// Regenerate JSON extraction files from the source workbook.
// Reads: ./data/source/original.xlsx
// Writes: ./data/source/{assets,machines,mrc,overheads}.json
//
// Usage: node scripts/extract.mjs
import XLSX from "xlsx";
import { writeFileSync } from "node:fs";
import { fileURLToPath } from "node:url";
import { dirname, resolve } from "node:path";

const __dirname = dirname(fileURLToPath(import.meta.url));
const SRC = resolve(__dirname, "..", "data", "source", "original.xlsx");

console.log(`→ Reading ${SRC}`);
const wb = XLSX.readFile(SRC, { cellFormula: false });

function find(ws, range) {
  const out = [];
  const r = XLSX.utils.decode_range(range);
  for (let R = r.s.r; R <= r.e.r; R++) {
    const row = [];
    for (let C = r.s.c; C <= r.e.c; C++) {
      const addr = XLSX.utils.encode_cell({ r: R, c: C });
      row.push(ws[addr] ? ws[addr].v : null);
    }
    out.push(row);
  }
  return out;
}

// === ASSETS from Sheet1 ===
console.log("\n→ Extracting assets from Sheet1");
const sheet1 = wb.Sheets["Sheet1"];
const assets = [];
// Plant Overheads (rows 6-42)
for (let r = 5; r <= 41; r++) {
  const name = sheet1[XLSX.utils.encode_cell({ r, c: 1 })]?.v;
  if (!name) continue;
  const code = sheet1[XLSX.utils.encode_cell({ r, c: 5 })]?.v ?? "";
  const excl = sheet1[XLSX.utils.encode_cell({ r, c: 6 })]?.v;
  const incl = sheet1[XLSX.utils.encode_cell({ r, c: 8 })]?.v;
  const pv = sheet1[XLSX.utils.encode_cell({ r, c: 9 })]?.v;
  const rate = sheet1[XLSX.utils.encode_cell({ r, c: 11 })]?.v;
  assets.push({
    name: String(name).trim(),
    code: String(code).trim(),
    category: "Plant",
    monthly_excl: typeof excl === "number" ? excl : 0,
    monthly_incl: typeof incl === "number" ? incl : 0,
    purchase_value: typeof pv === "number" ? pv : 0,
    site_rate: typeof rate === "number" ? rate : 0,
  });
}
// Vehicle Costs (rows 46-50)
for (let r = 45; r <= 49; r++) {
  const name = sheet1[XLSX.utils.encode_cell({ r, c: 1 })]?.v;
  if (!name) continue;
  const code = sheet1[XLSX.utils.encode_cell({ r, c: 5 })]?.v ?? "";
  const excl = sheet1[XLSX.utils.encode_cell({ r, c: 6 })]?.v;
  const incl = sheet1[XLSX.utils.encode_cell({ r, c: 8 })]?.v;
  const rate = sheet1[XLSX.utils.encode_cell({ r, c: 11 })]?.v;
  assets.push({
    name: String(name).trim(),
    code: String(code).trim(),
    category: "Vehicle",
    monthly_excl: typeof excl === "number" ? excl : 0,
    monthly_incl: typeof incl === "number" ? incl : 0,
    purchase_value: 0,
    site_rate: typeof rate === "number" ? rate : 0,
  });
}
// Bastion finance/insurance/tracking (rows 54-66)
for (let r = 53; r <= 65; r++) {
  const name = sheet1[XLSX.utils.encode_cell({ r, c: 1 })]?.v;
  if (!name) continue;
  const code = sheet1[XLSX.utils.encode_cell({ r, c: 5 })]?.v ?? "";
  const gl = sheet1[XLSX.utils.encode_cell({ r, c: 2 })]?.v;
  const sup = sheet1[XLSX.utils.encode_cell({ r, c: 3 })]?.v;
  const excl = sheet1[XLSX.utils.encode_cell({ r, c: 6 })]?.v;
  const incl = sheet1[XLSX.utils.encode_cell({ r, c: 8 })]?.v;
  assets.push({
    name: String(name).trim(),
    code: String(code).trim(),
    category: (gl ? String(gl).trim() : "Finance"),
    gl_account: gl ? String(gl).trim() : "",
    supplier: sup ? String(sup).trim() : "",
    monthly_excl: typeof excl === "number" ? excl : 0,
    monthly_incl: typeof incl === "number" ? incl : 0,
    purchase_value: 0,
    site_rate: 0,
  });
}
writeFileSync(resolve(__dirname, "..", "data", "source", "assets.json"), JSON.stringify(assets, null, 2));
console.log(`  ✓ ${assets.length} assets`);

