@hadoku/jobplatform-worker 2.32.37 → 2.32.39

package/dist/clients/scraper.d.ts

@@ -57,5 +57,20 @@ export declare function matchCompanies(env: AppEnv, names: string[], providers?:
  * worth revisiting if registry grows.
  */
 export declare function triggerSearch(env: AppEnv): Promise<void>;
+/** One posting read from its URL by the scraper, in the ingest-webhook shape. */
+export interface ExtractedPosting {
+    job: unknown;
+    ats: string | null;
+    board_slug: string | null;
+}
+/**
+ * Read one job posting from a URL a person pasted (scraper
+ * `POST /api/v1/jobboards/extract`). Writes nothing on either side.
+ *
+ * Throws ScraperClientError carrying the scraper's status and its own words —
+ * "LinkedIn postings can't be imported…", "That posting no longer exists" —
+ * which are written for the person and passed through to them.
+ */
+export declare function extractPosting(env: AppEnv, url: string, greenhouseSlug?: string): Promise<ExtractedPosting>;
 export {};
 //# sourceMappingURL=scraper.d.ts.map

package/dist/clients/scraper.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"scraper.d.ts","sourceRoot":"","sources":["../../src/clients/scraper.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,aAAa,CAAC;AAI1C,+EAA+E;AAC/E,UAAU,WAAW;IACpB,GAAG,EAAE,MAAM,CAAC;IACZ,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,MAAM,EAAE,MAAM,CAAC;IACf,aAAa,EAAE,MAAM,EAAE,CAAC;CACxB;AAED,MAAM,WAAW,eAAe;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,WAAW,EAAE,CAAC;CACpB;AAQD,8EAA8E;AAC9E,MAAM,WAAW,YAAY;IAC5B,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,OAAO,CAAC;IACjB,GAAG,EAAE,MAAM,GAAG,IAAI,CAAC;IACnB,IAAI,EAAE,MAAM,GAAG,IAAI,CAAC;IACpB,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,MAAM,EAAE,MAAM,CAAC;IACf,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;CACtB;AAWD,qBAAa,kBAAmB,SAAQ,KAAK;IAC5C,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;gBAEV,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM;CAMzD;AA0BD;;;;;GAKG;AACH,wBAAsB,UAAU,CAC/B,GAAG,EAAE,MAAM,EACX,KAAK,EAAE,MAAM,EAAE,EACf,SAAS,CAAC,EAAE,MAAM,EAAE,GAClB,OAAO,CAAC,eAAe,EAAE,CAAC,CAqB5B;AAED;;;;GAIG;AACH,wBAAsB,cAAc,CACnC,GAAG,EAAE,MAAM,EACX,KAAK,EAAE,MAAM,EAAE,EACf,SAAS,CAAC,EAAE,MAAM,EAAE,GAClB,OAAO,CAAC,YAAY,EAAE,CAAC,CAqBzB;AAED;;;;;;GAMG;AACH,wBAAsB,aAAa,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAY9D"}
+{"version":3,"file":"scraper.d.ts","sourceRoot":"","sources":["../../src/clients/scraper.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,aAAa,CAAC;AAI1C,+EAA+E;AAC/E,UAAU,WAAW;IACpB,GAAG,EAAE,MAAM,CAAC;IACZ,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,MAAM,EAAE,MAAM,CAAC;IACf,aAAa,EAAE,MAAM,EAAE,CAAC;CACxB;AAED,MAAM,WAAW,eAAe;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,WAAW,EAAE,CAAC;CACpB;AAQD,8EAA8E;AAC9E,MAAM,WAAW,YAAY;IAC5B,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,OAAO,CAAC;IACjB,GAAG,EAAE,MAAM,GAAG,IAAI,CAAC;IACnB,IAAI,EAAE,MAAM,GAAG,IAAI,CAAC;IACpB,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,MAAM,EAAE,MAAM,CAAC;IACf,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;CACtB;AAWD,qBAAa,kBAAmB,SAAQ,KAAK;IAC5C,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;gBAEV,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM;CAMzD;AA0BD;;;;;GAKG;AACH,wBAAsB,UAAU,CAC/B,GAAG,EAAE,MAAM,EACX,KAAK,EAAE,MAAM,EAAE,EACf,SAAS,CAAC,EAAE,MAAM,EAAE,GAClB,OAAO,CAAC,eAAe,EAAE,CAAC,CAqB5B;AAED;;;;GAIG;AACH,wBAAsB,cAAc,CACnC,GAAG,EAAE,MAAM,EACX,KAAK,EAAE,MAAM,EAAE,EACf,SAAS,CAAC,EAAE,MAAM,EAAE,GAClB,OAAO,CAAC,YAAY,EAAE,CAAC,CAqBzB;AAED;;;;;;GAMG;AACH,wBAAsB,aAAa,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAY9D;AAED,iFAAiF;AACjF,MAAM,WAAW,gBAAgB;IAChC,GAAG,EAAE,OAAO,CAAC;IACb,GAAG,EAAE,MAAM,GAAG,IAAI,CAAC;IACnB,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;CAC1B;AAcD;;;;;;;GAOG;AACH,wBAAsB,cAAc,CACnC,GAAG,EAAE,MAAM,EACX,GAAG,EAAE,MAAM,EACX,cAAc,CAAC,EAAE,MAAM,GACrB,OAAO,CAAC,gBAAgB,CAAC,CAkB3B"}

