openings 0.1.51 → 0.1.53

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "openings",
-  "version": "0.1.51",
+  "version": "0.1.53",
   "description": "A free, candidate-safe job-search substrate for AI agents",
   "license": "MIT",
   "repository": { "type": "git", "url": "git+https://github.com/abhay-avagama/hiring-agent.git" },

package/src/catalog.ts

@@ -35,6 +35,9 @@ interface LeverJob {
   hostedUrl: string;
   categories?: { location?: string };
   descriptionPlain?: string;
+  /** The requirements and responsibilities, which Lever keeps apart from the opening paragraph. */
+  lists?: Array<{ text?: string; content?: string }>;
+  additionalPlain?: string;
   workplaceType?: string;
   createdAt?: number;
 }
@@ -458,6 +461,19 @@ function normalizeAshby(company: Company, job: AshbyJob): Job {
   });
 }
 
+/**
+ * Lever splits an advert across three fields and a board may fill any of them. Reading only the opening
+ * paragraph left more than half of Lever's roles with no description at all: one aggregator alone publishes
+ * 4,076 postings whose `descriptionPlain` is empty and whose requirements sit entirely in `lists`. Those
+ * requirements are also where the skills a candidate searches for are written, so this is the part that counts.
+ */
+function leverDescription(job: LeverJob): string {
+  const sections = [job.descriptionPlain ?? ""];
+  for (const list of job.lists ?? []) sections.push([list.text ?? "", stripHtml(list.content ?? "")].filter(Boolean).join("\n"));
+  sections.push(job.additionalPlain ?? "");
+  return sections.map((section) => section.trim()).filter(Boolean).join("\n\n");
+}
+
 function normalizeLever(company: Company, job: LeverJob): Job {
   const location = job.categories?.location ?? "Unspecified";
   return classifyJob({
@@ -473,7 +489,7 @@ function normalizeLever(company: Company, job: LeverJob): Job {
     eligibilityConfidence: "unknown",
     url: job.hostedUrl,
     updatedAt: job.createdAt ? new Date(job.createdAt).toISOString() : undefined,
-    description: job.descriptionPlain ?? "",
+    description: leverDescription(job),
   });
 }
 

package/src/crawler.ts

@@ -22,6 +22,11 @@ interface CrawlerOptions {
   describe?(source: Company, job: Job): Promise<string>;
   /** Only roles open to these countries, posted in the last 30 days, get a description fetch. Empty: none do. */
   describeCountries?: string[];
+  /** How many descriptions one source may be asked for in a run, and how long that may take. The nightly crawl
+   * keeps both small so no employer holds up the pass; the description catch-up raises them to reach an employer
+   * whose backlog is in the thousands, which 150 a night would never finish. */
+  describeLimit?: number;
+  describeBudgetMs?: number;
   pacingNow?: () => number;
   pacingSleep?: (delayMs: number) => Promise<void>;
   now?: () => Date;
@@ -76,8 +81,8 @@ export function createCrawler(options: CrawlerOptions): Crawler {
     // listed its jobs. Without this, an employer with 4 India roles among 11,581 American ones never reads the four.
     // Ties keep the board's own order, which is newest first on every provider that dates its postings.
     pending.sort((left, right) => describeRank(jobs[left]!) - describeRank(jobs[right]!) || left - right);
-    const deadline = Date.now() + DESCRIBE_BUDGET_MS;
-    let budget = DESCRIBE_LIMIT;
+    const deadline = Date.now() + (options.describeBudgetMs ?? DESCRIBE_BUDGET_MS);
+    let budget = options.describeLimit ?? DESCRIBE_LIMIT;
     for (const index of pending) {
       if (budget <= 0 || Date.now() >= deadline) break;
       budget -= 1;
@@ -191,6 +196,7 @@ export function createCrawler(options: CrawlerOptions): Crawler {
 }
 
 // ponytail: per-source caps keep one crawl bounded; a big backlog (first run) drains over a few nights.
+/** Defaults for the nightly crawl, where the point is to keep every source moving rather than finish any one. */
 const DESCRIBE_LIMIT = 150;
 const DESCRIBE_BUDGET_MS = 180_000;
 

package/src/local-jobs.ts

@@ -17,6 +17,8 @@ interface LocalJobsOptions {
   workdayCountries?: string[];
   describe?(source: Company, job: Job): Promise<string>;
   describeCountries?: string[];
+  describeLimit?: number;
+  describeBudgetMs?: number;
   now?: () => Date;
   onCrawled?(source: Company, partition: JobPartition): void | Promise<void>;
 }

package/src/runtime.ts

@@ -37,6 +37,9 @@ export function createRuntime(options: { dataDir?: string; concurrency?: number;
     // Server crawls set OPENINGS_DESCRIBE_COUNTRIES (IN) to read required experience from Workday detail pages; installs skip the extra requests.
     describe: (source, job) => fetchJobDescription(source, job, globalThis.fetch),
     describeCountries: (process.env.OPENINGS_DESCRIBE_COUNTRIES ?? "").split(",").map((code) => code.trim().toUpperCase()).filter((code) => /^[A-Z]{2}$/.test(code)),
+    // The description catch-up pass raises these; unset, the nightly crawl keeps the crawler's own small defaults.
+    describeLimit: positiveEnv("OPENINGS_DESCRIBE_LIMIT"),
+    describeBudgetMs: positiveEnv("OPENINGS_DESCRIBE_BUDGET_MS"),
     onCrawled,
   });
   const recommender = createJobRecommender({ sources: companies, store, crawl: local.crawl });
@@ -96,3 +99,11 @@ export function createHostedRuntime(options: { sources: Company[]; store: Snapsh
     prepareJobSearch: preparation.prepare, getJobCoverage: coverage.getCoverage, recommend: recommender.recommend, analyzeJobFit: analyzer.analyze, optimizeResume: optimizer.optimize,
   };
 }
+
+/** A positive whole number from the environment, or undefined when unset or nonsense. */
+function positiveEnv(name: string): number | undefined {
+  const raw = process.env[name];
+  if (raw === undefined) return undefined;
+  const value = Number(raw);
+  return Number.isInteger(value) && value > 0 ? value : undefined;
+}

package/src/version.ts

@@ -1 +1 @@
-export const VERSION = "0.1.51";
+export const VERSION = "0.1.53";