feat(qajobsboard): fetch job detail pages for concrete location text

The jobs.json feed often labels roles Remote/Worldwide while the public
job page JSON-LD and description include the real city (e.g. Mumbai/Nagpur).
Enrich vague rows by reading each QAJobsBoard detail URL before import.

Co-authored-by: Cursor <cursoragent@cursor.com>

extractors/qajobsboard/manifest.ts

@@ -9,6 +9,11 @@ import type {
   ExtractorRunResult,
 } from "@shared/types/extractors";
 import type { CreateJobInput } from "@shared/types/jobs";
+import {
+  extractJobLocationFromText,
+  fetchQaJobDetailEnrichment,
+  stripHtml,
+} from "./src/detail-page.js";
 
 const JOBS_URL = "https://qajobsboard.jobboardly.com/jobs.json";
 
@@ -46,21 +51,6 @@ function asString(value: unknown): string | undefined {
   return trimmed ? trimmed : undefined;
 }
 
-function decodeHtmlEntities(value: string): string {
-  return value
-    .replace(/&/g, "&")
-    .replace(/&lt;/g, "<")
-    .replace(/&gt;/g, ">")
-    .replace(/&quot;/g, '"')
-    .replace(/&#39;/g, "'")
-    .replace(/&nbsp;/g, " ");
-}
-
-function stripHtml(html: string): string {
-  const noTags = html.replace(/<[^>]+>/g, " ");
-  return decodeHtmlEntities(noTags).replace(/\s+/g, " ").trim();
-}
-
 function salaryLabel(raw: SalaryBand | undefined): string | undefined {
   if (!raw) return undefined;
   const schedule = raw.schedule ? `${raw.schedule}: ` : "";
@@ -93,19 +83,6 @@ function isVagueLocationLabel(value: string): boolean {
   return VAGUE_LOCATION_LABELS.has(value.trim().toLowerCase());
 }
 
-function extractJobLocationFromDescription(
-  html: string | undefined,
-): string | undefined {
-  if (!html) return undefined;
-  const text = stripHtml(html);
-  const match = text.match(
-    /\bjob\s+location\s*:\s*([^\n.]{2,120})/i,
-  );
-  if (!match?.[1]) return undefined;
-  const extracted = match[1].trim();
-  return extracted || undefined;
-}
-
 function locationLabel(job: QaJobBoardlyJob): string {
   const limits = Array.isArray(job.location_limits)
     ? job.location_limits
@@ -117,9 +94,9 @@ function locationLabel(job: QaJobBoardlyJob): string {
   const loc = asString(job.location);
   if (loc && !isVagueLocationLabel(loc)) return loc;
 
-  const fromDescription = extractJobLocationFromDescription(
-    job.description?.html,
-  );
+  const fromDescription = job.description?.html
+    ? extractJobLocationFromText(stripHtml(job.description.html))
+    : undefined;
   if (fromDescription) return fromDescription;
 
   if (loc) return loc;
@@ -179,6 +156,47 @@ function mapJob(raw: QaJobBoardlyJob): CreateJobInput | null {
   };
 }
 
+const DETAIL_ENRICH_CONCURRENCY = 4;
+
+function needsDetailEnrichment(location: string | undefined): boolean {
+  if (!location?.trim()) return true;
+  return isVagueLocationLabel(location);
+}
+
+async function enrichJobsFromDetailPages(
+  jobs: CreateJobInput[],
+  shouldCancel?: () => boolean,
+): Promise<CreateJobInput[]> {
+  const enriched = jobs.map((job) => ({ ...job }));
+  const targets = enriched
+    .map((job, index) => ({ job, index }))
+    .filter(({ job }) => needsDetailEnrichment(job.location));
+
+  for (let offset = 0; offset < targets.length; offset += DETAIL_ENRICH_CONCURRENCY) {
+    if (shouldCancel?.()) break;
+    const batch = targets.slice(offset, offset + DETAIL_ENRICH_CONCURRENCY);
+    await Promise.all(
+      batch.map(async ({ job, index }) => {
+        try {
+          const detail = await fetchQaJobDetailEnrichment(job.jobUrl);
+          if (!detail) return;
+          enriched[index] = {
+            ...job,
+            ...(detail.location ? { location: detail.location } : {}),
+            ...(detail.jobDescription
+              ? { jobDescription: detail.jobDescription }
+              : {}),
+          };
+        } catch {
+          // keep feed row when detail page fetch fails
+        }
+      }),
+    );
+  }
+
+  return enriched;
+}
+
 export const manifest: ExtractorManifest = {
   id: "qajobsboard",
   displayName: "QAJobsBoard",
@@ -230,16 +248,21 @@ export const manifest: ExtractorManifest = {
         out.push(mapped);
       }
 
+      const withDetails = await enrichJobsFromDetailPages(
+        out,
+        context.shouldCancel,
+      );
+
       context.onProgress?.({
         phase: "list",
         termsProcessed: 1,
         termsTotal: 1,
         currentUrl: JOBS_URL,
-        jobPagesProcessed: out.length,
-        detail: `QAJobsBoard: ${out.length} matched (${rows.length} total listings)`,
+        jobPagesProcessed: withDetails.length,
+        detail: `QAJobsBoard: ${withDetails.length} matched (${rows.length} total listings, detail pages for vague locations)`,
       });
 
-      return { success: true, jobs: out };
+      return { success: true, jobs: withDetails };
     } catch (error) {
       const message = error instanceof Error ? error.message : "Unknown error";
       return { success: false, jobs: [], error: message };