	″	✔	gap
12	ui-chart "Flow (m³/h)" — Net	″	✔	gap
13	ui-template "Raw output table"	whole object (array)	✔	gap
14	ui-chart "percControl"	{topic:'percControl',payload:number}	✔	gap

Populated/structure coverage: test/integration/basic-dashboard-flow.test.js (asserts output count = 15 and routes outputs 0–14). Degraded/empty-input coverage (no payload:null reaching any ui-chart) is still a gap — see below.

Known coverage gaps (tracked, prospective per the rule)

The output-coverage rule applies prospectively. Outstanding items for this node:

• Dedicated test/basic/output-port0.test.js exercising every key above in both populated and degraded (pre-tick / null) states.
• Port-1 line-protocol assertion (field names + tag).
• Port-2 child.register payload-shape test.
• fn_status_split degraded/empty-input fan-out test (no payload:null to any ui-chart) — the failure mode the rule was written for. The structure test in basic-dashboard-flow.test.js covers the populated path only.