feat: add skills

2026-11-08 16:07:38 +00:00 · 2026-08-26 10:35:34 -04:00
1 parent 6984212b17
commit c50aa92862
@@ -0,0 +1,11 @@
+{
+  "name": "jobsdb-cli",
+  "version": "1.0.0",
+  "description": "CLI for searching a local SQLite snapshot of job postings (documents/postings/jobs.db). Zero runtime dependencies: bun:sqlite only.",
+  "type": "module",
+  "main": "src/cli.ts",
+  "bin": { "jobsdb-search": "src/cli.ts" },
+  "scripts": { "start": "bun run src/cli.ts", "typecheck": "tsc --noEmit" },
+  "dependencies": {},
+  "devDependencies": { "typescript": "^5.4.0", "@types/bun": "1.3.14" }
+}
@@ -0,0 +1,86 @@
+#!/usr/bin/env bun
+// jobsdb-cli: search a local SQLite snapshot of job postings. Zero dependencies (bun:sqlite).
+import { runSearch, type SearchOpts } from "./commands/search.js"
+import { runDetail } from "./commands/detail.js"
+import { writeError } from "./helpers.js"
+
+const HELP = `jobsdb-cli - search the local postings snapshot (documents/postings/jobs.db)
+
+Usage:
+  cli.ts search [flags]
+  cli.ts detail <posting_id|url> [--format json|plain]
+
+Search flags:
+  -q, --query <text>        Full-text match on title, company, location (FTS5)
+  --company <text>          Company name contains <text>
+  --location <text>         Location text contains <text> (e.g. "New York", "San Francisco")
+  --country <code>          location_country equals <code> (e.g. US)
+  --remote <list>           Comma list of: remote, remote-regional, hybrid, onsite
+  --seniority <list>        Comma list of: senior, staff_principal, manager, director, head_of, vp, c_level_founder, mid_unclassified, junior_associate, intern, unclassified
+  --sector <list>           Comma list of company target_sector values (crypto, fintech-nc, ai-ml, devtools, enterprise-saas, hardware, other)
+  --crypto-native           Only companies flagged crypto-native
+  --tier <n>                Company priority_tier <= n (1 is best)
+  --jobage <days>           Posted within N days of the row's scrape snapshot (not today)
+  --page <n>                1-indexed page. Default 1
+  -n, --limit <n>           Results per page. Default 20
+  --format json|table       Default json
+
+Env:
+  JOBSDB_PATH               Override the database path
+`
+
+type Flags = Record<string, string | boolean>
+
+function parse(argv: string[]): { cmd: string; pos: string[]; flags: Flags } {
+  const [cmd = "", ...rest] = argv
+  const pos: string[] = []
+  const flags: Flags = {}
+  for (let i = 0; i < rest.length; i++) {
+    const a = rest[i]
+    if (a === "-q") { flags.query = rest[++i] ?? ""; continue }
+    if (a === "-n") { flags.limit = rest[++i] ?? ""; continue }
+    if (a.startsWith("--")) {
+      const [k, inline] = a.slice(2).split("=", 2)
+      if (inline !== undefined) { flags[k] = inline; continue }
+      const next = rest[i + 1]
+      if (next === undefined || next.startsWith("-")) flags[k] = true
+      else flags[k] = rest[++i]
+      continue
+    }
+    pos.push(a)
+  }
+  return { cmd, pos, flags }
+}
+
+function num(flags: Flags, name: string, dflt?: number): number | undefined {
+  const raw = flags[name]
+  if (raw === undefined) return dflt
+  const n = Number(raw)
+  if (!Number.isFinite(n) || n < 0) { writeError(`--${name} must be a non-negative number, got "${raw}"`, "BAD_ARG"); process.exit(1) }
+  return n
+}
+const list = (v: string | boolean | undefined) => (typeof v === "string" && v ? v.split(",").map((s) => s.trim()).filter(Boolean) : undefined)
+const str = (v: string | boolean | undefined) => (typeof v === "string" && v ? v : undefined)
