import { fetchJobs, isEnglishContent } from "../fetch.ts";
import type { SearchConfig } from "../config.ts";

const CSV = "./jobs.csv";

const existingLinks = new Set<string>();
try {
  const text = await Bun.file(CSV).text();
  for (const line of text.split("\n").slice(1)) {
    const url = line.match(/"(https?:\/\/[^"]+)"/)?.[1];
    if (url) existingLinks.add(url);
  }
} catch {}

const RAILS_KEYWORDS = [
  "Ruby on Rails",
  "Ruby Developer",
  "Rails Developer",
  "Ruby Engineer",
  "Rails Engineer",
  "Full Stack Ruby",
  "Backend Ruby",
  "RoR Developer",
  "RoR Engineer",
].join(" OR ");

const searches: SearchConfig[] = [
  {
    label: "Switzerland Remote Rails (Zürich)",
    keywords: RAILS_KEYWORDS,
    geoId: "102436504",
    location: "Zürich, Switzerland",
    workType: ["remote"],
    timePosted: "month",
  },
  {
    label: "Switzerland Remote Rails (by name)",
    keywords: RAILS_KEYWORDS,
    location: "Switzerland",
    workType: ["remote"],
    timePosted: "month",
  },
  {
    label: "Switzerland Remote Rails (Basel)",
    keywords: RAILS_KEYWORDS,
    geoId: "103383283",
    location: "Basel",
    workType: ["remote"],
    timePosted: "month",
  },
];

type Result = { title: string; company: string; id: string; link: string };
const results: Result[] = [];
const seen = new Set<string>();

for (const config of searches) {
  console.log(`\n[${config.label}]`);
  const jobs = await fetchJobs(config, { raw: true }); // raw: bypass SOFTWARE_TITLE_RE, we want all Rails hits

  for (const job of jobs) {
    if (seen.has(job.id) || existingLinks.has(job.link)) continue;
    seen.add(job.id);

    process.stdout.write(`  checking content: ${job.title} @ ${job.company} … `);
    const english = await isEnglishContent(job.id);
    if (!english) {
      console.log("dropped (non-English)");
      continue;
    }
    console.log("ok");
    results.push({ title: job.title, company: job.company, id: job.id, link: job.link });
  }
}

console.log(`\n${results.length} new Rails remote jobs found.\n`);
for (const r of results) {
  console.log(`${r.title} | ${r.company} | ${r.link}`);
}