openings 0.1.52 → 0.1.54

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "openings",
-  "version": "0.1.52",
+  "version": "0.1.54",
   "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

@@ -27,6 +27,10 @@ interface CrawlerOptions {
    * whose backlog is in the thousands, which 150 a night would never finish. */
   describeLimit?: number;
   describeBudgetMs?: number;
+  /** How old a posting may be and still be worth reading, in days. 0 reads any live role, whatever its date,
+   * including one whose board states no date at all. The nightly crawl keeps the 30-day default so it spends
+   * its small budget on what is new; the catch-up pass lifts it to reach roles still listed but long posted. */
+  describeMaxAgeDays?: number;
   pacingNow?: () => number;
   pacingSleep?: (delayMs: number) => Promise<void>;
   now?: () => Date;
@@ -66,14 +70,18 @@ export function createCrawler(options: CrawlerOptions): Crawler {
   /** Required experience for every job: from its description, the last crawl's reading, or a paced detail fetch for recent roles in describeCountries. */
   async function withExperience(source: Company, jobs: Job[], previous: Job[] | undefined): Promise<Job[]> {
     const known = new Map((previous ?? []).map(normalizeJobExperience).filter((job) => job.experienceVersion === EXPERIENCE_VERSION && job.experience !== undefined).map((job) => [job.id, job]));
-    const since = now().getTime() - 30 * 86_400_000;
+    const maxAgeDays = options.describeMaxAgeDays ?? DESCRIBE_MAX_AGE_DAYS;
+    const since = maxAgeDays > 0 ? now().getTime() - maxAgeDays * 86_400_000 : 0;
+    // An undated posting is still a live posting. Inside a window it cannot be placed, so it waits; with the
+    // window lifted there is no reason to leave it unread.
+    const recentEnough = (job: Job) => (maxAgeDays <= 0 ? true : Date.parse(job.updatedAt ?? "") >= since);
     const out: Job[] = [];
     const pending: number[] = [];
     for (const job of jobs) {
       if (job.description.trim() || job.experienceVersion === EXPERIENCE_VERSION) out.push(normalizeJobExperience(job));
       else if (known.has(job.id)) out.push({ ...job, description: known.get(job.id)!.description, experience: known.get(job.id)!.experience, experienceVersion: EXPERIENCE_VERSION });
       else {
-        if (options.describe && describeRank(job) < describeOrder.length && Date.parse(job.updatedAt ?? "") >= since) pending.push(out.length);
+        if (options.describe && describeRank(job) < describeOrder.length && recentEnough(job)) pending.push(out.length);
         out.push(normalizeJobExperience(job));
       }
     }
@@ -198,6 +206,7 @@ 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_MAX_AGE_DAYS = 30;
 const DESCRIBE_BUDGET_MS = 180_000;
 
 function partitionTime(value: string | undefined): number {

package/src/local-jobs.ts

@@ -19,6 +19,7 @@ interface LocalJobsOptions {
   describeCountries?: string[];
   describeLimit?: number;
   describeBudgetMs?: number;
+  describeMaxAgeDays?: number;
   now?: () => Date;
   onCrawled?(source: Company, partition: JobPartition): void | Promise<void>;
 }

package/src/runtime.ts

@@ -40,6 +40,7 @@ export function createRuntime(options: { dataDir?: string; concurrency?: number;
     // 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"),
+    describeMaxAgeDays: wholeEnv("OPENINGS_DESCRIBE_MAX_AGE_DAYS"),
     onCrawled,
   });
   const recommender = createJobRecommender({ sources: companies, store, crawl: local.crawl });
@@ -107,3 +108,11 @@ function positiveEnv(name: string): number | undefined {
   const value = Number(raw);
   return Number.isInteger(value) && value > 0 ? value : undefined;
 }
+
+/** A whole number from the environment, zero included, or undefined when unset or nonsense. */
+function wholeEnv(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 = "0.1.52";
+export const VERSION = "0.1.54";