package/dist/index.js

@@ -2581,6 +2581,74 @@ function scheduleRankBuild(c, db, profileId, profile) {
     rebuildRank(db, profileId, profile).finally(() => building.delete(profileId))
   );
 }
+function normalizeWorkplaceType(wt) {
+  if (wt === "on_site") return "onsite";
+  return wt;
+}
+function insertJobStatement(db, job, { source, board_slug, now }) {
+  const workplaceType = normalizeWorkplaceType(job.workplace_type);
+  const structuredMin = job.salary?.min ?? null;
+  const structuredMax = job.salary?.max ?? null;
+  const prose = structuredMin === null && structuredMax === null ? parseSalaryRange(job.description) : null;
+  const salaryMin = structuredMin ?? prose?.min ?? null;
+  const salaryMax = structuredMax ?? prose?.max ?? null;
+  const parsed = parseAtsSlug(job.url);
+  const ats = parsed.ats ?? (board_slug ? source : null);
+  const slug2 = board_slug ?? parsed.slug;
+  const slugSource = slug2 === null ? null : board_slug ? "scraped" : "guessed";
+  const role = classifyRole(job.title, job.description);
+  return db.prepare(
+    `INSERT OR IGNORE INTO jobs (
+				id, url, source_site, title, company, location,
+				job_type, workplace_type, salary_min, salary_max,
+				description, posted_date, application_url, department,
+				scraper_used, raw, scraped_at, last_seen_at, ats, slug,
+				role_track, role_level, slug_source, geo_keys
+			) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`
+  ).bind(
+    job.id,
+    job.url,
+    job.source_site,
+    job.title,
+    job.company,
+    job.location,
+    job.job_type,
+    workplaceType,
+    salaryMin,
+    salaryMax,
+    job.description,
+    job.posted_date ?? null,
+    job.application_url ?? null,
+    job.department ?? null,
+    job.scraper_used ?? null,
+    JSON.stringify(job.raw),
+    now,
+    now,
+    ats,
+    slug2,
+    role.track,
+    role.level,
+    slugSource,
+    geoKeys(job.location)
+  );
+}
+async function rankNewJobs(db) {
+  const ranked = await db.prepare("SELECT profile_id, criteria_hash, max_job_rowid FROM job_profile_rank_state").all();
+  if (ranked.results.length === 0) return;
+  const head2 = await jobsHead(db);
+  for (const state of ranked.results) {
+    const from = Number(state.max_job_rowid);
+    if (from >= head2) continue;
+    const profile = await loadScorableProfile(db, state.profile_id);
+    if (criteriaHash(profile) !== state.criteria_hash) continue;
+    const rows = await db.prepare(
+      `SELECT id, title, location, workplace_type, salary_max, role_level
+				 FROM jobs WHERE rowid > ? AND rowid <= ?`
+    ).bind(from, head2).all();
+    await writeRanks(db, state.profile_id, profile, rows.results);
+    await markRankBuilt(db, state.profile_id, profile, head2);
+  }
+}
 const app$4 = new OpenAPIHono();
 const CULL_DAYS = 60;
 const CULL_BATCH = 500;
@@ -2614,23 +2682,6 @@ async function stripStaleDescriptions(db) {
   ).bind(cutoff, cutoff).run();
   return res.meta?.changes ?? 0;
 }
-async function rankNewJobs(db) {
-  const ranked = await db.prepare("SELECT profile_id, criteria_hash, max_job_rowid FROM job_profile_rank_state").all();
