openclaw-cortex-memory 0.1.0-Alpha.9 → 0.1.1

package/dist/src/store/archive_store.js

@@ -36,22 +36,197 @@ Object.defineProperty(exports, "__esModule", { value: true });
 exports.createArchiveStore = createArchiveStore;
 const fs = __importStar(require("fs"));
 const path = __importStar(require("path"));
+const http_post_1 = require("../net/http_post");
 const ontology_1 = require("../graph/ontology");
+const llm_output_validator_1 = require("../quality/llm_output_validator");
 function normalizeBaseUrl(value) {
     if (!value)
         return "";
     return value.endsWith("/") ? value.slice(0, -1) : value;
 }
-function scoreQuality(text) {
-    const length = text.length;
-    const score = Math.max(0, Math.min(1, Number((Math.min(length, 320) / 320).toFixed(2))));
-    if (score >= 0.75) {
-        return { score, level: "high" };
+function resolveArchiveSourceCharLimit(value) {
+    if (typeof value === "number" && Number.isFinite(value)) {
+        return Math.max(1000, Math.floor(value));
     }
-    if (score >= 0.4) {
-        return { score, level: "medium" };
+    return 500000;
+}
+function clampTailText(text, maxChars) {
+    const source = (text || "").trim();
+    if (!source)
+        return "";
+    if (!Number.isFinite(maxChars) || maxChars <= 0 || source.length <= maxChars) {
+        return source;
+    }
+    return source.slice(-Math.floor(maxChars)).trim();
+}
+const ARCHIVE_LOW_INFORMATION_LINE = /^(ok|okay|got it|roger|noted|sure|thanks|thank you|received|copy that|understood|好的|收到|明白|了解|谢谢|感谢|可以|行|嗯|嗯嗯|没问题)(?:\b|$)/i;
+function denoiseArchiveSourceText(text) {
+    const raw = (text || "").trim();
+    if (!raw)
+        return "";
+    const output = [];
+    const seen = new Set();
+    for (const line of raw.split(/\r?\n/)) {
+        const trimmed = line.trim();
+        if (!trimmed)
+            continue;
+        const content = trimmed.replace(/^\[[^\]]+\]\s*/, "").trim();
+        if (!content)
+            continue;
+        const hasSignal = /(https?:\/\/|www\.|[A-Za-z0-9._-]+\.[A-Za-z]{2,}|[`#/:\\]|@\w+|\b\d{2,}\b)/.test(content);
+        if (!hasSignal && ARCHIVE_LOW_INFORMATION_LINE.test(content)) {
+            continue;
+        }
+        const dedupKey = content.toLowerCase();
+        if (!hasSignal && seen.has(dedupKey)) {
+            continue;
+        }
+        seen.add(dedupKey);
+        output.push(trimmed);
+    }
+    return output.length > 0 ? output.join("\n") : raw;
+}
+function normalizeOneLineText(value) {
+    return String(value || "").replace(/\s+/g, " ").trim();
+}
+const ARCHIVE_TASK_INSTRUCTION_PATTERNS = [
+    /请|需要|帮我|麻烦|任务|需求|实现|修复|排查|优化|上线|部署|整理|编写|启用|查看|检查|分析|导入|生成|提高|改进/,
+    /please|can you|need to|task|implement|fix|investigate|optimi[sz]e|deploy|enable|review/i,
+];
+const ARCHIVE_COMPLETION_REPORT_PATTERNS = [
+    /已完成|完成了|处理完|搞定|已修复|修复了|已实现|已上线|已部署|结果|汇报|完成情况|报告|已通过|验证通过|测试通过/,
+    /done|completed|fixed|implemented|deployed|resolved|report|summary|finished/i,
+];
+const ARCHIVE_USER_ACCEPTANCE_PATTERNS = [
+    /确认|认可|通过|验收|OK|可以|好的|收到|辛苦|谢谢|没问题|就这样|接受|效果可以/,
+    /approved|accepted|looks good|great|works|thank you|confirmed/i,
+];
+const ARCHIVE_ACTION_PATTERNS = [
+    /决定|完成|修复|发布|上线|部署|提交|交付|验证|关闭|推进|落地|实施|启用|导入|生成|优化|改进/,
+    /decide|complete|fix|release|deploy|ship|deliver|verify|close|implement|enable|migrate/i,
+];
+const ARCHIVE_FAILURE_PATTERNS = [
+    /失败|报错|错误|异常|阻塞|卡住|不行|超时|回滚|故障|不通过|无法/,
+    /failed|error|exception|blocked|timeout|rollback|incident/i,
+];
+const ARCHIVE_SUCCESS_PATTERNS = [
+    /成功|完成|修复|解决|通过|已上线|稳定|正常|恢复|可用|生效/,
+    /success|completed|fixed|resolved|passed|stable|recovered|works/i,
+];
+function matchesAnyPattern(text, patterns) {
+    return patterns.some(pattern => pattern.test(text));
+}
+function firstMatchIndex(text, patterns) {
+    let minIndex = -1;
+    for (const pattern of patterns) {
+        const idx = text.search(pattern);
+        if (idx < 0) {
+            continue;
+        }
+        if (minIndex < 0 || idx < minIndex) {
+            minIndex = idx;
+        }
+    }
+    return minIndex;
+}
+function scoreQuality(args) {
+    const summary = (args.summary || "").trim();
+    const cause = (args.cause || "").trim();
+    const process = (args.process || "").trim();