+
+const { cmd, pos, flags } = parse(process.argv.slice(2))
+const KNOWN = new Set(["query", "company", "location", "country", "remote", "seniority", "sector", "crypto-native", "tier", "jobage", "page", "limit", "format"])
+for (const k of Object.keys(flags)) if (!KNOWN.has(k)) { writeError(`Unknown flag --${k}`, "BAD_ARG"); process.exit(1) }
+
+if (!cmd || cmd === "help" || cmd === "--help" || cmd === "-h") { process.stdout.write(HELP); process.exit(0) }
+
+if (cmd === "search") {
+  const format = (str(flags.format) ?? "json") as SearchOpts["format"]
+  if (format !== "json" && format !== "table") { writeError(`--format must be json or table`, "BAD_ARG"); process.exit(1) }
+  runSearch({
+    query: str(flags.query), company: str(flags.company), location: str(flags.location), country: str(flags.country),
+    remote: list(flags.remote), seniority: list(flags.seniority), sector: list(flags.sector),
+    cryptoNative: flags["crypto-native"] === true, tier: num(flags, "tier"), jobage: num(flags, "jobage"),
+    page: Math.max(1, num(flags, "page", 1)!), limit: Math.max(1, num(flags, "limit", 20)!), format,
+  })
+} else if (cmd === "detail") {
+  if (!pos[0]) { writeError("detail requires a <posting_id|url>", "NO_ID"); process.exit(1) }
+  const format = (str(flags.format) ?? "json") as "json" | "plain"
+  runDetail(pos[0], format)
+} else {
+  writeError(`Unknown command "${cmd}"`, "BAD_CMD"); process.exit(1)
+}
@@ -0,0 +1,17 @@
+import { openDb, toJob, writeError, type Row } from "../helpers.js"
+
+export function runDetail(idOrUrl: string, format: "json" | "plain"): void {
+  const db = openDb()
+  const row = db
+    .query("SELECT p.*, c.target_sector, c.crypto_native, c.priority_tier FROM postings p LEFT JOIN companies c ON c.company_id = p.company_id WHERE p.posting_id = ? OR p.url = ? LIMIT 1")
+    .get(idOrUrl, idOrUrl) as Row | null
+  if (!row) { writeError(`no posting with id or url ${idOrUrl}`, "NOT_FOUND"); process.exit(1) }
+  const job = toJob(row)
+  if (format === "plain") {
+    process.stdout.write(
+      [`${job.title} - ${job.company}`, `${job.location} | ${job.remote ?? "?"} | ${job.seniority ?? "?"}`, `Posted ${job.date ?? "?"}${job.date_approximate ? " (approx.)" : ""}, snapshot ${job.snapshot}`, job.salary ? `Salary ${job.salary}` : "", `URL: ${job.url}`, "", "The snapshot stores no description. Fetch the URL above for the full posting text."].filter(Boolean).join("\n") + "\n",
+    )
+    return
+  }
+  process.stdout.write(JSON.stringify(job, null, 2) + "\n")
+}
@@ -0,0 +1,72 @@
+import { openDb, toJob, renderTable, writeError, REMOTE_CLASS, type Row } from "../helpers.js"
+
+export interface SearchOpts {
+  query?: string
+  company?: string
+  location?: string
+  country?: string
+  remote?: string[]
+  seniority?: string[]
+  sector?: string[]
+  cryptoNative?: boolean
+  tier?: number
+  jobage?: number
+  page: number
+  limit: number
+  format: "json" | "table"
+}
+
+function ftsMatch(q: string): string {
+  // Quote every token so user text can never be parsed as FTS5 syntax.
+  return q.split(/\s+/).filter(Boolean).map((t) => `"${t.replace(/"/g, '""')}"`).join(" ")
+}
+
+export function runSearch(opts: SearchOpts): void {
+  const db = openDb()
+  const where: string[] = []
+  const params: (string | number)[] = []
+  let from = "postings p LEFT JOIN companies c ON c.company_id = p.company_id"
+  if (opts.query) {
+    from = "postings_fts f JOIN postings p ON p.rowid = f.rowid LEFT JOIN companies c ON c.company_id = p.company_id"
+    where.push("postings_fts MATCH ?")
