@remnic/core 9.69.35 → 9.69.37

package/dist/{chunk-LEMD7WZJ.js → chunk-KUK56OPR.js}

@@ -3,7 +3,7 @@ import {
   headerConnector,
   renderExtractionConversation,
   resolveSourceConnector
-} from "./chunk-XPJNT6PI.js";
+} from "./chunk-2PUE5P7C.js";
 import {
   ConsolidationResultSchema,
   DaySummaryResultSchema,
@@ -13,23 +13,30 @@ import {
   ProactiveExtractionResultSchema,
   ProactiveQuestionsResultSchema,
   buildProfileConsolidationResultSchema
-} from "./chunk-AJTG3CP7.js";
+} from "./chunk-JHU4EYDV.js";
 import {
   normalizeReasoningTrace
 } from "./chunk-54V4BZWP.js";
 import {
   shouldRunProactivePass
 } from "./chunk-2CGLBUW3.js";
+import {
+  applyExtractionSpanMaterialization,
+  bindLocalExtractionPrompt,
+  buildSpanMaterializeTurns,
+  stripUntrustedFactSpans,
+  truncateLocalExtractionConversation
+} from "./chunk-3SL52JKD.js";
 import {
   attachExtractionProvenance
-} from "./chunk-MPAJFFP3.js";
+} from "./chunk-PPTOAHTZ.js";
 import {
   containsExtractionPlaceholder,
   extractionAttributes,
   extractionCueAnchors,
   extractionText,
   isPlainRecord
-} from "./chunk-R4OCGPSH.js";
+} from "./chunk-FJMYMJLR.js";
 import {
   classifyExtractionThrownError,
   classifyFallbackParseFailure
@@ -45,14 +52,15 @@ import {
 import {
   CONSOLIDATION_RESPONSE_SCHEMA,
   CUE_ANCHOR_PROMPT_INSTRUCTION,
-  EXTRACTION_RESPONSE_SHAPE,
   OUTPUT_LANGUAGE_POLICY,
   PROFILE_CONSOLIDATION_RESPONSE_SCHEMA,
   buildConsolidationSystemPrompt,
   buildExtractionInstructions,
   buildProfileConsolidationSystemPrompt,
-  eventTimePromptInstruction
-} from "./chunk-7CBKAQIG.js";
+  eventTimePromptInstruction,
+  extractionResponseShape,
+  spanModePromptSection
+} from "./chunk-BOADKFN4.js";
 import {
   AMBIENT_CAPTURE_PROMPT_SECTION_COMPACT,
   clampAmbientCaptureConfidence
@@ -204,6 +212,20 @@ function normalizeQuestion(question) {
     priority
   };
 }
+function normalizeExtractedSpanRef(value) {
+  if (!isPlainRecord(value)) return void 0;
+  const { sourceMessageIndex, charStart, charEnd, frame, sourceHash, sourceLength } = value;
+  if (typeof sourceMessageIndex !== "number" || typeof charStart !== "number" || typeof charEnd !== "number" || typeof frame !== "string") {
+    return void 0;
+  }
+  return {
+    sourceMessageIndex,
+    charStart,
+    charEnd,
+    frame,
+    ...typeof sourceHash === "string" && typeof sourceLength === "number" ? { sourceHash, sourceLength } : {}
+  };
+}
 function normalizeFactKey(fact) {
   return `${fact.category}:${fact.content.trim().toLowerCase()}`;
 }
@@ -270,6 +292,18 @@ var ExtractionEngine = class {
   get shouldUseDirectClient() {
     return !this.useGatewayModelSource && this.client !== null;
   }
+  /**
+   * Span-mode materialization (issue #2333 Phase B): sibling module, called
+   * after parse and before sanitize/grounding so downstream sees one content
+   * form. Offsets index the exact prompt turn texts and die at materialization.
+   */
+  applySpanMaterialization(result, spanTurns) {
+    return applyExtractionSpanMaterialization(
+      result,
+      spanTurns,
+      this.config.extraction?.spanMode ?? "off"
+    );
+  }
   /**
    * Build FallbackLlmOptions with the configured gateway agent ID injected.
    */
@@ -342,6 +376,7 @@ var ExtractionEngine = class {
         procedureSteps,
         reasoningTrace,
         quote: extractionText(f.quote),
+        span: normalizeExtractedSpanRef(f.span),
         eventTime: extractionText(f.eventTime) ?? extractionText(f.event_time)
       };
     }).filter((fact) => fact !== void 0 && fact.content.length > 0) : [];
@@ -529,7 +564,7 @@ var ExtractionEngine = class {
         ambient,