-  if (ranked.results.length === 0) return;
-  const head2 = await jobsHead(db);
-  for (const state of ranked.results) {
-    const from = Number(state.max_job_rowid);
-    if (from >= head2) continue;
-    const profile = await loadScorableProfile(db, state.profile_id);
-    if (criteriaHash(profile) !== state.criteria_hash) continue;
-    const rows = await db.prepare(
-      `SELECT id, title, location, workplace_type, salary_max, role_level
-				 FROM jobs WHERE rowid > ? AND rowid <= ?`
-    ).bind(from, head2).all();
-    await writeRanks(db, state.profile_id, profile, rows.results);
-    await markRankBuilt(db, state.profile_id, profile, head2);
-  }
-}
 app$4.use("/ingest", requireMinTier("friend"));
 app$4.use("/ingest/backfill-slugs", requireMinTier("friend"));
 app$4.use("/ingest/backfill-roles", requireMinTier("friend"));
@@ -2640,10 +2691,6 @@ app$4.use("/ingest/cull", requireMinTier("friend"));
 app$4.use("/ingest/rebuild-rank", requireMinTier("friend"));
 app$4.use("/ingest/reconcile-mail", requireMinTier("friend"));
 app$4.use("/directives", requireMinTier("friend"));
-function normalizeWorkplaceType(wt) {
-  if (wt === "on_site") return "onsite";
-  return wt;
-}
 const ingestRoute = createRoute({
   method: "post",
   path: "/ingest",
@@ -2667,53 +2714,7 @@ app$4.openapi(ingestRoute, async (c) => {
   const { jobs, batch_number, is_final, source, board_slug } = c.req.valid("json");
   const db = c.env.JOB_PLATFORM_DB;
   const now = (/* @__PURE__ */ new Date()).toISOString();
-  const stmts = jobs.map((job) => {
-    const workplaceType = normalizeWorkplaceType(job.workplace_type);
-    const structuredMin = job.salary?.min ?? null;
-    const structuredMax = job.salary?.max ?? null;
-    const prose = structuredMin === null && structuredMax === null ? parseSalaryRange(job.description) : null;
-    const salaryMin = structuredMin ?? prose?.min ?? null;
-    const salaryMax = structuredMax ?? prose?.max ?? null;
-    const parsed = parseAtsSlug(job.url);
-    const ats = parsed.ats ?? (board_slug ? source : null);
-    const slug2 = board_slug ?? parsed.slug;
-    const slugSource = slug2 === null ? null : board_slug ? "scraped" : "guessed";
-    const role = classifyRole(job.title, job.description);
-    return db.prepare(
-      `INSERT OR IGNORE INTO jobs (
-					id, url, source_site, title, company, location,
-					job_type, workplace_type, salary_min, salary_max,
-					description, posted_date, application_url, department,
-					scraper_used, raw, scraped_at, last_seen_at, ats, slug,
-					role_track, role_level, slug_source, geo_keys
-				) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`
-    ).bind(
-      job.id,
-      job.url,
-      job.source_site,
-      job.title,
-      job.company,
-      job.location,
-      job.job_type,
-      workplaceType,
-      salaryMin,
-      salaryMax,
-      job.description,
-      job.posted_date ?? null,
-      job.application_url ?? null,
-      job.department ?? null,
-      job.scraper_used ?? null,
-      JSON.stringify(job.raw),
-      now,
-      now,
-      ats,
-      slug2,
-      role.track,
-      role.level,
-      slugSource,
-      geoKeys(job.location)
-    );
-  });
+  const stmts = jobs.map((job) => insertJobStatement(db, job, { source, board_slug, now }));
   const inserted = [];
   startTime(c, "insert");
   for (let i = 0; i < stmts.length; i += INGEST_BATCH_CHUNK) {
@@ -2724,20 +2725,34 @@ app$4.openapi(ingestRoute, async (c) => {
   const accepted = inserted.filter(Boolean).length;
