fold: autoCaptureContextTurns rolling pair window onto current tips (PR 966, composed with master's reconcile-with-kept-texts and protected-prefix contracts)

Author: gorkem2020

dist/index.js

@@ -41,7 +41,7 @@ import { buildFallbackCandidate, gateRegexFallbackCapture } from "./src/autocapt
 import { gateMappedReflectionEntries, resolveMappedRowAdmissionController } from "./src/reflection-mapped-admission.js";
 import { createMemoryCLI } from "./cli.js";
 import { isNoise } from "./src/noise-filter.js";
-import { buildConversationTurnsForExtraction, formatConversationTranscript, neutralizeSpeakerTagSpoof, nextAutoCaptureMessageId, normalizeAutoCaptureText, reconcileTurnsWithKeptTexts, } from "./src/auto-capture-cleanup.js";
+import { buildConversationTurnsForExtraction, dedupePairWindow, formatConversationTranscript, neutralizeSpeakerTagSpoof, nextAutoCaptureMessageId, normalizeAutoCaptureText, reconcileTurnsWithKeptTexts, trimTurnsToUserCap, } from "./src/auto-capture-cleanup.js";
 // Import smart extraction & lifecycle components
 import { SmartExtractor, createExtractionRateLimiter, stripEnvelopeMetadata } from "./src/smart-extractor.js";
 import { compressTexts, estimateConversationValue } from "./src/session-compressor.js";
@@ -2069,6 +2069,8 @@ function _initPluginState(api) {
     const autoCaptureRecentTurns = new Map();
     const autoCaptureDeferredFlushTurns = new Map();
     const autoCaptureSessionIdToKey = new Map();
+    const autoCaptureSessionIds = new Map();
+    const autoCaptureRecentPairTurns = new Map();
     const autoCaptureInFlightRuns = new Map();
     return {
         config,
@@ -2100,6 +2102,8 @@ function _initPluginState(api) {
         autoCaptureRecentTurns,
         autoCaptureDeferredFlushTurns,
         autoCaptureSessionIdToKey,
+        autoCaptureSessionIds,
+        autoCaptureRecentPairTurns,
         autoCaptureInFlightRuns,
         captureAdmissionController,
         captureAdmissionAudit,
@@ -2213,7 +2217,7 @@ const memoryLanceDBProPlugin = {
             _registeredApisMap.delete(api); // dual-track rollback: Map un-claim
             throw err;
         }
-        const { config, resolvedDbPath, vectorDim, store, embedder, retriever, canonicalCorpusIndexer, dreamingEngine, dreamingScheduler, scopeManager, migrator, smartExtractor, mdMirror, decayEngine, tierManager, extractionRateLimiter, reflectionErrorStateBySession, reflectionDerivedBySession, reflectionDerivedSuppressionBySession, reflectionByAgentCache, reflectionByAgentCacheGeneration, recallHistory, turnCounter, autoCaptureSeenTextCount, autoCapturePendingIngressTexts, autoCaptureCountedPendingCount, autoCaptureRecentTurns, autoCaptureDeferredFlushTurns, autoCaptureSessionIdToKey, autoCaptureInFlightRuns, captureAdmissionController, captureAdmissionAudit, captureReflectionAdmissionController, admissionRejectionAuditWriter, } = singleton;
+        const { config, resolvedDbPath, vectorDim, store, embedder, retriever, canonicalCorpusIndexer, dreamingEngine, dreamingScheduler, scopeManager, migrator, smartExtractor, mdMirror, decayEngine, tierManager, extractionRateLimiter, reflectionErrorStateBySession, reflectionDerivedBySession, reflectionDerivedSuppressionBySession, reflectionByAgentCache, reflectionByAgentCacheGeneration, recallHistory, turnCounter, autoCaptureSeenTextCount, autoCapturePendingIngressTexts, autoCaptureCountedPendingCount, autoCaptureRecentTurns, autoCaptureDeferredFlushTurns, autoCaptureSessionIdToKey, autoCaptureSessionIds, autoCaptureRecentPairTurns, autoCaptureInFlightRuns, captureAdmissionController, captureAdmissionAudit, captureReflectionAdmissionController, admissionRejectionAuditWriter, } = singleton;
         const learnAutoCaptureSessionAlias = (sessionId, sessionKey) => {
             if (typeof sessionId !== "string" || !sessionId
                 || typeof sessionKey !== "string" || !sessionKey
@@ -3590,7 +3594,36 @@ const memoryLanceDBProPlugin = {
                                 // texts the selectors dropped back into extraction. Kept indices
