openings 0.1.10 → 0.1.12

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "openings",
-  "version": "0.1.10",
+  "version": "0.1.12",
   "description": "Find evidence-grounded jobs, including relevant roles you may not have searched for, without accounts or API keys.",
   "author": { "name": "Openings contributors" },
   "license": "MIT",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "openings",
-  "version": "0.1.10",
+  "version": "0.1.12",
   "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

@@ -116,6 +116,8 @@ export function searchJobs(jobs: Job[], query: SearchQuery): JobSummary[] {
 export interface FetchJobsObserver {
   onBackoff?(event: { status: number; delayMs: number }): void;
   workdayPageDelayMs?: number;
+  /** Countries to fetch with Workday's country facet after a capped crawl, so roles beyond the 2,000-posting cap are not lost. */
+  workdayCountries?: string[];
   pacingNow?: () => number;
   pacingSleep?: (delayMs: number, signal?: AbortSignal) => Promise<void>;
 }
@@ -147,6 +149,11 @@ export async function fetchSourceJobs(company: Company, fetcher: Fetch = globalT
         : (body as { offers: RecruiteeJob[] }).offers.map((job) => normalizeRecruitee(company, job));
 }
 
+const WORKDAY_LISTING_CAP = 2000;
+/** Workday's country facet ids are the same on every tenant. Verified live on 2026-09-07. */
+const WORKDAY_COUNTRY_FACETS: Record<string, string> = { IN: "c4f78be1a8f14da0ab49ce1162348a5e", US: "bc33aa3152ec42d4995f4791a106ed09" };
+function workdayJobKey(job: WorkdayJob): string { return job.bulletFields?.[0] ?? job.externalPath; }
+
 async function fetchWorkdayJobs(company: Company, fetcher: Fetch, signal?: AbortSignal, observer?: FetchJobsObserver): Promise<Job[]> {
   const source = parseWorkdayToken(company.token);
   const endpoint = `https://${source.host}/wday/cxs/${encodeURIComponent(source.tenant)}/${encodeURIComponent(source.site)}/jobs`;
@@ -168,10 +175,10 @@ async function fetchWorkdayJobs(company: Company, fetcher: Fetch, signal?: Abort
       previousPageStart = pacingNow();
     } finally { release(); }
   }
-  async function page(offset: number): Promise<{ total: number; jobs: WorkdayJob[] }> {
+  async function page(offset: number, appliedFacets: Record<string, string[]> = {}): Promise<{ total: number; jobs: WorkdayJob[] }> {
     for (let attempt = 0; attempt < 3; attempt += 1) {
       await pacePageStart();
-      const response = await fetcher(endpoint, { method: "POST", headers: { "content-type": "application/json" }, body: JSON.stringify({ appliedFacets: {}, limit, offset, searchText: "" }), signal });
+      const response = await fetcher(endpoint, { method: "POST", headers: { "content-type": "application/json" }, body: JSON.stringify({ appliedFacets, limit, offset, searchText: "" }), signal });
       if (response.ok) {
         const body = await response.json() as { total?: unknown; jobPostings?: unknown };
         if (!Number.isInteger(body.total) || !Array.isArray(body.jobPostings)) throw new Error(`${company.name} Workday source returned an invalid payload`);
@@ -203,6 +210,18 @@ async function fetchWorkdayJobs(company: Company, fetcher: Fetch, signal?: Abort
   await Promise.all(Array.from({ length: workerCount }, worker));
   const jobs = [first.jobs, ...pages].flat().slice(0, first.total);
   if (jobs.length !== first.total) throw new Error(`${company.name} Workday source returned ${jobs.length} of ${first.total} jobs`);
+  // Workday stops an unfiltered listing at 2,000 postings. For capped tenants, a per-country pass sees past the cap.
+  if (first.total >= WORKDAY_LISTING_CAP) {
+    const seen = new Set(jobs.map((job) => workdayJobKey(job)));
+    for (const code of observer?.workdayCountries ?? []) {
+      const facet = WORKDAY_COUNTRY_FACETS[code.toUpperCase()];
+      if (!facet) continue;
+      const head = await page(0, { locationCountry: [facet] });
+      const extra = [head.jobs];
+      for (let offset = limit; offset < Math.min(head.total, WORKDAY_LISTING_CAP); offset += limit) extra.push((await page(offset, { locationCountry: [facet] })).jobs);
+      for (const job of extra.flat()) { const key = workdayJobKey(job); if (!seen.has(key)) { seen.add(key); jobs.push(job); } }
+    }
+  }
   return jobs.map((job) => normalizeWorkday(company, source, job));
 }
 

package/src/cli.ts

@@ -22,7 +22,7 @@ import { mergeAttemptedRoundLeads, prepareRecruiteeRoundArtifacts } from "./recr
 const HELP = `Openings — search public company job boards
 