+    params.push(ftsMatch(opts.query))
+  }
+  if (opts.company) { where.push("p.company_name LIKE ?"); params.push(`%${opts.company}%`) }
+  if (opts.location) { where.push("p.location_text LIKE ?"); params.push(`%${opts.location}%`) }
+  if (opts.country) { where.push("p.location_country = ?"); params.push(opts.country) }
+  if (opts.remote?.length) {
+    const classes = opts.remote.map((r) => REMOTE_CLASS[r.toLowerCase()]).filter(Boolean)
+    if (classes.length !== opts.remote.length) {
+      writeError(`--remote accepts ${Object.keys(REMOTE_CLASS).join(", ")}`, "BAD_ARG"); process.exit(1)
+    }
+    where.push(`p.location_class IN (${classes.map(() => "?").join(",")})`); params.push(...classes)
+  }
+  if (opts.seniority?.length) { where.push(`p.seniority IN (${opts.seniority.map(() => "?").join(",")})`); params.push(...opts.seniority) }
+  if (opts.sector?.length) { where.push(`c.target_sector IN (${opts.sector.map(() => "?").join(",")})`); params.push(...opts.sector) }
+  if (opts.cryptoNative) where.push("c.crypto_native = 'True'")
+  if (opts.tier !== undefined) { where.push("c.priority_tier <= ?"); params.push(opts.tier) }
+
+  const sql = `SELECT p.*, c.target_sector, c.crypto_native, c.priority_tier FROM ${from}
+    ${where.length ? "WHERE " + where.join(" AND ") : ""}
+    ORDER BY COALESCE(c.priority_tier, 9), p.scrape_timestamp DESC`
+  let rows: Row[]
+  try { rows = db.query(sql).all(...params) as Row[] } catch (e) {
+    writeError(`query failed: ${(e as Error).message}`, "QUERY_FAILED"); process.exit(1)
+  }
+  let jobs = rows.map(toJob)
+  if (opts.jobage !== undefined) {
+    // Recency relative to each row's own scrape snapshot, not today's calendar: the DB is a static export.
+    jobs = jobs.filter((j) => {
+      if (!j.date || !j.snapshot) return false
+      const age = (Date.parse(j.snapshot) - Date.parse(j.date)) / 86_400_000
+      return age <= opts.jobage!
+    })
+  }
+  const total = jobs.length
+  const start = (opts.page - 1) * opts.limit
+  const pageJobs = jobs.slice(start, start + opts.limit)
+  const snapshot = pageJobs.reduce<string | null>((m, j) => (j.snapshot && (!m || j.snapshot > m) ? j.snapshot : m), null)
+
+  if (opts.format === "table") { process.stdout.write(renderTable(pageJobs) + "\n"); return }
+  process.stdout.write(JSON.stringify({ meta: { count: pageJobs.length, page: opts.page, total, snapshot, db: process.env.JOBSDB_PATH ?? "documents/postings/jobs.db" }, results: pageJobs }, null, 2) + "\n")
+}
@@ -0,0 +1,116 @@
+import { Database } from "bun:sqlite"
+import { existsSync } from "node:fs"
+
+export const DB_PATH = process.env.JOBSDB_PATH ?? "documents/postings/jobs.db"
+
+export const REMOTE_CLASS: Record<string, string> = { remote: "R1", "remote-global": "R1", "remote-regional": "R2", hybrid: "H", onsite: "O" }
+export const CLASS_LABEL: Record<string, string> = { R1: "Remote (global)", R2: "Remote (regional)", H: "Hybrid", O: "Onsite" }
+
+export interface Row {
+  posting_id: string
+  company_id: string | null
+  company_name: string | null
+  title: string | null
+  seniority: string | null
+  posted_date: string | null
+  location_text: string | null
+  location_class: string | null
+  location_country: string | null
+  salary_min: string | null
+  salary_max: string | null
+  salary_currency: string | null
+  url: string | null
+  source_ats: string | null
+  scrape_timestamp: string | null
+  target_sector: string | null
+  crypto_native: string | null
+  priority_tier: number | null
+}
+
+export interface Job {
+  id: string
+  title: string | null
+  company: string | null
+  location: string | null
+  date: string | null
+  date_approximate: boolean
+  url: string | null
+  description: null
+  remote: string | null
+  seniority: string | null
+  salary: string | null
+  sector: string | null
+  crypto_native: boolean
+  priority_tier: number | null
+  ats: string | null
+  snapshot: string | null
+}
+
+export function openDb(): Database {
+  if (!existsSync(DB_PATH)) {
+    writeError(`database not found at ${DB_PATH} (set JOBSDB_PATH)`, "NO_DB")
+    process.exit(1)
+  }
+  return new Database(DB_PATH, { readonly: true })
+}
+
+/** "Posted 8 Days Ago" + scrape_timestamp -> ISO date. "30+" is a floor, flagged approximate. */
+export function resolveDate(posted: string | null, scraped: string | null): { date: string | null; approximate: boolean } {
+  if (!posted) return { date: null, approximate: false }
+  // greenhouse/ashby/lever store ISO timestamps; workday stores relative text.
