EventRecord

---

tags: [domain-model] store: events keyPath: "id" db: wodwiki-db (v29)

EventRecord

[!info] Current vs Future State This document distinguishes the current implementation in code from proposed/future-state enhancements.

Current State (Implemented in Code)

• Store: events
• Key path: id
• Type source: packages/core/src/types/storage.ts
• Database version: wodwiki-db (v23)

THE single stored record for workout data (renamed from UnifiedEventRecord). Grain event rows are the archival statement stream: the engine's output statements land here directly (V21), and Session rows carry execution metadata only. Grain summary rows hold derived aggregates; wellness rows are reconciled from note content. projectEventToFacts folds rows into query facts; eventsToStoredLogs rebuilds the statement stream for display and replay.

Fields (Current)

Field	Type	Notes
id	string	${resultId}:${seq} event | ${resultId}:summary:${metricKey}[:k=v…] summary | wellness:${noteId}:${key}
resultId	string	Session for workout rows; synthetic wellness:<noteId> source for note-derived wellness
noteId	string	FK → Note
blockContentId?	string	Content-stable cross-workout join key → BlockIndexRow
pageId?	string	@deprecated V10 legacy placement field (placements now owned by PageNote)
origin?	ResultOrigin
timestamp	number	Canonical time — when the workout happened
grain	EventGrain	'event' raw statement | 'summary' folded row
effortSlug?	string	FK → Effort.slug
outputType	string	Open vocabulary — KNOWN_OUTPUT_TYPES: segment, system, load, event, compiler, completion, analytics, wellness
metrics	StoredOutputStatement['metrics']	Typed metric array; EXACTLY ONE entry on summary rows
timeSpan?	{ started: number; ended?: number }
sourceBlockKey?	string
stackLevel?	number
completionReason?	string
blockId? / blockVersion?	string / number	Occurrence pin → BlockRevision [noteId, blockId, version]
metricTemporal?	MetricTemporal[]	Per-metric temporal anchors — instant or civil-date (ticket 12)
metricDateKeys?	string[]	d:YYYY-MM-DD keys — multiEntry index source (ticket 14)
representationKind?	'direct' | 'calculated' | 'substitute_summary'
summaryCoverage?	SummaryCoverage	Scope a substitute summary covers
reducerStats?	ReducerStats	Retained stats for substitution proofs

Indexes (Current)

Index	Key path	Unique	Purpose
by-timestamp	timestamp	no	the one proven culling index
by-result-grain	[resultId, grain]	no	finalize clear, per-result fetch, orphan GC
by-content-grain	[blockContentId, grain]	no	blockContentId join hot path
by-effort	effortSlug	no
by-outputType	outputType	no
by-grain	grain	no
by-metric-date	metricDateKeys	no	multiEntry civil-date keys (V17)

Relationships (Current)

Outgoing (this row references)

• resultId → Session for workout rows; wellness uses a synthetic per-note source ID, not a result foreign key
• noteId → Note
• pageId → Page — legacy denormalized context
• effortSlug → Effort
• blockId + blockVersion → BlockRevision — pinned saved block state (workout rows)
• blockContentId → BlockIndexRow — content-hash join

---

Future State (Proposed)

Typed-note composition recommendations

• Keep event/query representation separate from authored source and recorded-result ownership. A reusable graph consumes query results; it must not write event rows as an alternative editor-save path.
• Preserve source identity and provenance through note deletion/re-key or projection rebuild. Runtime output type segment is not a new BlockRevision kind.
• Do not create another page-membership projection here. Existing pageId semantics follow Page#Recommended composition contract; query-derived membership does not rewrite events.

Feedback case: a workout note appears in a second collection or switches to read presentation. Its recorded facts must remain attached to the original run, not be duplicated or reassigned because presentation changed. See Session#Typed-note composition recommendations.

These are review proposals; the session-renaming proposal below is independent and remains unimplemented.

Proposed Structure (Session Renaming)

In the proposed future state:

• resultId transitions to sessionId pointing to Session.
• Event ID formatting updates from ${resultId}:${seq} to ${sessionId}:${seq}.
• Index by-result-grain updates from [resultId, grain] to [sessionId, grain] (by-session-grain).

Map

!