Collapse topic saturation into semantic graph
This commit is contained in:
Meta Vens
parent
fc29ddccb2
commit
f5788502be
6 changed files with 19 additions and 125 deletions
|
|
@ -2940,8 +2940,6 @@ function renderRepoFaceRoomTextureFacts(
|
|||
"- These are not stored needs and not orders. Project whether this character gets mischievous, bored, sharp, withdrawn, socially hungry, status-testing, or still work-focused.",
|
||||
]
|
||||
: [];
|
||||
const topicSaturationFacts = renderRepoFaceTopicSaturationFacts(identity, input.recentMessages);
|
||||
|
||||
return [
|
||||
"Room texture facts:",
|
||||
`- Observed messages: ${stats.total}; humans: ${stats.humanMessages}; agents/bots: ${stats.agentMessages}; distinct speakers: ${stats.speakerCount}.`,
|
||||
|
|
@ -2949,7 +2947,6 @@ function renderRepoFaceRoomTextureFacts(
|
|||
`- This Face's own recent messages in the attached window: ${stats.ownMessages}.`,
|
||||
`- Structural texture: ${stats.texture}. This is evidence about conversational weight, not a command to speak or joke.`,
|
||||
...temporaryPressures,
|
||||
...(topicSaturationFacts ? [topicSaturationFacts] : []),
|
||||
].join("\n");
|
||||
}
|
||||
|
||||
|
|
@ -2969,6 +2966,8 @@ interface RepoFaceCuriosityCluster {
|
|||
prominence: number;
|
||||
saturation: number;
|
||||
novelty: number;
|
||||
recentOverlap: number;
|
||||
stateOverlap: number;
|
||||
clusterDensity: number;
|
||||
jurisdictionFit: number;
|
||||
evidence: string[];
|
||||
|
|
@ -3231,6 +3230,8 @@ function decodeRepoFaceCuriosityGraph(
|
|||
prominence,
|
||||
saturation,
|
||||
novelty,
|
||||
recentOverlap,
|
||||
stateOverlap,
|
||||
clusterDensity: density,
|
||||
jurisdictionFit,
|
||||
evidence: cluster
|
||||
|
|
@ -3261,11 +3262,11 @@ function renderRepoFaceCuriosityClusters(
|
|||
: "local vector shards";
|
||||
return [
|
||||
"Curiosity graph attractors:",
|
||||
`- Source: semantic retrieval from ${backend}; local graph decoding over retrieved chunks. This is curiosity weather, not orders or consensus.`,
|
||||
...clusters.map((cluster) => {
|
||||
`- Source: semantic retrieval from ${backend}; local graph decoding over retrieved chunks. Saturation is derived here from semantic overlap with recent conversation and Face state, not from a separate repeated-term detector. This is curiosity weather, not orders or consensus.`,
|
||||
...clusters.map((cluster, index) => {
|
||||
const suggestedMotion = suggestCuriosityMotion(identity, cluster);
|
||||
return [
|
||||
`- ${cluster.label || "unnamed cluster"}: prominence ${formatSignal(cluster.prominence)}, saturation ${formatSignal(cluster.saturation)}, novelty ${formatSignal(cluster.novelty)}, cluster density ${formatSignal(cluster.clusterDensity)}, jurisdiction fit ${formatSignal(cluster.jurisdictionFit)}.`,
|
||||
`- Semantic cluster ${index + 1}: prominence ${formatSignal(cluster.prominence)}, semantic saturation ${formatSignal(cluster.saturation)}, novelty ${formatSignal(cluster.novelty)}, recent-conversation overlap ${formatSignal(cluster.recentOverlap)}, Face-state overlap ${formatSignal(cluster.stateOverlap)}, cluster density ${formatSignal(cluster.clusterDensity)}, jurisdiction fit ${formatSignal(cluster.jurisdictionFit)}.`,
|
||||
` Suggested motion: ${suggestedMotion}`,
|
||||
` Evidence: ${cluster.evidence.join("; ")}.`,
|
||||
].join("\n");
|
||||
|
|
@ -4681,115 +4682,6 @@ function renderComedyImprovDirective(identity: RepoDiscordIdentity): string {
|
|||
});
|
||||
}
|
||||
|
||||
interface RoomTopicSaturationSignal {
|
||||
messageCount: number;
|
||||
coveredMessages: number;
|
||||
terms: Array<{ term: string; count: number }>;
|
||||
}
|
||||
|
||||
interface TopicRelationToIdentity {
|
||||
isHomeAdjacent: boolean;
|
||||
matchedTerms: string[];
|
||||
}
|
||||
|
||||
function renderRepoFaceTopicSaturationFacts(
|
||||
identity: RepoDiscordIdentity,
|
||||
messages: SourceMessage[],
|
||||