 Usage:
-  openings crawl [--country CODE | --companies FILE] [--concurrency N] [--source-cache-hours N] [--source-limit N] [--delay-ms N] [--workday-page-delay-ms N] [--data-dir PATH]
+  openings crawl [--workday-countries IN,US] [--country CODE | --companies FILE] [--concurrency N] [--source-cache-hours N] [--source-limit N] [--delay-ms N] [--workday-page-delay-ms N] [--data-dir PATH]
   openings snapshot export [--input FILE] [--output-dir PATH]
   openings coverage report --country CODE [--snapshot FILE] [--catalog FILE] [--candidates FILE] [--registry FILE] [--output FILE] [--as-of ISO]
   openings sources discover-subdomains SEEDS.json [--output FILE] [--limit N] [--delay-ms N] [--report FILE]
@@ -69,7 +69,7 @@ export async function run(args: string[]): Promise<number> {
     if (typeof parsed === "string") return fail(parsed);
     const runtime = createRuntime({
       dataDir: parsed.dataDir, concurrency: parsed.concurrency, sourceCacheHours: parsed.sourceCacheHours,
-      sourceLimit: parsed.sourceLimit, crawlDelayMs: parsed.delayMs, workdayPageDelayMs: parsed.workdayPageDelayMs,
+      sourceLimit: parsed.sourceLimit, crawlDelayMs: parsed.delayMs, workdayPageDelayMs: parsed.workdayPageDelayMs, workdayCountries: parsed.workdayCountries,
     });
     const slugs = parsed.companiesFile ? await readCompanyFile(parsed.companiesFile) : undefined;
     console.log(JSON.stringify(await runtime.crawl({ country: parsed.country, slugs }), null, 2));
@@ -514,6 +514,7 @@ function parseCrawl(args: string[]) {
   let concurrency = 10;
   let delayMs = 0;
   let workdayPageDelayMs = 0;
+  let workdayCountries: string[] | undefined;
   let sourceCacheHours = 24;
   let sourceLimit = 0;
   for (let index = 0; index < args.length; index += 1) {
@@ -526,6 +527,10 @@ function parseCrawl(args: string[]) {
     else if (arg === "--concurrency") {
       concurrency = Number(args[++index]);
       if (!Number.isInteger(concurrency) || concurrency < 1 || concurrency > 100) return "--concurrency must be an integer from 1 to 100";
+    } else if (arg === "--workday-countries") {
+      const value = args[++index] ?? "";
+      workdayCountries = value.split(",").map((code) => code.trim().toUpperCase()).filter(Boolean);
+      if (!workdayCountries.every((code) => /^[A-Z]{2}$/.test(code))) return "--workday-countries must be a comma-separated list of two-letter codes";
     } else if (arg === "--delay-ms") {
       delayMs = Number(args[++index]);
       if (!Number.isInteger(delayMs) || delayMs < 0 || delayMs > 60_000) return "--delay-ms must be an integer from 0 to 60000";
@@ -543,7 +548,7 @@ function parseCrawl(args: string[]) {
   if (country && companiesFile) return "Use either --country or --companies, not both";
   if (args.includes("--companies") && !companiesFile) return "--companies requires a file";
   if (args.includes("--data-dir") && !dataDir) return "--data-dir requires a value";
-  return { country, companiesFile, dataDir, concurrency, delayMs, workdayPageDelayMs, sourceCacheHours, sourceLimit };
+  return { country, companiesFile, dataDir, concurrency, delayMs, workdayPageDelayMs, sourceCacheHours, sourceLimit , workdayCountries };
 }
 
 function parseCountry(value: string | undefined): string | undefined {

package/src/crawler.ts

@@ -16,6 +16,7 @@ interface CrawlerOptions {
   sourceFreshnessMs?: number;
   sourceLimit?: number;
   workdayPageDelayMs?: number;
+  workdayCountries?: string[];
   pacingNow?: () => number;
   pacingSleep?: (delayMs: number) => Promise<void>;
   now?: () => Date;
@@ -94,6 +95,7 @@ export function createCrawler(options: CrawlerOptions): Crawler {
               const jobs = await options.fetchJobs(source, controller.signal, {
                 onBackoff: ({ status, delayMs }) => { metric.backoffMs += delayMs; if (status === 429) metric.throttles += 1; },
                 workdayPageDelayMs: options.workdayPageDelayMs,
+                workdayCountries: options.workdayCountries,
               });
               partitions[source.slug] = { fetchedAt: now().toISOString(), jobs };
               succeeded += 1;