         signal
       );
-      const sanitizedAdditions = this.sanitizeExtractionResult(proactiveAdditions, ctx.messageTimestamp);
+      const sanitizedAdditions = this.sanitizeExtractionResult({ ...proactiveAdditions, facts: stripUntrustedFactSpans(proactiveAdditions.facts) }, ctx.messageTimestamp);
       const groundedAdditions = applyGroundingWithConnector(this.config, sanitizedAdditions, ctx);
       if (!this.hasExtractionOutputs(groundedAdditions)) return base;
       return this.mergeProactiveExtractionPass(base, groundedAdditions, maxAdditional);
@@ -891,6 +926,7 @@ var ExtractionEngine = class {
     );
     const sourceConnector = lifecycleCaps.extractionScopeClassification ? resolvedConnector : void 0;
     const { conversation, renderedConversation } = renderExtractionConversation(boundedTurns, headerConnector(sourceConnector));
+    const spanTurns = buildSpanMaterializeTurns(boundedTurns.map((turn) => turn.content));
     const groundingSource = boundedTurns.map((turn) => turn.content).join("\n\n");
     const assertionSource = boundedTurns.filter((turn) => turn.extractionContextOnly !== true).map((turn) => turn.content).join("\n\n");
     const roleAssertionSources = {
@@ -953,7 +989,16 @@ var ExtractionEngine = class {
             this.profiler.endSpan("local-llm", extractionTraceId);
             this.emit({ kind: "llm_end", traceId, model: this.config.localLlmModel, operation: "extraction", durationMs });
             log.debug(`extraction: used local LLM \u2014 ${localResult.facts.length} facts, ${localResult.entities.length} entities`);
-            const sanitized = this.sanitizeExtractionResult(localResult, messageTimestamp);
+            const localSpanTurns = bindLocalExtractionPrompt(
+              conversation,
+              boundedTurns,
+              this.modelRegistry.calculateContextSizes(
+                this.config.localLlmModel,
+                this.config.localLlmMaxContext
+              ).maxInputChars,
+              renderedConversation
+            ).spanTurns;
+            const sanitized = this.sanitizeExtractionResult(this.applySpanMaterialization(localResult, localSpanTurns), messageTimestamp);
             const grounded = applyGroundingWithConnector(this.config, sanitized, groundingContext);
             const finalResult = await this.applyProactiveQuestionPass(conversation, grounded, groundingContext, signal);
             return this.finalizeExtractionResult(finalResult, boundedTurns, resolvedConnector, ambientCapture);
@@ -1001,7 +1046,7 @@ var ExtractionEngine = class {
             this.profiler.endSpan("direct-client", extractionTraceId);
             this.emit({ kind: "llm_end", traceId, model: this.config.model, operation: "extraction", durationMs });
             log.debug(`extraction: used direct client (${this.config.model}) \u2014 ${directResult.facts.length} facts, ${directResult.entities.length} entities`);
-            const sanitized = this.sanitizeExtractionResult(directResult, messageTimestamp);
+            const sanitized = this.sanitizeExtractionResult(this.applySpanMaterialization(directResult, spanTurns), messageTimestamp);
             const grounded = applyGroundingWithConnector(this.config, sanitized, groundingContext);
             const finalResult = await this.applyProactiveQuestionPass(conversation, grounded, groundingContext, signal);
             return this.finalizeExtractionResult(finalResult, boundedTurns, resolvedConnector, ambientCapture);
@@ -1071,7 +1116,7 @@ var ExtractionEngine = class {
             content: buildExtractionInstructions(this.config, existingEntities, ambientCapture) + `
 