+    const result = (args.result || "").trim();
+    const outcome = (args.outcome || "").trim();
+    const sourceText = (args.sourceText || "").trim();
+    const mergedText = [summary, cause, process, result, outcome, sourceText].filter(Boolean).join("\n");
+    const hasStructuredTriplet = cause.length > 0 && process.length > 0 && result.length > 0;
+    const hasTaskInstruction = matchesAnyPattern(mergedText, ARCHIVE_TASK_INSTRUCTION_PATTERNS);
+    const hasCompletionReport = matchesAnyPattern(mergedText, ARCHIVE_COMPLETION_REPORT_PATTERNS);
+    const hasUserAcceptance = matchesAnyPattern(mergedText, ARCHIVE_USER_ACCEPTANCE_PATTERNS);
+    const hasAction = matchesAnyPattern(mergedText, ARCHIVE_ACTION_PATTERNS);
+    const hasFailure = matchesAnyPattern(mergedText, ARCHIVE_FAILURE_PATTERNS);
+    const hasSuccess = matchesAnyPattern(mergedText, ARCHIVE_SUCCESS_PATTERNS);
+    const hasOutcome = outcome.length >= 6 || hasSuccess;
+    const firstFailureIdx = hasFailure ? firstMatchIndex(mergedText, ARCHIVE_FAILURE_PATTERNS) : -1;
+    const firstSuccessIdx = hasSuccess ? firstMatchIndex(mergedText, ARCHIVE_SUCCESS_PATTERNS) : -1;
+    const failThenSuccess = hasFailure && hasSuccess && firstFailureIdx >= 0 && firstSuccessIdx > firstFailureIdx;
+    const workflowComplete = hasStructuredTriplet || (hasTaskInstruction && hasCompletionReport && hasUserAcceptance);
+    let score = 0;
+    if (summary.length >= 24)
+        score += 0.1;
+    if (summary.length >= 60)
+        score += 0.1;
+    if (summary.length >= 120)
+        score += 0.06;
+    if (summary.length >= 180)
+        score += 0.04;
+    if (hasStructuredTriplet)
+        score += 0.22;
+    if (hasAction)
+        score += 0.14;
+    if (hasOutcome)
+        score += 0.12;
+    if (hasTaskInstruction)
+        score += 0.12;
+    if (hasCompletionReport)
+        score += 0.12;
+    if (hasUserAcceptance)
+        score += 0.14;
+    if (workflowComplete)
+        score += 0.12;
+    if (failThenSuccess)
+        score += 0.1;
+    const normalizedScore = Math.max(0, Math.min(1, Number(score.toFixed(2))));
+    if (normalizedScore >= 0.75) {
+        return {
+            score: normalizedScore,
+            level: "high",
+            signals: { hasStructuredTriplet, hasTaskInstruction, hasCompletionReport, hasUserAcceptance, workflowComplete, failThenSuccess },
+        };
     }
-    return { score, level: "low" };
+    if (normalizedScore >= 0.4) {
+        return {
+            score: normalizedScore,
+            level: "medium",
+            signals: { hasStructuredTriplet, hasTaskInstruction, hasCompletionReport, hasUserAcceptance, workflowComplete, failThenSuccess },
+        };
+    }
+    return {
+        score: normalizedScore,
+        level: "low",
+        signals: { hasStructuredTriplet, hasTaskInstruction, hasCompletionReport, hasUserAcceptance, workflowComplete, failThenSuccess },
+    };
+}
+function normalizeArchiveDedupText(value) {
+    return String(value || "")
+        .toLowerCase()
+        .replace(/[^\p{L}\p{N}:./\\#@_-]+/gu, " ")
+        .replace(/\s+/g, " ")
+        .trim();
+}
+function buildArchiveDedupText(args) {
+    const relationText = Array.isArray(args.relations)
+        ? args.relations
+            .map(relation => `${relation.source || ""}|${relation.type || ""}|${relation.target || ""}`)
+            .filter(item => item.replace(/\|/g, "").trim().length > 0)
+            .sort()
+            .join("\n")
+        : "";
+    const entityText = Array.isArray(args.entities)
+        ? args.entities.map(item => String(item || "").trim()).filter(Boolean).sort().join("\n")
+        : "";
+    return [
+        `canonical:${args.canonicalId || ""}`,
+        `event_type:${args.eventType}`,
+        `summary:${args.summary}`,
+        `cause:${args.cause || ""}`,
+        `process:${args.process || ""}`,
+        `result:${args.result || args.outcome || ""}`,
+        entityText ? `entities:${entityText}` : "",
+        relationText ? `relations:${relationText}` : "",
+    ]
+        .filter(Boolean)
+        .map(normalizeArchiveDedupText)
+        .filter(Boolean)
+        .join("\n");
 }
 async function requestEmbedding(args) {
     const endpoint = args.baseUrl.endsWith("/embeddings") ? args.baseUrl : `${args.baseUrl}/embeddings`;