   const skipped = jobs.length - accepted;
   const seen = jobs.filter((_, i) => !inserted[i]);
-  const seenUrls = seen.map((j) => j.url);
-  const seenStmts = [];
-  const strippedStmts = [];
-  for (let i = 0; i < seenUrls.length; i += 90) {
-    const chunk = seenUrls.slice(i, i + 90);
+  const SEEN_CHUNK = 45;
+  const storedReads = [];
+  for (let i = 0; i < seen.length; i += SEEN_CHUNK) {
+    const chunk = seen.slice(i, i + SEEN_CHUNK);
     const inList = chunk.map(() => "?").join(",");
-    seenStmts.push(
-      db.prepare(`UPDATE jobs SET last_seen_at = ? WHERE url IN (${inList})`).bind(now, ...chunk)
+    storedReads.push(
+      db.prepare(
+        `SELECT id, url, description = '' AS blank FROM jobs
+					 WHERE url IN (${inList}) OR id IN (${inList})`
+      ).bind(...chunk.map((j) => j.url), ...chunk.map((j) => j.id))
     );
-    strippedStmts.push(
-      db.prepare(`SELECT url FROM jobs WHERE url IN (${inList}) AND description = ''`).bind(...chunk)
+  }
+  startTime(c, "seen");
+  const stored = storedReads.length ? (await db.batch(storedReads)).flatMap(
+    (r) => r.results
+  ) : [];
+  const byUrl = new Map(stored.map((r) => [r.url, r]));
+  const byId = new Map(stored.map((r) => [r.id, r]));
+  const writes = [];
+  const sameUrl = seen.filter((j) => byUrl.has(j.url)).map((j) => j.url);
+  for (let i = 0; i < sameUrl.length; i += 90) {
+    const chunk = sameUrl.slice(i, i + 90);
+    const inList = chunk.map(() => "?").join(",");
+    writes.push(
+      db.prepare(`UPDATE jobs SET last_seen_at = ? WHERE url IN (${inList})`).bind(now, ...chunk)
     );
     if (board_slug) {
-      seenStmts.push(
+      writes.push(
         db.prepare(
           `UPDATE jobs SET slug = ?, slug_source = 'scraped'
 						 WHERE url IN (${inList})
@@ -2746,21 +2761,24 @@ app$4.openapi(ingestRoute, async (c) => {
       );
     }
   }
-  startTime(c, "seen");
-  if (seenStmts.length) await db.batch(seenStmts);
-  endTime(c, "seen");
-  startTime(c, "restore");
-  if (strippedStmts.length) {
-    const found = await db.batch(strippedStmts);
-    const stripped = new Set(found.flatMap((r) => r.results.map((row) => row.url)));
-    const restores = seen.filter((j) => stripped.has(j.url) && j.description !== "").map(
-      (j) => db.prepare(`UPDATE jobs SET description = ? WHERE url = ? AND description = ''`).bind(j.description, j.url)
-    );
-    for (let i = 0; i < restores.length; i += INGEST_BATCH_CHUNK) {
-      await db.batch(restores.slice(i, i + INGEST_BATCH_CHUNK));
+  for (const j of seen) {
+    const row = byId.get(j.id);
+    if (!row || row.url === j.url) continue;
+    writes.push(db.prepare("UPDATE jobs SET last_seen_at = ? WHERE id = ?").bind(now, row.id));
+    writes.push(db.prepare("UPDATE OR IGNORE jobs SET url = ? WHERE id = ?").bind(j.url, row.id));
+  }
+  for (const j of seen) {
+    const row = byUrl.get(j.url) ?? byId.get(j.id);
+    if (row?.blank && j.description !== "") {
+      writes.push(
+        db.prepare(`UPDATE jobs SET description = ? WHERE id = ? AND description = ''`).bind(j.description, row.id)
+      );
     }
   }
-  endTime(c, "restore");
+  for (let i = 0; i < writes.length; i += INGEST_BATCH_CHUNK) {
+    await db.batch(writes.slice(i, i + INGEST_BATCH_CHUNK));
+  }
+  endTime(c, "seen");
   if (accepted > 0) {
     startTime(c, "rank");
     await rankNewJobs(db);
@@ -3401,6 +3419,34 @@ async function triggerSearch(env) {
     );
   }
 }
+function scraperMessage(body) {
+  try {
+    const parsed = JSON.parse(body);
