/**
 * Série de odds Polymarket + divergência mercado × pesquisa do Chile... Alemanha 2025.
 * Mercado "Germany Parliamentary Election Winner" (P de o partido vencer MAIS cadeiras) × voto por
 * partido nas pesquisas. Via função Vercel/EUA. Molde Chile/Peru. Reprodutível e local.
 *
 * Saída: data/germany-market-odds-timeseries.csv, data/germany-divergence-timeseries.csv
 */
import { readFileSync, writeFileSync, mkdirSync } from 'fs'
import { join } from 'path'

const ROOT = process.cwd()
const OUT = join(ROOT, 'data'); mkdirSync(OUT, { recursive: true })
const csv = (rows) => rows.map((r) => r.map((v) => { const s = String(v ?? ''); return /[",\n]/.test(s) ? `"${s.replace(/"/g, '""')}"` : s }).join(',')).join('\n') + '\n'
const num = (s) => { const m = String(s).replace(/,/g, '').match(/-?\d+(?:\.\d+)?/); return m ? parseFloat(m[0]) : null }

const CANON = [
  ['cdu', 'CDU/CSU'], ['union', 'CDU/CSU'], ['spd', 'SPD'], ['afd', 'AfD'],
  ['green', 'Grüne'], ['grüne', 'Grüne'], ['bsw', 'BSW'], ['linke', 'Die Linke'], ['left', 'Die Linke'], ['fdp', 'FDP'],
]
const canon = (s) => { const t = String(s || '').toLowerCase(); for (const [k, v] of CANON) if (t.includes(k)) return v; return null }
const isoOf = (t) => new Date(t * 1000).toISOString().slice(0, 10)

const raw = JSON.parse(readFileSync(join(OUT, 'germany-poly-raw.json'), 'utf-8'))
const odds = []
for (const m of raw.odds || []) {
  const name = canon(m.candidate); if (!name || !Array.isArray(m.history) || !m.history.length) continue
  const byDay = new Map()
  for (const pt of m.history) byDay.set(isoOf(pt.t), pt.p)
  for (const [date, p] of byDay) odds.push({ date, candidate: name, pct: Math.round(p * 1000) / 10, volume: Math.round(m.volume || 0) })
}
odds.sort((a, b) => a.date === b.date ? b.pct - a.pct : a.date.localeCompare(b.date))
writeFileSync(join(OUT, 'germany-market-odds-timeseries.csv'), csv([['date', 'party', 'polymarket_pct', 'volume_usd'], ...odds.map((o) => [o.date, o.candidate, o.pct, o.volume])]))
console.log(`📈 market-odds: ${odds.length} linhas, ${new Set(odds.map((o) => o.date)).size} datas, ${new Set(odds.map((o) => o.candidate)).size} partidos (${raw.startDate?.slice(0, 10)}→${raw.endDate?.slice(0, 10)})`)

const idx = {}
for (const o of odds) (idx[o.candidate] ||= []).push({ date: o.date, pct: o.pct })
for (const k in idx) idx[k].sort((a, b) => a.date.localeCompare(b.date))
const mAt = (k, d) => { const a = idx[k]; if (!a) return null; let best = null; for (const e of a) { if (e.date <= d) best = e; else break } return best }

const lines = readFileSync(join(ROOT, 'polls', 'germany-polls.csv'), 'utf-8').trim().split('\n').slice(1)
const dv = [['poll_date', 'pollster', 'party', 'poll_pct', 'polymarket_pct', 'polymarket_date', 'divergence_pp']]
let matched = 0
for (const ln of lines) {
  const c = ln.split(','); const poll_date = c[0], pollster = c[2], party = c[4], poll = num(c[6])
  const k = canon(party); if (!k || poll == null) continue
  const m = mAt(k, poll_date); if (!m) continue
  dv.push([poll_date, pollster, k, poll, m.pct, m.date, Math.round((m.pct - poll) * 100) / 100]); matched++
}
writeFileSync(join(OUT, 'germany-divergence-timeseries.csv'), csv(dv))
console.log(`📊 divergence: ${matched} linhas (voto por partido × mercado "vence mais cadeiras")`)