@@ -70,24 +245,18 @@ async function requestEmbedding(args) {
         : 4;
     let lastError = null;
     for (let attempt = 0; attempt < maxRetries; attempt += 1) {
-        const controller = new AbortController();
-        const timeoutId = setTimeout(() => controller.abort(), timeoutMs);
+        const response = await (0, http_post_1.postJsonWithTimeout)({
+            endpoint,
+            apiKey: args.apiKey,
+            body,
+            timeoutMs,
+        });
+        if (!response.ok) {
+            lastError = new Error(response.status > 0 ? `embedding_http_${response.status}` : (response.error || "embedding_network_error"));
+            continue;
+        }
         try {
-            const response = await fetch(endpoint, {
-                method: "POST",
-                headers: {
-                    "content-type": "application/json",
-                    authorization: `Bearer ${args.apiKey}`,
-                },
-                body: JSON.stringify(body),
-                signal: controller.signal,
-            });
-            clearTimeout(timeoutId);
-            if (!response.ok) {
-                lastError = new Error(`embedding_http_${response.status}`);
-                continue;
-            }
-            const json = await response.json();
+            const json = (response.json || {});
             const embedding = json?.data?.[0]?.embedding;
             if (Array.isArray(embedding) && embedding.length > 0) {
                 return embedding.filter(item => Number.isFinite(item));
@@ -95,7 +264,6 @@ async function requestEmbedding(args) {
             lastError = new Error("embedding_empty");
         }
         catch (error) {
-            clearTimeout(timeoutId);
             lastError = error;
         }
         if (attempt < maxRetries - 1) {
@@ -131,34 +299,6 @@ function inferGateSource(event) {
     }
     return "manual";
 }
-function buildArchiveVectorText(event, normalizedEventType, entities) {
-    const lines = [
-        `event_type: ${normalizedEventType}`,
-        `summary: ${(event.summary || "").trim()}`,
-        `outcome: ${(event.outcome || "").trim()}`,
-        `entities: ${entities.join(", ")}`,
-        `source_file: ${(event.source_file || "").trim()}`,
-        `actor: ${(event.actor || "").trim()}`,
-    ].map(line => line.trim()).filter(Boolean);
-    if (Array.isArray(event.relations) && event.relations.length > 0) {
-        const relationLines = event.relations
-            .map(relation => {
-            if (!relation || typeof relation !== "object")
-                return "";
-            const source = typeof relation.source === "string" ? relation.source.trim() : "";
-            const target = typeof relation.target === "string" ? relation.target.trim() : "";
-            const type = typeof relation.type === "string" ? relation.type.trim() : "related_to";
-            if (!source || !target)
-                return "";
-            return `${source} -[${type}]-> ${target}`;
-        })
-            .filter(Boolean);
-        if (relationLines.length > 0) {
-            lines.push(`relations: ${relationLines.join(" ; ")}`);
-        }
-    }
-    return lines.join("\n").trim();
-}
 function splitTextChunks(text, chunkSize, chunkOverlap) {
     const normalizedSize = Number.isFinite(chunkSize) && chunkSize >= 200 ? Math.floor(chunkSize) : 600;
     const normalizedOverlap = Number.isFinite(chunkOverlap) && chunkOverlap >= 0
@@ -210,6 +350,32 @@ function splitTextChunks(text, chunkSize, chunkOverlap) {
     }
     return output;
 }
+function pickEvidenceChunks(chunks, maxCount) {
+    if (!chunks.length || maxCount <= 0)
+        return [];
+    if (chunks.length <= maxCount)
+        return chunks;
+    const picked = new Map();
+    picked.set(chunks[0].index, chunks[0]);
+    if (maxCount >= 2) {
+        const mid = chunks[Math.floor(chunks.length / 2)];
+        picked.set(mid.index, mid);
+    }
+    if (maxCount >= 3) {
+        const last = chunks[chunks.length - 1];
+        picked.set(last.index, last);
+    }
+    if (picked.size < maxCount) {
+        for (const chunk of chunks) {
+            if (!picked.has(chunk.index)) {
+                picked.set(chunk.index, chunk);
+            }