                                 // pin each surviving copy to its own turn; occurrence counting
                                 // stays as the fallback when positional alignment is unavailable.
-                                const finalConversationTurns = reconcileTurnsWithKeptTexts(thisCallTurns, cleanTexts, cleanTurnIndices);
+                                let finalConversationTurns = reconcileTurnsWithKeptTexts(thisCallTurns, cleanTexts, cleanTurnIndices);
+                                // Rolling PAIR window sized by autoCaptureContextTurns (0 =
+                                // disabled: each extraction sees only its own call's turns, and
+                                // nothing is retained between calls). When enabled, this call's
+                                // reconciled pairs extend what earlier calls buffered, bounded
+                                // to autoCaptureContextTurns user turns (or this call's own
+                                // new-user count when larger, so unextracted user turns are
+                                // never trimmed out of their own transcript). The buffer holds
+                                // the FILTERED window, so selector-dropped texts can never
+                                // re-enter a later transcript as retained context. A remember
+                                // flow (prepended referent) bypasses the prepend for its own
+                                // call: the extractor's protected-prefix contract counts
+                                // referent turns from position zero.
+                                const contextTurns = config.autoCaptureContextTurns ?? 0;
+                                if (contextTurns > 0 && rememberPrependedTurns.length === 0) {
+                                    const priorPairTurns = autoCaptureRecentPairTurns.get(sessionKey) || [];
+                                    finalConversationTurns = trimTurnsToUserCap(dedupePairWindow([...priorPairTurns, ...finalConversationTurns]), Math.max(contextTurns, finalConversationTurns.filter((turn) => turn.role === "user").length));
+                                }
+                                if (contextTurns === 0) {
+                                    autoCaptureRecentPairTurns.delete(sessionKey);
+                                }
+                                else if (thisCallTurns.length > 0) {
+                                    // Deliberately retained across successful extractions:
+                                    // deleting it here would mean steady-state captures (one
+                                    // extraction per turn) always see a bare current pair. The
+                                    // set-time trim bounds it; the watermark keeps retained
+                                    // turns from re-becoming sources.
+                                    autoCaptureRecentPairTurns.set(sessionKey, finalConversationTurns);
+                                    pruneMapIfOver(autoCaptureRecentPairTurns, AUTO_CAPTURE_MAP_MAX_ENTRIES);
+                                }
                                 // The referent is the OLDEST turn of the prepended window, which is
                                 // exactly what the extractor's newest-first budget walk sacrifices
                                 // first, so it needs a guaranteed share. Only the referent RUN gets
@@ -5655,6 +5688,7 @@ export function parsePluginConfig(value) {
             })()
             : undefined,
         extractMinMessages: parsePositiveInt(cfg.extractMinMessages) ?? 4,
+        autoCaptureContextTurns: Math.min(10, Math.max(0, Math.floor(Number(cfg.autoCaptureContextTurns)) || 0)),
         extractMaxChars: parsePositiveInt(cfg.extractMaxChars) ?? 8000,
         batchChunkSize: (() => { const raw = parsePositiveInt(cfg.batchChunkSize); return raw === undefined ? undefined : Math.min(50, raw); })(),
         scopes: typeof cfg.scopes === "object" && cfg.scopes !== null ? cfg.scopes : undefined,