 Return only valid JSON matching this shape. Placeholder text describes field shape only and is never source evidence:
-${EXTRACTION_RESPONSE_SHAPE}`
+${extractionResponseShape(this.config.extraction?.spanMode ?? "off")}`
           },
           { role: "user", content: conversation }
         ];
@@ -1106,7 +1151,7 @@ ${EXTRACTION_RESPONSE_SHAPE}`
             `extracted ${result.facts.length} facts, ${result.entities.length} entities, ${(result.questions ?? []).length} questions via fallback (${detailed.modelUsed})`
           );
           const normalized = this.normalizeExtractionResultPayload(result);
-          const sanitized = this.sanitizeExtractionResult(normalized, messageTimestamp);
+          const sanitized = this.sanitizeExtractionResult(this.applySpanMaterialization(normalized, spanTurns), messageTimestamp);
           const grounded = applyGroundingWithConnector(this.config, sanitized, groundingContext);
           const finalResult = await this.applyProactiveQuestionPass(conversation, grounded, groundingContext, signal);
           return this.finalizeExtractionResult(finalResult, boundedTurns, resolvedConnector, ambientCapture);
@@ -1178,7 +1223,7 @@ ${EXTRACTION_RESPONSE_SHAPE}`
     );
     log.debug(`Model context: ${contextSizes.description}`);
     const maxConversationChars = contextSizes.maxInputChars;
-    const truncatedConversation = conversation.length > maxConversationChars ? conversation.slice(0, maxConversationChars) + "\n\n[truncated]" : conversation;
+    const truncatedConversation = truncateLocalExtractionConversation(conversation, maxConversationChars);
     const localPrompt = `You are a memory extraction system. Extract durable, reusable memories from this conversation.
 
 Use the most specific category:
@@ -1211,9 +1256,9 @@ ${CUE_ANCHOR_PROMPT_INSTRUCTION}
 - Each fact must include a quote copied verbatim from one contiguous conversation span.` : ""}${lifecycleCaps.extractionScopeClassification ? `
 - Set each fact scope to "global" for cross-project knowledge or "project" for codebase-specific knowledge. Tool, command, or CLI-flag instructions tied to one agent are "project" (the same tool name can mean different things across agents); when keeping one, begin the fact with a leading "In <agent>," clause naming that agent.` : ""}${this.config.subjectClassification?.enabled === true ? `
 - Set each fact subject to "user" when it models the person (preferences, relationships, biography, moments, commitments about their life) or "agent" when it is reusable operating knowledge (procedures, principles, tool-usage lessons, debugging strategies, environment facts with no personal content).` : ""}${ambientCapture ? AMBIENT_CAPTURE_PROMPT_SECTION_COMPACT : ""}
-${eventTimePromptInstruction(this.config)}
+${eventTimePromptInstruction(this.config)}${spanModePromptSection(this.config.extraction?.spanMode ?? "off")}
 Return only valid JSON matching this shape. Placeholder text describes field shape only and is never source evidence:
-${EXTRACTION_RESPONSE_SHAPE}
+${extractionResponseShape(this.config.extraction?.spanMode ?? "off")}
 
 Conversation:
 ${truncatedConversation}`;
@@ -1284,7 +1329,7 @@ ${truncatedConversation}`;
             content: buildExtractionInstructions(this.config, existingEntities, ambientCapture) + `
 
 Return only valid JSON matching this shape. Placeholder text describes field shape only and is never source evidence:
-${EXTRACTION_RESPONSE_SHAPE}`
+${extractionResponseShape(this.config.extraction?.spanMode ?? "off")}`
           },
           { role: "user", content: conversation }
         ],
@@ -2387,4 +2432,4 @@ export {
   shouldEnableLocalExtractionThinking,
   ExtractionEngine
 };
-//# sourceMappingURL=chunk-LEMD7WZJ.js.map
+//# sourceMappingURL=chunk-KUK56OPR.js.map