+  if (/^\d{4}-\d{2}-\d{2}/.test(posted)) return { date: posted.slice(0, 10), approximate: false }
+  if (!scraped) return { date: null, approximate: false }
+  const base = new Date(scraped)
+  if (isNaN(base.getTime())) return { date: null, approximate: false }
+  const m = /(\d+)(\+?)\s*Days?/i.exec(posted)
+  let days: number
+  let approximate = false
+  if (/Today/i.test(posted)) days = 0
+  else if (/Yesterday/i.test(posted)) days = 1
+  else if (m) { days = Number(m[1]); approximate = m[2] === "+" }
+  else return { date: null, approximate: false }
+  base.setUTCDate(base.getUTCDate() - days)
+  return { date: base.toISOString().slice(0, 10), approximate }
+}
+
+export function toJob(r: Row): Job {
+  const { date, approximate } = resolveDate(r.posted_date, r.scrape_timestamp)
+  const salary = r.salary_min && r.salary_max ? `${r.salary_min}-${r.salary_max} ${r.salary_currency ?? ""}`.trim() : null
+  return {
+    id: r.posting_id,
+    title: r.title || null,
+    company: r.company_name || null,
+    location: r.location_text || null,
+    date,
+    date_approximate: approximate,
+    url: r.url || null,
+    description: null,
+    remote: r.location_class ? CLASS_LABEL[r.location_class] ?? null : null,
+    seniority: r.seniority || null,
+    salary,
+    sector: r.target_sector || null,
+    crypto_native: r.crypto_native === "True",
+    priority_tier: r.priority_tier ?? null,
+    ats: r.source_ats || null,
+    snapshot: r.scrape_timestamp ? r.scrape_timestamp.slice(0, 10) : null,
+  }
+}
+
+export function writeError(error: string, code: string): void {
+  process.stderr.write(JSON.stringify({ error, code }) + "\n")
+}
+
+export function renderTable(jobs: Job[]): string {
+  if (jobs.length === 0) return "No results."
+  const rows = jobs.map((j) =>
+    [
+      (j.title ?? "").slice(0, 44).padEnd(44),
+      (j.company ?? "-").slice(0, 22).padEnd(22),
+      (j.location ?? "-").slice(0, 22).padEnd(22),
+      (j.remote ?? "-").slice(0, 16).padEnd(16),
+      j.date ? j.date + (j.date_approximate ? "~" : " ") : "-".padEnd(11),
+      j.url ?? "",
+    ].join("  "),
+  )
+  return rows.join("\n")
+}
@@ -0,0 +1,14 @@
+{
+  "compilerOptions": {
+    "target": "ESNext",
+    "module": "ESNext",
+    "moduleResolution": "bundler",
+    "types": ["bun"],
+    "strict": true,
+    "noEmit": true,
+    "skipLibCheck": true,
+    "allowImportingTsExtensions": true,
+    "verbatimModuleSyntax": true
+  },
+  "include": ["src/**/*.ts"]
+}