dist/src/auto-capture-cleanup.js

@@ -357,6 +357,103 @@ function keepRenderedTail(blocks, rendered, start, end, budget) {
     }
     return kept;
 }
+/**
+ * Bounds a rolling pair window to at most `maxUserTurns` user turns, keeping
+ * the newest ones with their interleaved assistant replies, and never leaving
+ * an orphan assistant turn ahead of the window's first user turn. The caller
+ * passes max(autoCaptureContextTurns, this call's new user turns), so the
+ * transcript always contains every not-yet-extracted user turn, padded with
+ * earlier still-buffered pairs up to the configured window.
+ */
+export function trimTurnsToUserCap(turns, maxUserTurns) {
+    const cap = Math.max(1, maxUserTurns);
+    let userCount = 0;
+    let start = turns.length;
+    for (let i = turns.length - 1; i >= 0; i--) {
+        if (turns[i].role === "user") {
+            userCount++;
+            if (userCount > cap)
+                break;
+            start = i;
+        }
+    }
+    if (userCount === 0) {
+        // All-assistant window (possible under captureAssistant=true when the
+        // delta carries only assistant turns): no user anchor exists, so keep
+        // the newest `cap` turns instead of silently dropping everything.
+        return turns.slice(-cap);
+    }
+    return turns.slice(start);
+}
+/**
+ * Repairs a pair window that double-preserved deferred turns. A below-threshold
+ * deferral keeps content alive on two independent paths -- the rolling pair
+ * buffer, and the watermark rollback (or pending-ingress re-queue) whose next
+ * slice re-includes the same turns -- so the assembled window can carry the
+ * same exchange twice. Collapse duplicates by user text at pair granularity:
+ * a pair-shaped copy (user turn plus its replies) beats a flat re-queued copy,
+ * copies of an identical exchange collapse to the latest, and a repeated user
+ * text whose replies differ is a real conversation and is kept whole.
+ */
+export function dedupePairWindow(turns) {
+    const groups = [];
+    let current = null;
+    for (const turn of turns) {
+        if (turn.role === "user") {
+            current = { turns: [turn], userText: turn.text, replies: "" };
+            groups.push(current);
+        }
+        else if (current) {
+            current.turns.push(turn);
+            current.replies = JSON.stringify(current.turns.slice(1).map((t) => t.text));
+        }
+        else {
+            groups.push({ turns: [turn], userText: null, replies: "" });
+        }
+    }
+    const kept = [];
+    for (const group of groups) {
+        if (group.userText === null) {
+            kept.push(group);
+            continue;
+        }
+        let prevIndex = -1;
+        for (let i = kept.length - 1; i >= 0; i--) {
+            if (kept[i].userText === group.userText) {
+                prevIndex = i;
+                break;
+            }
+        }
+        if (prevIndex < 0) {
+            kept.push(group);
+            continue;
+        }
+        const prev = kept[prevIndex];
+        const prevPaired = prev.turns.length > 1;
+        const currPaired = group.turns.length > 1;
+        if (currPaired && prevPaired) {
+            if (prev.replies === group.replies) {
+                kept.splice(prevIndex, 1);
+                kept.push(group);
+            }
+            else {
+                kept.push(group);
+            }
+        }
+        else if (currPaired && !prevPaired) {
+            kept.splice(prevIndex, 1);
+            kept.push(group);
+        }
+        else if (!currPaired && prevPaired) {
+            continue;
+        }
+        else {
+            kept.splice(prevIndex, 1);
+            kept.push(group);
+        }
+    }
+    return kept.flatMap((group) => group.turns);
+}
 /**
  * Assembles the ordered turn sequence for the extraction prompt's transcript
  * from this call's true message-loop order, without recomputing any

index.ts

@@ -74,11 +74,13 @@ import { isNoise } from "./src/noise-filter.js";
 import {
   type ConversationTurn,
   buildConversationTurnsForExtraction,
+  dedupePairWindow,
   formatConversationTranscript,
   neutralizeSpeakerTagSpoof,
   nextAutoCaptureMessageId,
   normalizeAutoCaptureText,
   reconcileTurnsWithKeptTexts,
+  trimTurnsToUserCap,
 } from "./src/auto-capture-cleanup.js";
 
 // Import smart extraction & lifecycle components
@@ -277,6 +279,8 @@ interface PluginConfig {
     thinkLevel?: string;
   };
   extractMinMessages?: number;
+  /** Rolling extraction context window in retained user turns (0 = disabled, max 10). */
+  autoCaptureContextTurns?: number;
   extractMaxChars?: number;
   batchChunkSize?: number;
   scopes?: {
@@ -2489,6 +2493,9 @@ interface PluginSingletonState {
   autoCaptureRecentTurns: Map<string, ConversationTurn[]>;
   autoCaptureDeferredFlushTurns: Map<string, ConversationTurn[]>;
   autoCaptureSessionIdToKey: Map<string, string>;
+
+  autoCaptureSessionIds: Map<string, string>;
+  autoCaptureRecentPairTurns: Map<string, ConversationTurn[]>;
   autoCaptureInFlightRuns: Map<string, Set<Promise<void>>>;
   captureAdmissionController: () => AdmissionController | null;
   captureAdmissionAudit: () => boolean;
@@ -2797,6 +2804,9 @@ function _initPluginState(api: OpenClawPluginApi): PluginSingletonState {
   const autoCaptureRecentTurns = new Map<string, ConversationTurn[]>();
   const autoCaptureDeferredFlushTurns = new Map<string, ConversationTurn[]>();
   const autoCaptureSessionIdToKey = new Map<string, string>();
+
+  const autoCaptureSessionIds = new Map<string, string>();
+  const autoCaptureRecentPairTurns = new Map<string, ConversationTurn[]>();
   const autoCaptureInFlightRuns = new Map<string, Set<Promise<void>>>();
 
