ai-job-search-app/scripts/import-legacy.ts

/**
 * One-time import of the legacy markdown job files + resumes into the DB.
 * Usage: npx tsx scripts/import-legacy.ts
 */
import { PrismaClient } from "@prisma/client";
import matter from "gray-matter";
import fs from "node:fs";
import path from "node:path";
import os from "node:os";

const prisma = new PrismaClient();
const LEGACY = path.join(os.homedir(), "development/personal/ai-job-search");

function section(body: string, name: string): string | undefined {
  const re = new RegExp(`##\\s*${name}\\s*\\n([\\s\\S]*?)(?=\\n##\\s|$)`, "i");
  return body.match(re)?.[1]?.trim();
}

function firstMatch(body: string, re: RegExp): string | undefined {
  return body.match(re)?.[1]?.trim();
}

async function importJobs() {
  const dir = path.join(LEGACY, "jobs");
  const files = fs.readdirSync(dir).filter((f) => f.endsWith(".md"));
  for (const file of files) {
    const raw = fs.readFileSync(path.join(dir, file), "utf8");
    const { data, content } = matter(raw);
    const title = (data.title as string) ?? file.replace(/\.md$/, "");
    const company = (data.company as string) ?? "Unknown";
    const existing = await prisma.job.findFirst({ where: { title, company } });
    if (existing) {
      console.log(`skip (exists): ${file}`);
      continue;
    }
    const url = firstMatch(content, /\*\*Link:\*\*\s*(\S+)/);
    const applyUrl = firstMatch(content, /\*\*Apply(?: link)?:\*\*\s*(\S+)/);
    const details = section(content, "Details");
    const payLine = details ? firstMatch(details, /\*\*Pay(?:\s*Range)?:\*\*\s*(.+)/) : undefined;
    const location = details ? firstMatch(details, /\*\*Location:\*\*\s*(.+)/) : undefined;

    // crude pay parse: "$15-20/hr", "₱150-170k/mo", "$65-120/hr"
    let payMin: number | undefined, payMax: number | undefined, payCurrency: string | undefined, payPeriod: string | undefined;
    if (payLine) {
      const m = payLine.match(/([$₱P])\s*([\d,.]+)\s*[-–]\s*([\d,.]+)(k?)\s*\/?\s*(hr|hour|mo|month|yr|year)?/i);
      if (m) {
        const mult = m[4] === "k" ? 1000 : 1;
        payMin = parseFloat(m[2].replace(/,/g, "")) * mult;
        payMax = parseFloat(m[3].replace(/,/g, "")) * mult;
        payCurrency = m[1] === "$" ? "USD" : "PHP";
        payPeriod = /hr|hour/i.test(m[5] ?? "") ? "hour" : /mo/i.test(m[5] ?? "") ? "month" : /yr|year/i.test(m[5] ?? "") ? "year" : undefined;
      }
    }

    const phEligibility = /philippines|ph eligible|170\+ countries|worldwide/i.test(content) ? "yes" : "unclear";
    const status = (data.status as string)?.replace(/\s+/g, "-") === "not-applied" ? "prospect" : "prospect";

    await prisma.job.create({
      data: {
        title,
        company,
        source: file.replace(/\.md$/, ""),
        url,
        applyUrl,
        payMin,
        payMax,
        payCurrency,
        payPeriod,
        location,
        remoteType: "remote",
        phEligibility,
        description: section(content, "Description"),
        requirements: section(content, "Requirements"),
        notes: section(content, "Notes") ?? section(content, "Tailoring notes"),
        status,
      },
    });
    console.log(`imported: ${file}`);
  }
}

async function importResumes() {
  const base = path.join(LEGACY, "resume/resume.md");
  if (fs.existsSync(base)) {
    const md = fs.readFileSync(base, "utf8");
    const exists = await prisma.resume.findFirst({ where: { kind: "base" } });
    if (!exists) {
      await prisma.resume.create({ data: { name: "Base Resume", kind: "base", markdown: md } });
      console.log("imported: base resume");
    }
  }
  const tailoredDir = path.join(LEGACY, "resume/tailored");
  if (fs.existsSync(tailoredDir)) {
    for (const f of fs.readdirSync(tailoredDir).filter((f) => f.endsWith(".md"))) {
      const name = `Tailored: ${f.replace(/\.md$/, "")}`;
      const exists = await prisma.resume.findFirst({ where: { name } });
      if (exists) continue;
      await prisma.resume.create({
        data: { name, kind: "tailored", markdown: fs.readFileSync(path.join(tailoredDir, f), "utf8") },
      });
      console.log(`imported: tailored ${f}`);
    }
  }
}

async function seedProfile() {
  const exists = await prisma.profile.findUnique({ where: { id: "main" } });
  if (exists) return;
  await prisma.profile.create({
    data: {
      id: "main",
      fullName: "Joshua C. Sarmiento",
      email: "joshuacsarmiento@gmail.com",
      location: "Philippines",
      headline: "Software Developer / AI Engineer — Prompt Engineering, Conversational AI, Backend Systems",
      summary:
        "Software developer with 8+ years of experience spanning enterprise IT, full-stack web development, and applied AI: building LLM-powered APIs and RAG systems with FastAPI and Amazon Bedrock.",
      skills: JSON.stringify([
        "Prompt engineering", "Amazon Bedrock", "RAG", "Conversational AI", "Python", "FastAPI",
        "REST API design", "Active Directory", "Full-stack web development", "Bash/PowerShell automation",
      ]),
      workHistory: JSON.stringify([
        {
          position: "Project Technical Specialist III",
          company: "Philippine Institute of Volcanology and Seismology (PHIVOLCS)",
          unit: "Geomatics and Hazard Assessment Services Section",
          start: "2025-03", end: null, supervisor: "Eric Santos",
          bullets: [
            "Contributed to CLUP and CDRA systems",
            "Gates Handa: AI APIs with FastAPI + Amazon Bedrock, chatbot, RAG document/report analyzer",
            "Design system diagrams and plans; support project management",
          ],
        },
        {
          position: "Computer Programmer II",
          company: "Presidential Commission for the Urban Poor (PCUP)",
          unit: "IT Unit",
          start: "2019-08", end: "2024-02", supervisor: "Jason Ronidel",
          bullets: [
            "Developed major web apps: Attendance Monitoring, Service Workflow Management, UPO Accreditation, Reservation, Directory",
            "Managed commission network infrastructure and devices",
            "MS Access + VBA database applications",
          ],
        },
        {
          position: "IT Associate",
          company: "Manulife Business Process Services",
          unit: "Global Access Management (GAM) Division",
          start: "2016-11", end: "2017-10", supervisor: "Emory Lopez",
          bullets: [
            "Automated Active Directory user management with scripts (.sh, .ps1, .bat)",
            "Managed AD, mainframe, Oracle, FTP, and email accounts (Exchange, Lotus Notes, O365)",
            "ServiceNow request handling; VPN token creation",
          ],
        },
      ]),
    },
  });
  console.log("seeded: profile");
}

async function main() {
  await importJobs();
  await importResumes();
  await seedProfile();
  console.log("done");
}

main().finally(() => prisma.$disconnect());