+            if (picked.size >= maxCount)
+                break;
+        }
+    }
+    return [...picked.values()].sort((a, b) => a.index - b.index).slice(0, maxCount);
+}
 async function mapWithConcurrency(items, maxConcurrency, mapper) {
     if (items.length === 0) {
         return [];
@@ -232,8 +398,9 @@ async function mapWithConcurrency(items, maxConcurrency, mapper) {
 }
 function createArchiveStore(options) {
     const archivePath = path.join(options.memoryRoot, "sessions", "archive", "sessions.jsonl");
-    const mutationLogPath = path.join(options.memoryRoot, "graph", "mutation_log.jsonl");
+    const mutationLogPath = path.join(options.memoryRoot, "sessions", "archive", "mutation_log.jsonl");
     const graphSchema = (0, ontology_1.loadGraphSchema)(options.projectRoot);
+    const archiveSourceTextMaxChars = resolveArchiveSourceCharLimit(options.writePolicy?.archiveSourceTextMaxChars);
     async function storeEvents(events) {
         const stored = [];
         const skipped = [];
@@ -243,46 +410,112 @@ function createArchiveStore(options) {
         const lines = [];
         const mutationLines = [];
         for (const event of events) {
-            const summary = (event.summary || "").trim();
-            if (!summary) {
+            const rawSummary = normalizeOneLineText(event.summary || "");
+            if (!rawSummary) {
                 skipped.push({ summary: "", reason: "empty_summary" });
                 options.logger.info("archive_skip reason=empty_summary");
                 continue;
             }
+            const cause = normalizeOneLineText(event.cause || "");
+            const process = normalizeOneLineText(event.process || "");
+            const result = normalizeOneLineText(event.result || event.outcome || "");
+            const summary = rawSummary;
             const confidence = typeof event.confidence === "number"
                 ? Math.max(0, Math.min(1, event.confidence))
                 : undefined;
-            if (typeof confidence === "number" && confidence < 0.35) {
-                skipped.push({ summary, reason: "low_confidence" });
-                options.logger.info("archive_skip reason=filtered_low_quality detail=low_confidence");
+            const quality = scoreQuality({
+                summary,
+                cause,
+                process,
+                result,
+                outcome: event.outcome,
+                sourceText: event.source_text,
+            });
+            const gateSource = inferGateSource(event);
+            const lifecycleComplete = quality.signals.workflowComplete;
+            if (gateSource === "sync" && !quality.signals.hasStructuredTriplet) {
+                skipped.push({ summary, reason: "incomplete_cause_process_result" });
+                options.logger.info("archive_skip reason=incomplete_cause_process_result gate_source=sync");
                 continue;
             }
-            const quality = scoreQuality(summary);
-            if (quality.level === "low") {
-                skipped.push({ summary, reason: "low_quality" });
-                options.logger.info("archive_skip reason=filtered_low_quality detail=low_quality");
-                continue;
+            const archiveMinConfidence = typeof options.writePolicy?.archiveMinConfidence === "number"
+                ? Math.max(0, Math.min(1, options.writePolicy.archiveMinConfidence))
+                : 0.35;
+            if (typeof confidence === "number" && confidence < archiveMinConfidence) {
+                if (!lifecycleComplete) {
+                    skipped.push({ summary, reason: "low_confidence" });
+                    options.logger.info("archive_skip reason=filtered_low_quality detail=low_confidence");
+                    continue;
+                }
+                options.logger.info(`archive_confidence_override reason=workflow_complete confidence=${confidence.toFixed(2)} threshold=${archiveMinConfidence.toFixed(2)}`);
+            }
+            const archiveMinQualityScore = typeof options.writePolicy?.archiveMinQualityScore === "number"
+                ? Math.max(0, Math.min(1, options.writePolicy.archiveMinQualityScore))
+                : 0.4;
+            if (quality.score < archiveMinQualityScore) {