package/src/local-jobs.ts

@@ -14,6 +14,7 @@ interface LocalJobsOptions {
   sourceFreshnessMs?: number;
   sourceLimit?: number;
   workdayPageDelayMs?: number;
+  workdayCountries?: string[];
   now?: () => Date;
   onCrawled?(source: Company, partition: JobPartition): void | Promise<void>;
 }

package/src/locations.ts

@@ -104,8 +104,8 @@ function detectRegions(location: string, description: string): string[] {
 /** US state and district postal codes. Many collide with ISO country codes (IN, CA, DE, CO, GA, ID, IL, LA, MA, MD, MO, MT, NE, PA, SC, SD, TN, VA). */
 const US_STATE_CODES = new Set(["AL", "AK", "AZ", "AR", "CA", "CO", "CT", "DE", "FL", "GA", "HI", "ID", "IL", "IN", "IA", "KS", "KY", "LA", "ME", "MD", "MA", "MI", "MN", "MS", "MO", "MT", "NE", "NV", "NH", "NJ", "NM", "NY", "NC", "ND", "OH", "OK", "OR", "PA", "RI", "SC", "SD", "TN", "TX", "UT", "VT", "VA", "WA", "WV", "WI", "WY", "DC"]);
 const US_CITY_STATE = /,\s*([A-Z]{2})(?=\s*(?:$|,|\d{5}|\(|\/|-))/g;
-/** Workday's "ST-CITY, street" shape, e.g. "IN-INDIANAPOLIS, 220 VIRGINIA AVE". */
-const US_STATE_PREFIX = /^([A-Z]{2})-(?=[A-Za-z])/;
+/** Workday's "ST-CITY, street" and "ST - City" shapes, e.g. "IN-INDIANAPOLIS, 220 VIRGINIA AVE" or "IN - Indianapolis". */
+const US_STATE_PREFIX = /^([A-Z]{2})\s*-\s*(?=[A-Za-z])/;
 
 function detectCountries(location: string): string[] {
   const byName: string[] = [];

package/src/runtime.ts

@@ -13,7 +13,8 @@ import { createFileSnapshotStore } from "./snapshot-store.ts";
 import { createJobCoverageReader } from "./job-coverage.ts";
 import { createJobSearchPreparer } from "./job-search-preparation.ts";
 
-export function createRuntime(options: { dataDir?: string; concurrency?: number; crawlDelayMs?: number; workdayPageDelayMs?: number; sourceCacheHours?: number; sourceLimit?: number } = {}) {
+export function createRuntime(options: { dataDir?: string; concurrency?: number; crawlDelayMs?: number; workdayPageDelayMs?: number; workdayCountries?: string[]; sourceCacheHours?: number; sourceLimit?: number } = {}) {
+  const workdayCountries = options.workdayCountries ?? (process.env.OPENINGS_WORKDAY_COUNTRIES ?? "IN,US").split(",").map((code) => code.trim().toUpperCase()).filter((code) => /^[A-Z]{2}$/.test(code));
   const dataDir = options.dataDir ?? process.env.OPENINGS_DATA_DIR ?? join(process.cwd(), ".openings");
   const store = createFileSnapshotStore(join(dataDir, "snapshot.json"));
   const aggregatorUrl = resolveAggregatorUrl(process.env.OPENINGS_AGGREGATOR_URL);
@@ -28,6 +29,7 @@ export function createRuntime(options: { dataDir?: string; concurrency?: number;
     sourceFreshnessMs: (options.sourceCacheHours ?? 0) * 60 * 60 * 1000,
     sourceLimit: options.sourceLimit,
     workdayPageDelayMs: options.workdayPageDelayMs,
+    workdayCountries,
     onCrawled,
   });
   const recommender = createJobRecommender({ sources: companies, store, crawl: local.crawl });
@@ -41,6 +43,7 @@ export function createRuntime(options: { dataDir?: string; concurrency?: number;
     maxAttempts: 1,
     sourceStartDelayMs: options.crawlDelayMs,
     workdayPageDelayMs: options.workdayPageDelayMs,
+    workdayCountries,
     onCrawled,
   });
   const preparation = createJobSearchPreparer({ sources: companies, store, crawl: preparationLocal.crawl, seed: aggregatorUrl ? (countries) => fetchSeedSnapshot(aggregatorUrl, globalThis.fetch, 20_000, countries) : undefined });

package/src/version.ts

@@ -1 +1 @@
-export const VERSION = "0.1.10";
+export const VERSION = "0.1.12";