): string | undefined {
|
||||
const signal = detectRoomTopicSaturation(messages);
|
||||
if (!signal) {
|
||||
return undefined;
|
||||
}
|
||||
const relation = estimateTopicRelationToIdentity(identity, signal);
|
||||
const base = `- Current-room topic saturation: ${signal.coveredMessages}/${signal.messageCount} recent messages orbit the same dominant topic. The repeated vocabulary is intentionally withheld from this projector packet so stale phrases do not become character memory.`;
|
||||
if (relation.isHomeAdjacent) {
|
||||
return [
|
||||
base,
|
||||
`- This saturation appears adjacent to ${identity.displayName}'s territory. Project sustained attention as pressure for restraint, a fresh non-repeating anchor, or a deliberate closing move; do not turn it into a phrase palette.`,
|
||||
].join("\n");
|
||||
}
|
||||
|
||||
return [
|
||||
base,
|
||||
`- This saturation does not obviously belong to ${identity.displayName}'s territory. Project it as social weather: another domain is absorbing the room, so this character may feel neglected, bored, crowded out, competitive, relieved, or tempted to pull attention back toward its own unfinished fascinations.`,
|
||||
].join("\n");
|
||||
}
|
||||
|
||||
function estimateTopicRelationToIdentity(
|
||||
identity: RepoDiscordIdentity,
|
||||
signal: RoomTopicSaturationSignal,
|
||||
): TopicRelationToIdentity {
|
||||
const identityTerms = new Set(significantTopicTerms([
|
||||
identity.id,
|
||||
identity.displayName,
|
||||
identity.repoName,
|
||||
identity.description ?? "",
|
||||
...identity.channelPermissions.flatMap((permission) => [
|
||||
permission.label ?? "",
|
||||
permission.topic ?? "",
|
||||
permission.posture ?? "",
|
||||
]),
|
||||
].join(" ")));
|
||||
const matchedTerms = signal.terms
|
||||
.map((term) => term.term)
|
||||
.filter((term) => identityTerms.has(term));
|
||||
|
||||
return {
|
||||
isHomeAdjacent: matchedTerms.length > 0,
|
||||
matchedTerms,
|
||||
};
|
||||
}
|
||||
|
||||
function detectRoomTopicSaturation(messages: SourceMessage[]): RoomTopicSaturationSignal | undefined {
|
||||
const recent = messages
|
||||
.filter((message) => collapseWhitespace(message.content).length > 0)
|
||||
.slice(-18);
|
||||
if (recent.length < 8) {
|
||||
return undefined;
|
||||
}
|
||||
|
||||
const termCounts = new Map<string, number>();
|
||||
const messageTerms = recent.map((message) => new Set(significantTopicTerms(message.content)));
|
||||
for (const terms of messageTerms) {
|
||||
for (const term of terms) {
|
||||
termCounts.set(term, (termCounts.get(term) ?? 0) + 1);
|
||||
}
|
||||
}
|
||||
|
||||
const minimumCount = Math.max(3, Math.ceil(recent.length * 0.25));
|
||||
const terms = Array.from(termCounts.entries())
|
||||
.map(([term, count]) => ({ term, count }))
|
||||
.filter((entry) => entry.count >= minimumCount)
|
||||
.sort((left, right) => {
|
||||
if (right.count !== left.count) {
|
||||
return right.count - left.count;
|
||||
}
|
||||
return left.term.localeCompare(right.term);
|
||||
})
|
||||
.slice(0, 8);
|
||||
|
||||
if (terms.length < 3) {
|
||||
return undefined;
|
||||
}
|
||||
|
||||
const repeatedTermSet = new Set(terms.slice(0, 6).map((entry) => entry.term));
|
||||
const coveredMessages = messageTerms.filter((termsForMessage) =>
|
||||
Array.from(termsForMessage).some((term) => repeatedTermSet.has(term)),
|
||||
).length;
|
||||
const topCount = terms[0]?.count ?? 0;
|
||||
const hasDominantTerm = topCount >= Math.ceil(recent.length * 0.35);
|
||||
const hasBroadCoverage = coveredMessages >= Math.ceil(recent.length * 0.68);
|
||||
if (!hasDominantTerm || !hasBroadCoverage) {
|
||||
return undefined;
|
||||
}
|
||||
|
||||
return {
|
||||
messageCount: recent.length,
|
||||
coveredMessages,
|
||||
terms,
|
||||
};
|
||||
}
|
||||
|
||||
const TOPIC_STOP_WORDS = new Set([
|
||||
"about",
|
||||
"actually",
|
||||
|
|
|
|||
Loading…
Reference in a new issue