feat(context,kernel): compression pipeline stage 3 — replay-safe tier summarizer

ContextSummarizedEvent (registered) keyed on stable session sequence, not the ephemeral pack ordinal, so replay reuses the recorded summary artifact and never re-calls the LLM (invariants #6/#8). TierContextSummarizer mirrors JournalCompactionService for freeform context (tool outputs/docs/retrieved text) the journal path doesn't cover. ContextState + DefaultContextReducer fold the event; reducer replay-safety unit-tested.
2026-07-01 14:03:47 +04:00
parent b50688df1d
commit e0c222392c
6 changed files with 151 additions and 1 deletions
@@ -0,0 +1,64 @@
+package com.correx.core.kernel.orchestration
+
+import com.correx.core.artifactstore.ArtifactStore
+import com.correx.core.context.model.ContextEntry
+import com.correx.core.events.events.ContextSummarizedEvent
+import com.correx.core.events.events.EventPayload
+import com.correx.core.events.types.SessionId
+
+/**
+ * Stage 5 (RECURSIVE_SUMMARIZE) for freeform context — the tier the decision-journal
+ * compaction path does NOT cover: long tool outputs, docs and retrieved text. Mirrors
+ * [JournalCompactionService]: summarize via injected inference, store the summary in the
+ * artifact store, emit [ContextSummarizedEvent] referencing it. Replay reads the recorded
+ * artifact and never re-calls the LLM (invariants #6/#8).
+ */
+class TierContextSummarizer(
+    private val artifactStore: ArtifactStore,
+    private val summarize: suspend (String) -> String,
+    private val tokenThreshold: () -> Int = { DEFAULT_TOKEN_THRESHOLD },
+) {
+    companion object {
+        private const val DEFAULT_TOKEN_THRESHOLD = 2000
+        private const val CHARS_PER_TOKEN = 4
+    }
+
+    /**
+     * Summarize [freeformEntries] (tier 2/3 candidates) when their combined size crosses the
+     * threshold. Returns true when a summary event was emitted. [tier] selects light summary
+     * (2) vs summary-of-summaries (3); protected spans are already preserved in the fact sheet,
+     * so the prompt only asks to keep intent and load-bearing facts.
+     */
+    suspend fun summarizeIfNeeded(
+        sessionId: SessionId,
+        freeformEntries: List<ContextEntry>,
+        coversThroughSequence: Long,
+        tier: Int,
+        emit: suspend (EventPayload) -> Unit,
+    ): Boolean {
+        if (freeformEntries.isEmpty()) return false
+        val tokenEstimate = freeformEntries.sumOf { it.content.length } / CHARS_PER_TOKEN
+        if (tokenEstimate < tokenThreshold()) return false
+
+        val prompt = buildString {
+            appendLine("Summarize the following context concisely (≤200 words).")
+            appendLine("Preserve user intent, decisions, constraints and every load-bearing fact.")
+            appendLine()
+            freeformEntries.forEach { appendLine("- ${it.content}") }
+        }
+        val summaryText = summarize(prompt)
+        val summaryArtifactId = artifactStore.put(summaryText.toByteArray(Charsets.UTF_8))
+        artifactStore.flushBefore {
+            emit(
+                ContextSummarizedEvent(
+                    sessionId = sessionId,
+                    coversThroughSequence = coversThroughSequence,
+                    summaryArtifactId = summaryArtifactId,
+                    tier = tier,
+                    entriesOmittedCount = freeformEntries.size,
+                ),
+            )
+        }
+        return true
+    }
+}