@semiont/jobs 0.5.6 → 0.5.8

package/dist/worker-main.js

@@ -1,6 +1,6 @@
+import { createTomlConfigLoader, softwareToAgent, baseUrl, getPrimaryMediaType, textExtractionOf, reconcileSelector, didToAgent, getLocaleEnglishName } from '@semiont/core';
 import { deriveStorageUri } from '@semiont/content';
 import { withSpan, SpanKind, recordJobOutcome } from '@semiont/observability';
-import { createTomlConfigLoader, softwareToAgent, baseUrl, reconcileSelector, didToAgent, getLocaleEnglishName } from '@semiont/core';
 import { generateAnnotationId } from '@semiont/event-sourcing';
 import { createInferenceClient } from '@semiont/inference';
 import { createServer } from 'http';
@@ -8,7 +8,7 @@ import { existsSync, readFileSync } from 'fs';
 import { homedir, hostname } from 'os';
 import { join } from 'path';
 import { InMemorySessionStorage, setStoredSession, kbBackendUrl, SemiontClient, SemiontSession } from '@semiont/sdk';
-import { HttpTransport, HttpContentTransport } from '@semiont/api-client';
+import { HttpTransport, HttpContentTransport } from '@semiont/http-transport';
 import { createProcessLogger } from '@semiont/observability/process-logger';
 
 var __create = Object.create;
@@ -18,7 +18,11 @@ var __getOwnPropNames = Object.getOwnPropertyNames;
 var __getProtoOf = Object.getPrototypeOf;
 var __hasOwnProp = Object.prototype.hasOwnProperty;
 var __commonJS = (cb, mod) => function __require() {
-  return mod || (0, cb[__getOwnPropNames(cb)[0]])((mod = { exports: {} }).exports, mod), mod.exports;
+  try {
+    return mod || (0, cb[__getOwnPropNames(cb)[0]])((mod = { exports: {} }).exports, mod), mod.exports;
+  } catch (e) {
+    throw mod = 0, e;
+  }
 };
 var __copyProps = (to, from, except, desc) => {
   if (from && typeof from === "object" || typeof from === "function") {
@@ -9811,21 +9815,6 @@ function logAnchorMethod(motivation, exact, anchorMethod) {
 
 // src/workers/annotation-detection.ts
 var AnnotationDetection = class {
-  /**
-   * Fetch content from a ContentFetcher and read the stream to a string.
-   * Shared helper for all workers.
-   */
-  static async fetchContent(contentFetcher, resourceId) {
-    const stream = await contentFetcher(resourceId);
-    if (!stream) {
-      throw new Error(`Could not load content for resource ${resourceId}`);
-    }
-    const chunks = [];
-    for await (const chunk of stream) {
-      chunks.push(Buffer.isBuffer(chunk) ? chunk : Buffer.from(chunk));
-    }
-    return Buffer.concat(chunks).toString("utf-8");
-  }
   /**
    * Detect comments in content.
    *
@@ -10071,10 +10060,19 @@ Knowledge graph context:
 ${parts.join("\n")}`;
     }
   }
+  let semanticContextSection = "";
+  const similar = context?.semanticContext?.similar ?? [];
+  if (similar.length > 0) {
+    const lines = [...similar].sort((a, b) => b.score - a.score).slice(0, 3).map((m) => `- (${m.score.toFixed(2)}) ${m.text.slice(0, 240)}`);
+    semanticContextSection = `
+
+Related passages from the knowledge base:
+${lines.join("\n")}`;
+  }
   const structureGuidance = finalMaxTokens >= 1e3 ? "organized into titled sections (## Section) with well-structured paragraphs" : "organized into well-structured paragraphs";
   const prompt = `Generate a concise, informative resource about "${topic}".
 ${entityTypes.length > 0 ? `Focus on these entity types: ${entityTypes.join(", ")}.` : ""}
-${userPrompt ? `Additional context: ${userPrompt}` : ""}${annotationSection}${contextSection}${graphContextSection}${sourceLanguageInstruction}${languageInstruction}
+${userPrompt ? `Additional context: ${userPrompt}` : ""}${annotationSection}${contextSection}${graphContextSection}${semanticContextSection}${sourceLanguageInstruction}${languageInstruction}
 
 Requirements:
 - Start with a clear heading (# Title)
@@ -10500,6 +10498,17 @@ async function handleJobInner(adapter, config, job) {
     adapter.failJob(jobId, `Worker not configured for job type: ${jobType}`);
     return;
   }
+  if (jobType !== "generation") {
+    const descriptor = await session.client.browse.resource(resourceId);
+    const mediaType = getPrimaryMediaType(descriptor);
+    const extraction = mediaType ? textExtractionOf(mediaType) : "none";
+    if (extraction === "pdf-text-layer") {
+      throw new Error(`Cannot run ${jobType} on resource ${resourceId}: PDF text-layer detection is not yet supported`);
+    }
+    if (extraction !== "decode") {
+      throw new Error(`Cannot run ${jobType} on resource ${resourceId}: media type '${mediaType ?? "unknown"}' has no extractable text to analyze`);
+    }
+  }
   const onProgress = (percentage, message, stage, extra) => {
     emitEvent(session, "job:report-progress", {
       ...lifecycleBase,