+    if (typeof parsed.detail === "string") return parsed.detail;
+    if (typeof parsed.error?.message === "string") return parsed.error.message;
+  } catch {
+  }
+  return null;
+}
+async function extractPosting(env, url, greenhouseSlug) {
+  const response = await scraperFetch(env, "/api/v1/jobboards/extract", {
+    method: "POST",
+    body: JSON.stringify(greenhouseSlug ? { url, greenhouse_slug: greenhouseSlug } : { url })
+  });
+  const text = await response.text();
+  if (!response.ok) {
+    throw new ScraperClientError(
+      scraperMessage(text) ?? `scraper /extract returned ${response.status}`,
+      response.status,
+      text
+    );
+  }
+  const json = JSON.parse(text);
+  if (!json.success || !json.data) {
+    throw new ScraperClientError("scraper /extract reported failure", 502, text);
+  }
+  return json.data;
+}
 const app$3 = new OpenAPIHono();
 app$3.use("/profiles", requireMinTier("friend"));
 app$3.use("/profiles/*", requireMinTier("friend"));
@@ -7106,6 +7152,175 @@ function registerLocationsRoute(app2) {
     }
   );
 }
+const ImportJobSchema = z.object({
+  url: z.string().trim().url().max(2e3),
+  /** Where it lands in the pipeline when it is new to the caller. */
+  state: JobStateSchema.exclude(["dismissed"]).default("saved")
+}).openapi("ImportJob");
+const ImportedSchema = z.object({
+  success: z.literal(true),
+  data: z.object({
+    job: z.object({
+      id: z.string(),
+      title: z.string(),
+      company: z.string(),
+      location: z.string(),
+      url: z.string()
+    }),
+    /** The posting was new to the corpus (false: it was already there). */
+    created: z.boolean(),
+    /** The caller's state for it now. */
+    state: JobStateSchema,
+    /** The caller already tracked it, so its state was left alone. */
+    already_tracked: z.boolean()
+  })
+}).openapi("ImportedJob");
+function failure(status, message) {
+  return {
+    success: false,
+    error: status === 422 ? "Unprocessable" : "Bad gateway",
+    message
+  };
+}
+const JOB_COLS = "id, title, company, location, url";
+async function greenhouseSlugFor(db, url) {
+  const parsed = new URL(url);
+  if (!parsed.searchParams.has("gh_jid")) return void 0;
+  const prefix = `${parsed.protocol}//${parsed.host}/`;
+  const row = await db.prepare(
+    `SELECT slug FROM jobs
+			 WHERE url >= ? AND url < ? AND ats = 'greenhouse' AND slug IS NOT NULL
+			   AND slug_source = 'scraped'
+			 LIMIT 1`
+  ).bind(prefix, `${prefix.slice(0, -1)}0`).first();
+  return row?.slug;
+}
+function registerImportRoute(app2) {
+  app2.post("/jobs/import", gateAuthed);
+  app2.openapi(
+    createRoute({
+      method: "post",
+      path: "/jobs/import",
+      tags: ["Pipeline"],
+      summary: "Import a posting from its URL into the corpus and the caller’s pipeline",
+      request: {
+        query: ownerNameQuery,
+        body: { content: { "application/json": { schema: ImportJobSchema } } }
+      },
+      responses: {
+        200: {
+          description: "Imported (or found, and added to the pipeline)",
+          content: { "application/json": { schema: ImportedSchema } }
+        },
+        403: {
+          description: "Forbidden",
+          content: { "application/json": { schema: IdentityErrorResponseSchema } }
+        },
+        404: {
+          description: "The board no longer lists that posting, or no such owner name",
+          content: { "application/json": { schema: IdentityErrorResponseSchema } }
+        },
+        409: {
