Jobber/extractors/greenhouse/manifest.ts

/**
 * Greenhouse public job boards API.
 *
 * https://developers.greenhouse.io/job-board.html
 *   GET https://boards-api.greenhouse.io/v1/boards/{company}/jobs?content=true
 *
 * No auth. Each entry in `greenhouseCompanies` is fetched independently.
 */

import type {
  ExtractorManifest,
  ExtractorRunResult,
} from "@shared/types/extractors";
import type { CreateJobInput } from "@shared/types/jobs";

interface GhDepartment {
  id?: number;
  name?: string;
}
interface GhMetadata {
  name?: string;
  value?: unknown;
}
interface GhJob {
  id?: number;
  title?: string;
  absolute_url?: string;
  internal_job_id?: number;
  updated_at?: string;
  requisition_id?: string | null;
  location?: { name?: string };
  content?: string; // HTML, may be entity-encoded
  metadata?: GhMetadata[];
  departments?: GhDepartment[];
  offices?: Array<{ name?: string }>;
}
interface GhResponse {
  jobs?: GhJob[];
}

function asString(value: unknown): string | undefined {
  if (typeof value !== "string") return undefined;
  const trimmed = value.trim();
  return trimmed ? trimmed : undefined;
}

function readCompanies(raw: string | undefined): string[] {
  if (!raw) return [];
  try {
    const parsed = JSON.parse(raw);
    if (Array.isArray(parsed)) {
      return parsed
        .map((entry) => (typeof entry === "string" ? entry.trim() : ""))
        .filter(Boolean);
    }
  } catch {
    // fall through
  }
  return raw
    .split(/[\n,;|]+/)
    .map((entry) => entry.trim())
    .filter(Boolean);
}

function decodeHtmlEntities(value: string): string {
  return value
    .replace(/&amp;/g, "&")
    .replace(/&lt;/g, "<")
    .replace(/&gt;/g, ">")
    .replace(/&quot;/g, '"')
    .replace(/&#39;/g, "'")
    .replace(/&nbsp;/g, " ");
}

function mapJob(job: GhJob, company: string): CreateJobInput | null {
  const jobUrl = asString(job.absolute_url);
  if (!jobUrl) return null;

  const employer = company
    .split(/[-_]/)
    .filter(Boolean)
    .map((part) => part.charAt(0).toUpperCase() + part.slice(1))
    .join(" ");
  const officeNames =
    job.offices
      ?.map((office) => asString(office.name))
      .filter((name): name is string => Boolean(name)) ?? [];
  const departmentNames =
    job.departments
      ?.map((dept) => asString(dept.name))
      .filter((name): name is string => Boolean(name)) ?? [];

  const description = job.content ? decodeHtmlEntities(job.content) : undefined;

  return {
    source: "greenhouse",
    sourceJobId: job.id != null ? String(job.id) : undefined,
    title: asString(job.title) ?? "Unknown Title",
    employer: employer || company,
    jobUrl,
    applicationLink: jobUrl,
    location:
      asString(job.location?.name) ?? (officeNames.join("; ") || undefined),
    jobFunction:
      departmentNames.length > 0 ? departmentNames.join(", ") : undefined,
    datePosted: asString(job.updated_at),
    jobDescription: description,
  };
}

async function fetchCompany(company: string): Promise<GhJob[]> {
  const url = `https://boards-api.greenhouse.io/v1/boards/${encodeURIComponent(company)}/jobs?content=true`;
  const response = await fetch(url, {
    headers: { Accept: "application/json" },
  });
  if (response.status === 404) return [];
  if (!response.ok) {
    throw new Error(
      `Greenhouse request for "${company}" failed with status ${response.status}`,
    );
  }
  const body = (await response.json()) as GhResponse;
  return Array.isArray(body.jobs) ? body.jobs : [];
}

export const manifest: ExtractorManifest = {
  id: "greenhouse",
  displayName: "Greenhouse (ATS)",
  providesSources: ["greenhouse"],
  async run(context): Promise<ExtractorRunResult> {
    if (context.shouldCancel?.()) return { success: true, jobs: [] };

    const companies = readCompanies(context.settings.greenhouseCompanies);
    if (companies.length === 0) {
      return {
        success: true,
        jobs: [],
        error:
          "No Greenhouse companies configured. Set GREENHOUSE_COMPANIES or the greenhouseCompanies setting (comma- or newline-separated slugs).",
      };
    }

    const seen = new Set<string>();
    const out: CreateJobInput[] = [];

    try {
      for (let i = 0; i < companies.length; i += 1) {
        if (context.shouldCancel?.()) break;
        const company = companies[i];
        context.onProgress?.({
          phase: "list",
          termsProcessed: i,
          termsTotal: companies.length,
          currentUrl: company,
          detail: `Greenhouse: ${company} (${i + 1}/${companies.length})`,
        });

        let added = 0;
        const jobs = await fetchCompany(company);
        for (const job of jobs) {
          const mapped = mapJob(job, company);
          if (!mapped) continue;
          const key = mapped.sourceJobId || mapped.jobUrl;
          if (seen.has(key)) continue;
          seen.add(key);
          out.push(mapped);
          added += 1;
        }

        context.onProgress?.({
          phase: "list",
          termsProcessed: i + 1,
          termsTotal: companies.length,
          currentUrl: company,
          jobPagesProcessed: out.length,
          detail: `Greenhouse: ${company} → ${added} jobs (${out.length} total)`,
        });
      }
    } catch (error) {
      const message = error instanceof Error ? error.message : "Unknown error";
      return { success: false, jobs: out, error: message };
    }

    return { success: true, jobs: out };
  },
};

export default manifest;