   return {
@@ -2829,6 +2839,9 @@ function _initPluginState(api: OpenClawPluginApi): PluginSingletonState {
     autoCaptureRecentTurns,
     autoCaptureDeferredFlushTurns,
     autoCaptureSessionIdToKey,
+
+    autoCaptureSessionIds,
+    autoCaptureRecentPairTurns,
     autoCaptureInFlightRuns,
     captureAdmissionController,
     captureAdmissionAudit,
@@ -2988,7 +3001,9 @@ const memoryLanceDBProPlugin = {
       autoCaptureCountedPendingCount,
       autoCaptureRecentTurns,
       autoCaptureDeferredFlushTurns,
-    autoCaptureSessionIdToKey,
+      autoCaptureSessionIdToKey,
+      autoCaptureSessionIds,
+      autoCaptureRecentPairTurns,
       autoCaptureInFlightRuns,
       captureAdmissionController,
       captureAdmissionAudit,
@@ -4651,7 +4666,38 @@ const memoryLanceDBProPlugin = {
               // texts the selectors dropped back into extraction. Kept indices
               // pin each surviving copy to its own turn; occurrence counting
               // stays as the fallback when positional alignment is unavailable.
-              const finalConversationTurns = reconcileTurnsWithKeptTexts(thisCallTurns, cleanTexts, cleanTurnIndices);
+              let finalConversationTurns = reconcileTurnsWithKeptTexts(thisCallTurns, cleanTexts, cleanTurnIndices);
+              // Rolling PAIR window sized by autoCaptureContextTurns (0 =
+              // disabled: each extraction sees only its own call's turns, and
+              // nothing is retained between calls). When enabled, this call's
+              // reconciled pairs extend what earlier calls buffered, bounded
+              // to autoCaptureContextTurns user turns (or this call's own
+              // new-user count when larger, so unextracted user turns are
+              // never trimmed out of their own transcript). The buffer holds
+              // the FILTERED window, so selector-dropped texts can never
+              // re-enter a later transcript as retained context. A remember
+              // flow (prepended referent) bypasses the prepend for its own
+              // call: the extractor's protected-prefix contract counts
+              // referent turns from position zero.
+              const contextTurns = config.autoCaptureContextTurns ?? 0;
+              if (contextTurns > 0 && rememberPrependedTurns.length === 0) {
+                const priorPairTurns = autoCaptureRecentPairTurns.get(sessionKey) || [];
+                finalConversationTurns = trimTurnsToUserCap(
+                  dedupePairWindow([...priorPairTurns, ...finalConversationTurns]),
+                  Math.max(contextTurns, finalConversationTurns.filter((turn) => turn.role === "user").length),
+                );
+              }
+              if (contextTurns === 0) {
+                autoCaptureRecentPairTurns.delete(sessionKey);
+              } else if (thisCallTurns.length > 0) {
+                // Deliberately retained across successful extractions:
+                // deleting it here would mean steady-state captures (one
+                // extraction per turn) always see a bare current pair. The
+                // set-time trim bounds it; the watermark keeps retained
+                // turns from re-becoming sources.
+                autoCaptureRecentPairTurns.set(sessionKey, finalConversationTurns);
+                pruneMapIfOver(autoCaptureRecentPairTurns, AUTO_CAPTURE_MAP_MAX_ENTRIES);
+              }
               // The referent is the OLDEST turn of the prepended window, which is
               // exactly what the extractor's newest-first budget walk sacrifices
               // first, so it needs a guaranteed share. Only the referent RUN gets
@@ -7065,6 +7111,7 @@ export function parsePluginConfig(value: unknown): PluginConfig {
       })()
       : undefined,
     extractMinMessages: parsePositiveInt(cfg.extractMinMessages) ?? 4,
+    autoCaptureContextTurns: Math.min(10, Math.max(0, Math.floor(Number(cfg.autoCaptureContextTurns)) || 0)),
     extractMaxChars: parsePositiveInt(cfg.extractMaxChars) ?? 8000,
     batchChunkSize: (() => { const raw = parsePositiveInt(cfg.batchChunkSize); return raw === undefined ? undefined : Math.min(50, raw); })(),
     scopes: typeof cfg.scopes === "object" && cfg.scopes !== null ? cfg.scopes as any : undefined,

openclaw.plugin.json

@@ -438,6 +438,13 @@
         "default": 4,
         "description": "Minimum conversation messages required before smart extraction runs."
       },
+      "autoCaptureContextTurns": {
+        "type": "integer",
+        "minimum": 0,
+        "maximum": 10,
+        "default": 0,
+        "description": "Rolling context window for auto-capture extraction: how many recent user turns (with their assistant replies) stay in the transcript across extractions as context. 0 disables retention; each extraction then sees only its own call's turns."
+      },
       "batchChunkSize": {
         "type": "integer",
         "minimum": 1,
@@ -1939,6 +1946,11 @@
       "help": "Minimum conversation messages before smart extraction triggers",
       "advanced": true
     },
+    "autoCaptureContextTurns": {
+      "label": "Auto-Capture Context Turns",
+      "help": "Rolling window of recent user turns (with replies) kept as extraction context; 0 = off",
+      "advanced": true
+    },
     "extractMaxChars": {
       "label": "Max Chars for Extraction",
       "help": "Maximum conversation characters to process for extraction",