+          description: "That owner name has never signed in",
+          content: { "application/json": { schema: IdentityErrorResponseSchema } }
+        },
+        422: {
+          description: "No posting could be read from that page (or it is a LinkedIn URL)",
+          content: { "application/json": { schema: ErrorResponseSchema } }
+        },
+        502: {
+          description: "The site, or the scraper, could not be reached",
+          content: { "application/json": { schema: ErrorResponseSchema } }
+        },
+        503: {
+          description: "Identity could not be resolved right now — retry",
+          content: { "application/json": { schema: IdentityErrorResponseSchema } }
+        }
+      }
+    }),
+    async (c) => {
+      const { url, state } = c.req.valid("json");
+      const { ownerName } = c.req.valid("query");
+      const who = await effectiveUserId(c, ownerName);
+      if (isEffectiveUserError(who)) return c.json(who.error.body, who.error.status);
+      const db = c.env.JOB_PLATFORM_DB;
+      let row = await db.prepare(`SELECT ${JOB_COLS} FROM jobs WHERE url = ?`).bind(url).first();
+      let created = false;
+      if (!row) {
+        let extracted;
+        try {
+          extracted = await extractPosting(c.env, url, await greenhouseSlugFor(db, url));
+        } catch (err) {
+          if (!(err instanceof ScraperClientError)) throw err;
+          if (err.status === 404) {
+            return c.json(
+              { success: false, error: "Not found", message: err.message },
+              404
+            );
+          }
+          if (err.status === 400 || err.status === 422) {
+            return c.json(failure(422, err.message), 422);
+          }
+          return c.json(failure(502, `Could not read that posting: ${err.message}`), 502);
+        }
+        const job = IngestJobSchema.safeParse(extracted.job);
+        if (!job.success) {
+          return c.json(
+            failure(
+              502,
+              `The scraper returned a posting this API cannot store: ${job.error.message}`
+            ),
+            502
+          );
+        }
+        const now2 = (/* @__PURE__ */ new Date()).toISOString();
+        const [res] = await db.batch([
+          insertJobStatement(db, job.data, {
+            source: extracted.ats ?? "import",
+            board_slug: extracted.board_slug ?? void 0,
+            now: now2
+          })
+        ]);
+        created = (res.meta?.changes ?? 0) > 0;
+        row = await db.prepare(`SELECT ${JOB_COLS} FROM jobs WHERE url = ? OR id = ? LIMIT 1`).bind(job.data.url, job.data.id).first();
+        if (!row) throw new Error(`import: ${job.data.id} neither inserted nor found`);
+        if (created) await rankNewJobs(db);
+      }
+      const prev = await db.prepare("SELECT state FROM job_states WHERE job_id = ? AND user_id = ?").bind(row.id, who.userId).first();
+      if (prev) {
+        const kept = JobStateSchema.safeParse(prev.state);
+        return c.json(
+          {
+            success: true,
+            data: {
+              job: row,
+              created,
+              state: kept.success ? kept.data : state,
+              already_tracked: true
+            }
+          },
+          200
+        );
+      }
+      const now = (/* @__PURE__ */ new Date()).toISOString();
+      await db.batch([
+        db.prepare(
+          `INSERT INTO job_events (job_id, user_id, kind, from_state, to_state, occurred_at, created_at)
+						 SELECT ?, ?, 'state', NULL, ?, ?, ?