+                if (!lifecycleComplete) {
+                    skipped.push({ summary, reason: "low_quality" });
+                    options.logger.info("archive_skip reason=filtered_low_quality detail=low_quality");
+                    continue;
+                }
+                options.logger.info(`archive_quality_override reason=workflow_complete quality=${quality.score.toFixed(2)} threshold=${archiveMinQualityScore.toFixed(2)}`);
             }
             const normalizedEventType = (0, ontology_1.normalizeEventType)(event.event_type || "insight", graphSchema);
-            const entities = Array.isArray(event.entities)
-                ? [...new Set(event.entities.map(item => (typeof item === "string" ? item.trim() : "")).filter(Boolean))]
-                : [];
-            const relationValidation = (0, ontology_1.validateRelations)({
-                relations: Array.isArray(event.relations) ? event.relations : [],
-                entities,
-                entityTypes: event.entity_types,
-                schema: graphSchema,
+            const id = `evt_${Date.now().toString(36)}_${Math.random().toString(36).slice(2, 8)}`;
+            const sourceTextRaw = typeof event.source_text === "string" ? event.source_text : "";
+            const sourceText = clampTailText(denoiseArchiveSourceText(sourceTextRaw), archiveSourceTextMaxChars);
+            const canonicalId = event.canonical_id || (0, ontology_1.buildCanonicalId)({
+                eventType: normalizedEventType,
+                summary,
+                outcome: event.outcome,
+            });
+            const dedupText = buildArchiveDedupText({
+                eventType: normalizedEventType,
+                summary,
+                cause,
+                process,
+                result,
+                outcome: event.outcome,
+                canonicalId,
+                entities: event.entities,
+                relations: event.relations,
+            });
+            const dedup = options.deduplicator.check({
+                id,
+                summary: dedupText || `${normalizedEventType}: ${summary}`,
             });
-            if (relationValidation.accepted.length === 0 && Array.isArray(event.relations) && event.relations.length > 0) {
-                skipped.push({ summary, reason: "relation_validation_failed" });
-                options.logger.info("archive_skip reason=relation_validation_failed");
+            if (dedup.duplicate) {
+                skipped.push({ summary, reason: `duplicate_${dedup.stage || "unknown"}` });
+                options.logger.info(`archive_skip reason=duplicate_dedup_stage_${dedup.stage || "unknown"}`);
                 continue;
             }
-            const id = `evt_${Date.now().toString(36)}_${Math.random().toString(36).slice(2, 8)}`;
+            const record = {
+                id,
+                timestamp: new Date().toISOString(),
+                layer: "archive",
+                event_type: normalizedEventType,
+                summary,
+                cause,
+                process,
+                result,
+                source_text: sourceText || undefined,
+                outcome: event.outcome,
+                session_id: event.session_id,
+                source_file: event.source_file,
+                gate_source: gateSource,
+                embedding_status: "pending",
+                quality_score: quality.score,
+                quality_level: quality.level,
+                char_count: (sourceText || summary).length,
+                token_count: estimateTokenCount(sourceText || summary),
+                vector_chunks_total: 0,
+                vector_chunks_ok: 0,
+                confidence,
+                source_event_id: event.source_event_id,
+                actor: event.actor || "system",
+                canonical_id: canonicalId,
+            };
             let embedding = undefined;
-            let embeddingStatus = "pending";
-            let vectorChunksTotal = 0;
-            let vectorChunksOk = 0;
             const vectorUpsertRows = [];
             const embeddingModel = options.embedding?.model || "";
             const embeddingApiKey = options.embedding?.apiKey || "";
@@ -291,10 +524,38 @@ function createArchiveStore(options) {
