openings 0.1.39 → 0.1.42

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "openings",
-  "version": "0.1.39",
+  "version": "0.1.42",
   "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": "0.1.39",
+  "version": "0.1.42",
   "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

@@ -1,5 +1,6 @@
 import { CASCADE_MINIMUM, CASCADE_WINDOWS, type Company, type Job, type JobAge, type JobSummary, type SearchQuery, type SearchWindow } from "./types.ts";
 import { decodeEntities, providerSpec, type JsonGet } from "./providers.ts";
+import { matchesSearchTerms, searchHaystack, searchTerms } from "./text-match.ts";
 import { crawlSite, sitePostingsToJobs } from "./jobposting-site.ts";
 import { classifyJob, isEligibleForCountry, normalizeLocation } from "./locations.ts";
 import { experienceMatches, normalizeJobExperience } from "./experience.ts";
@@ -514,10 +515,11 @@ function normalizeRecruitee(company: Company, job: RecruiteeJob): Job {
 }
 
 function matches(job: Job, query: SearchQuery): boolean {
-  const terms = query.query?.trim().toLocaleLowerCase().split(/\s+/).filter(Boolean) ?? [];
+  const terms = searchTerms(query.query);
   const location = query.location ? normalizeLocation(query.location) : undefined;
-  const searchable = `${job.title} ${job.company}`.toLocaleLowerCase();
-  return (terms.length === 0 || terms.every((term) => searchable.includes(term)))
+  // Whole tokens only: a substring match once made "ios" find "Axio Biosolutions".
+  const searchable = searchHaystack(`${job.title} ${job.company}`);
+  return (terms.length === 0 || matchesSearchTerms(searchable, terms))
     && (!location || normalizeLocation(job.location).includes(location))
     && (!query.country || isEligibleForCountry(job, query.country))
     && (query.remote === undefined || job.remote === query.remote);

package/src/index.ts

@@ -16,8 +16,9 @@ export const companies: Company[] = Object.entries(companyData).map(([slug, valu
 }));
 
 export const catalog = createCatalog({ companies });
-export { createCatalog } from "./catalog.ts";
+export { createCatalog, searchJobs } from "./catalog.ts";
 export { EXPERIENCE_VERSION, experienceLabel, experienceMatches, normalizeJobExperience, statedExperience, titleExperience, type Experience } from "./experience.ts";
+export { matchesSearchTerms, searchHaystack, searchTerms, searchTokens } from "./text-match.ts";
 export type { Catalog } from "./catalog.ts";
 export type { Ats, Company, CrawlReport, Job, JobPartition, JobSnapshot, JobSummary, SearchQuery } from "./types.ts";
 export { createCrawlReporter, fetchSeedSnapshot, resolveAggregatorUrl } from "./crawl-reporting.ts";

package/src/text-match.ts

@@ -0,0 +1,41 @@
+/**
+ * Search matching on whole tokens, so "ios" finds an iOS role and never "Axio Biosolutions".
+ * Technology names keep the characters that carry their meaning: c++, c#, node.js, .net.
+ */
+
+/** Everything outside this set separates tokens, so "full-stack" and "full stack" agree. */
+const SEPARATORS = /[^a-z0-9.+#]+/g;
+
+/** A trailing plural "s" is dropped on both sides, so "engineers" and "engineer" meet in the middle. */
+function stem(token: string): string {
+  if (token.length < 4 || /[+#]/.test(token)) return token;
+  return token.endsWith("s") && !token.endsWith("ss") ? token.slice(0, -1) : token;
+}
+
+export function searchTokens(text: string): string[] {
+  return text.toLocaleLowerCase().replace(SEPARATORS, " ").split(" ")
+    .map((token) => token.replace(/^\.+/, "").replace(/\.+$/, ""))
+    .filter(Boolean)
+    .map(stem);
+}
+
+/** The text a job is matched against, padded so a term only matches a whole token. A dotted name is also
+ * indexed by its parts, so "node" finds "Node.js" while "node.js" still does. */
+export function searchHaystack(text: string): string {
+  const indexed = new Set<string>();
+  for (const token of searchTokens(text)) {
+    indexed.add(token);
+    if (token.includes(".")) for (const part of token.split(".")) if (part) indexed.add(stem(part));
+  }
+  return ` ${[...indexed].join(" ")} `;
+}
+
+/** The terms of a query; an empty query matches everything. */
+export function searchTerms(query: string | undefined, limit = 8): string[] {
+  return searchTokens(query ?? "").slice(0, limit);
+}
+
+/** Every term must appear as a whole token. */
+export function matchesSearchTerms(haystack: string, terms: string[]): boolean {
+  return terms.every((term) => haystack.includes(` ${term} `));
+}

package/src/version.ts

@@ -1 +1 @@
-export const VERSION = "0.1.39";
+export const VERSION = "0.1.42";