+						 WHERE NOT EXISTS (SELECT 1 FROM job_states WHERE job_id = ? AND user_id = ?)`
+        ).bind(row.id, who.userId, state, now, now, row.id, who.userId),
+        db.prepare(
+          `INSERT INTO job_states (job_id, user_id, state, updated_at) VALUES (?, ?, ?, ?)
+						 ON CONFLICT (job_id, user_id) DO NOTHING`
+        ).bind(row.id, who.userId, state, now)
+      ]);
+      return c.json(
+        { success: true, data: { job: row, created, state, already_tracked: false } },
+        200
+      );
+    }
+  );
+}
 const app$2 = new OpenAPIHono();
 registerFeedRoute(app$2);
 registerPreflightRoute(app$2);
@@ -7121,6 +7336,7 @@ registerFillRoutes(app$2);
 registerBlockedRoutes(app$2);
 registerInsightsRoute(app$2);
 registerLocationsRoute(app$2);
+registerImportRoute(app$2);
 registerAnswerRoutes(app$2);
 registerOwnerRoutes(app$2);
 const jobRoutes = app$2;

package/dist/ingestWrite.d.ts

@@ -0,0 +1,41 @@
+/**
+ * Writing postings into the corpus — the one path both POST /ingest (the
+ * scraper's nightly batches) and POST /jobs/import (one posting a person
+ * pasted) go through, so an imported posting is classified, salaried, placed
+ * and ranked exactly as a scraped one is.
+ */
+import type { z } from '@hono/zod-openapi';
+import type { AppEnv } from './types.js';
+import type { IngestJobSchema } from './schemas.js';
+type DB = AppEnv['JOB_PLATFORM_DB'];
+export type IngestJob = z.infer<typeof IngestJobSchema>;
+/**
+ * The INSERT OR IGNORE for one posting. Ignored when its url (UNIQUE) or id
+ * is already stored — an already-known posting keeps the fields it has.
+ *
+ * `source` is the batch label; `board_slug` the board the scraper actually
+ * fetched, when it says so.
+ */
+export declare function insertJobStatement(db: DB, job: IngestJob, { source, board_slug, now }: {
+    source: string;
+    board_slug?: string;
+    now: string;
+}): import("@cloudflare/workers-types").D1PreparedStatement;
+/**
+ * Extend every current ranking to cover the jobs inserted since it was built.
+ *
+ * Selected by rowid past each profile's marker rather than from the payload:
+ * ingest only ever INSERTs (an already-known posting is ignored, its fields
+ * untouched), so a job that was ranked once keeps the same bound and needs no
+ * new write. Re-ranking the whole payload whenever any of it was new used to
+ * upsert ~25 rows per profile per batch for one new job — 35% of the
+ * database's daily writes. The rowid range also picks up jobs a previous,
+ * failed extension missed, which a payload-only pass never would.
+ *
+ * A profile whose criteria changed since its ranking was built is left alone:
+ * extending it under the new criteria and moving its marker would bless the
+ * old rows as current. The feed sees the stale hash and rebuilds it.