             if (embeddingModel && embeddingApiKey && embeddingBaseUrl) {
                 const chunkSize = options.vectorChunking?.chunkSize ?? 600;
                 const chunkOverlap = options.vectorChunking?.chunkOverlap ?? 100;
-                const vectorText = buildArchiveVectorText(event, normalizedEventType, entities);
-                const chunks = splitTextChunks(vectorText, chunkSize, chunkOverlap);
-                vectorChunksTotal = chunks.length;
-                const chunkEmbeddings = await mapWithConcurrency(chunks, maxParallel, async (chunk) => {
+                const evidenceMaxChunks = typeof options.vectorChunking?.evidenceMaxChunks === "number"
+                    ? Math.max(0, Math.min(8, Math.floor(options.vectorChunking.evidenceMaxChunks)))
+                    : 2;
+                const summaryText = (record.summary || "").trim();
+                const evidenceChunks = record.source_text
+                    ? pickEvidenceChunks(splitTextChunks(record.source_text, chunkSize, chunkOverlap), evidenceMaxChunks)
+                    : [];
+                const summaryChunk = summaryText
+                    ? [
+                        {
+                            text: summaryText,
+                            source_field: "summary",
+                            index: 0,
+                            total: 1 + evidenceChunks.length,
+                            start: 0,
+                            end: summaryText.length,
+                        },
+                    ]
+                    : [];
+                const embeddingInputs = [
+                    ...summaryChunk,
+                    ...evidenceChunks.map((chunk, idx) => ({
+                        text: chunk.text,
+                        source_field: "evidence",
+                        index: idx + summaryChunk.length,
+                        total: summaryChunk.length + evidenceChunks.length,
+                        start: chunk.start,
+                        end: chunk.end,
+                    })),
+                ];
+                record.vector_chunks_total = embeddingInputs.length;
+                const chunkEmbeddings = await mapWithConcurrency(embeddingInputs, maxParallel, async (chunk) => {
                     try {
                         const chunkEmbedding = await requestEmbedding({
                             text: chunk.text,
@@ -314,14 +575,18 @@ function createArchiveStore(options) {
                         return null;
                     }
                     catch (error) {
-                        options.logger.warn(`Archive chunk embedding failed id=${id} chunk=${chunk.index} error=${error}`);
+                        options.logger.warn(`Archive chunk embedding failed id=${id} chunk=${chunk.index} field=${chunk.source_field} error=${error}`);
                         return null;
                     }
                 });
                 const validEmbeddings = chunkEmbeddings
                     .filter((item) => Boolean(item))
                     .sort((a, b) => a.chunk.index - b.chunk.index);
-                if (validEmbeddings.length > 0) {
+                const primary = validEmbeddings.find(item => item.chunk.source_field === "summary");
+                if (primary) {
+                    embedding = primary.embedding;
+                }
+                else if (validEmbeddings.length > 0) {
                     embedding = validEmbeddings[0].embedding;
                 }
                 for (const item of validEmbeddings) {
@@ -329,58 +594,24 @@ function createArchiveStore(options) {
                         id: `${id}_c${item.chunk.index}`,
                         summary: item.chunk.text,
                         embedding: item.embedding,
+                        source_field: item.chunk.source_field,
                         chunk_index: item.chunk.index,
-                        chunk_total: chunks.length,
+                        chunk_total: item.chunk.total,
                         chunk_start: item.chunk.start,
                         chunk_end: item.chunk.end,
                     });
                 }
-                vectorChunksOk = validEmbeddings.length;
-                embeddingStatus = vectorChunksTotal > 0 && vectorChunksOk === vectorChunksTotal ? "ok" : "failed";
+                record.vector_chunks_ok = validEmbeddings.length;
+                record.embedding_status = record.vector_chunks_total > 0 && record.vector_chunks_ok === record.vector_chunks_total
+                    ? "ok"
+                    : "failed";
             }
-            const dedup = options.deduplicator.check({
-                id,
-                summary: `${normalizedEventType}: ${summary}`,
-                embedding,
-            });
-            if (dedup.duplicate) {
-                skipped.push({ summary, reason: `duplicate_${dedup.stage || "unknown"}` });
-                options.logger.info(`archive_skip reason=duplicate_dedup_stage_${dedup.stage || "unknown"}`);
-                continue;
-            }
-            const gateSource = inferGateSource(event);
-            const record = {
-                ...event,
-                event_type: normalizedEventType,
-                entities,
-                relations: relationValidation.accepted,
-                canonical_id: event.canonical_id || (0, ontology_1.buildCanonicalId)({
-                    eventType: normalizedEventType,
-                    summary,
-                    entities,
-                    relations: relationValidation.accepted,
-                    outcome: event.outcome,
-                }),
-                actor: event.actor || "system",
-                confidence,
-                id,
-                timestamp: new Date().toISOString(),
-                layer: "archive",
-                gate_source: gateSource,
-                embedding_status: embeddingStatus,
-                quality_score: quality.score,
-                quality_level: quality.level,
-                char_count: buildArchiveVectorText(event, normalizedEventType, entities).length,
-                token_count: estimateTokenCount(buildArchiveVectorText(event, normalizedEventType, entities)),
-                vector_chunks_total: vectorChunksTotal,
-                vector_chunks_ok: vectorChunksOk,
-                embedding,
-            };
+            record.embedding = embedding;
             lines.push(JSON.stringify(record));
             stored.push(record);
             options.deduplicator.append({
                 id: record.id,
-                summary: `${record.event_type}: ${summary}`,
+                summary: dedupText || `${record.event_type}: ${summary}`,
                 embedding: embedding,
             });
             mutationLines.push(JSON.stringify({
@@ -405,13 +636,13 @@ function createArchiveStore(options) {
                             summary: chunkRow.summary,
                             timestamp: record.timestamp,
                             outcome: record.outcome,
-                            entities: record.entities,
-                            relations: record.relations,
                             embedding: chunkRow.embedding,
                             quality_score: record.quality_score,
                             layer: "archive",
                             source_memory_id: record.id,
                             source_memory_canonical_id: record.canonical_id,
+                            source_event_id: record.source_event_id || record.id,
+                            source_field: chunkRow.source_field,
                             char_count: chunkRow.summary.length,
                             token_count: estimateTokenCount(chunkRow.summary),
                             chunk_index: chunkRow.chunk_index,
@@ -435,6 +666,12 @@ function createArchiveStore(options) {
         if (lines.length > 0) {
             ensureDirForFile(archivePath);
             fs.appendFileSync(archivePath, `${lines.join("\n")}\n`, "utf-8");
+            for (let i = 0; i < lines.length; i++) {
+                const validation = (0, llm_output_validator_1.validateJsonlLine)(lines[i]);
+                if (!validation.valid && validation.errors.length > 0) {
+                    options.logger.warn(`archive_write_integrity_check_failed line=${i} errors=${validation.errors.join("|")}`);
+                }
+            }
             ensureDirForFile(mutationLogPath);
             fs.appendFileSync(mutationLogPath, `${mutationLines.join("\n")}\n`, "utf-8");
         }