+ */
+export declare function rankNewJobs(db: AppEnv['JOB_PLATFORM_DB']): Promise<void>;
+export {};
+//# sourceMappingURL=ingestWrite.d.ts.map

package/dist/ingestWrite.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ingestWrite.d.ts","sourceRoot":"","sources":["../src/ingestWrite.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AACH,OAAO,KAAK,EAAE,CAAC,EAAE,MAAM,mBAAmB,CAAC;AAC3C,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,YAAY,CAAC;AACzC,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC;AAQpD,KAAK,EAAE,GAAG,MAAM,CAAC,iBAAiB,CAAC,CAAC;AACpC,MAAM,MAAM,SAAS,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,eAAe,CAAC,CAAC;AAQxD;;;;;;GAMG;AACH,wBAAgB,kBAAkB,CACjC,EAAE,EAAE,EAAE,EACN,GAAG,EAAE,SAAS,EACd,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,EAAE,EAAE;IAAE,MAAM,EAAE,MAAM,CAAC;IAAC,UAAU,CAAC,EAAE,MAAM,CAAC;IAAC,GAAG,EAAE,MAAM,CAAA;CAAE,2DA6DjF;AAED;;;;;;;;;;;;;;GAcG;AACH,wBAAsB,WAAW,CAAC,EAAE,EAAE,MAAM,CAAC,iBAAiB,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,CAyB9E"}

package/dist/routes/ingest.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"ingest.d.ts","sourceRoot":"","sources":["../../src/routes/ingest.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAkB,MAAM,mBAAmB,CAAC;AAEhE,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,aAAa,CAAC;AAC1C,OAAO,EAAkB,KAAK,iBAAiB,EAAE,MAAM,sBAAsB,CAAC;AAwB9E,UAAU,YAAY;IACrB,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE;QAAE,WAAW,EAAE,iBAAiB,CAAA;KAAE,CAAC;CAC9C;AAkiCD,eAAO,MAAM,YAAY,oCAAM,CAAC"}
+{"version":3,"file":"ingest.d.ts","sourceRoot":"","sources":["../../src/routes/ingest.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAkB,MAAM,mBAAmB,CAAC;AAEhE,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,aAAa,CAAC;AAC1C,OAAO,EAAkB,KAAK,iBAAiB,EAAE,MAAM,sBAAsB,CAAC;AAiB9E,UAAU,YAAY;IACrB,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE;QAAE,WAAW,EAAE,iBAAiB,CAAA;KAAE,CAAC;CAC9C;AAq9BD,eAAO,MAAM,YAAY,oCAAM,CAAC"}

package/dist/routes/jobs/importUrl.d.ts

@@ -0,0 +1,3 @@
+import { type JobsApp } from './shared.js';
+export declare function registerImportRoute(app: JobsApp): void;
+//# sourceMappingURL=importUrl.d.ts.map

package/dist/routes/jobs/importUrl.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"importUrl.d.ts","sourceRoot":"","sources":["../../../src/routes/jobs/importUrl.ts"],"names":[],"mappings":"AA2BA,OAAO,EAKN,KAAK,OAAO,EACZ,MAAM,aAAa,CAAC;AA2ErB,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,OAAO,GAAG,IAAI,CAuJtD"}

package/dist/routes/jobs/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/routes/jobs/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AACH,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAuChD,eAAO,MAAM,SAAS,oCAAM,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/routes/jobs/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AACH,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAyChD,eAAO,MAAM,SAAS,oCAAM,CAAC"}

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@hadoku/jobplatform-worker",
-  "version": "2.32.37",
-  "gitHead": "808b4a95842c638f0d1840d97203dcc06383d710",
+  "version": "2.32.39",
+  "gitHead": "84b5623c402aac9159148daa3d454529dfea4407",
   "description": "Job listings aggregation and tracking dashboard - Worker package",
   "license": "MIT",
   "type": "module",