// === MACHINES from Machine Graph Data ===
console.log("\n→ Extracting machines from Machine Graph Data");
const wsMGD = wb.Sheets["Machine Graph Data"];
const monthNames = ["Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"];

// Part C.3: assign sequential codes for uncoded duplicates in the workbook.
// Maps "TYPE_<n>" to a counter so "SITE CONTANER #1" → SC01, etc.
// We also fix the spelling (CONTANER → CONTAINER) so it matches assets.json.
// Each entry: { prefix, test(name) → bool, normalised(name) → string }
const UNCODED_TYPES = [
  { prefix: "SC", test: (n) => /^SITE CONTANER/i.test(n) || /^Site Container/i.test(n) },
  { prefix: "OC", test: (n) => /^Office Container/i.test(n) },
  { prefix: "PC", test: (n) => /^Porta camp/i.test(n) },
  { prefix: "RV02", test: (n) => /^ELB 1TON RV/i.test(n) },
  { prefix: "GN100", test: (n) => /^10 KVA GENERATOR/i.test(n) },
  { prefix: "GN075", test: (n) => /^7,5 KVA GENERATOR/i.test(n) },
  { prefix: "GN300", test: (n) => /^30KVA GENERATOR/i.test(n) },
  { prefix: "TR01", test: (n) => /^TRAILOR PLANT/i.test(n) },
  { prefix: "TR02", test: (n) => /^DIESEL TRAILOR/i.test(n) },
  { prefix: "SO01", test: (n) => /^Solar office/i.test(n) },
];
const uncodedCounters = {};
function codeForUncoded(rawName) {
  // Normalise the workbook spelling to match assets.json
  const name = rawName.replace(/^SITE CONTANER/, "Site Container").trim();
  const match = UNCODED_TYPES.find((t) => t.test(name));
  if (!match) return null;
  // Fixed code (single instance) vs sequential
  if (/^\D+\d+$/.test(match.prefix)) return match.prefix;
  uncodedCounters[match.prefix] = (uncodedCounters[match.prefix] || 0) + 1;
  return `${match.prefix}${String(uncodedCounters[match.prefix]).padStart(2, "0")}`;
}

const machineRows = [];
let currentMachine = null;
let currentData = [];
const range = XLSX.utils.decode_range(wsMGD["!ref"]);
for (let r = range.s.r; r <= range.e.r; r++) {
  const a = wsMGD[XLSX.utils.encode_cell({ r, c: 0 })]?.v;
  const b = wsMGD[XLSX.utils.encode_cell({ r, c: 1 })]?.v;
  if (a == null && b == null) {
    if (currentMachine && currentData.length) {
      machineRows.push({ machine: currentMachine, data: currentData });
      currentMachine = null;
      currentData = [];
    }
    continue;
  }
  if (typeof a === "string" && a === "Month" && typeof b === "string" && b === "Income") continue;
  if (typeof a === "string" && monthNames.includes(a)) {
    if (!currentMachine) continue;
    const income = wsMGD[XLSX.utils.encode_cell({ r, c: 1 })]?.v;
    const fixed = wsMGD[XLSX.utils.encode_cell({ r, c: 2 })]?.v;
    const running = wsMGD[XLSX.utils.encode_cell({ r, c: 4 })]?.v;
    currentData.push({
      month: monthNames.indexOf(a) + 1,
      income: typeof income === "number" ? income : 0,
      fixed_expenses: typeof fixed === "number" ? fixed : 0,
      running_cost: typeof running === "number" ? running : 0,
    });
    continue;
  }
  if (typeof a === "string" && !monthNames.includes(a)) {
    if (currentMachine && currentData.length) machineRows.push({ machine: currentMachine, data: currentData });
    currentMachine = a;
    currentData = [];
  }
}
if (currentMachine && currentData.length) machineRows.push({ machine: currentMachine, data: currentData });

// Strip code from "NAME (CODE)" pattern OR assign sequential code for uncoded types
const machines = machineRows.map((m) => {
  const match = m.machine.match(/\(([^)]+)\)\s*$/);
  if (match) {
    return {
      machine: m.machine.replace(/\s*\([^)]+\)\s*$/, "").trim(),
      code: match[1].trim(),
      data: m.data,
    };
  }
  // Try sequential-code assignment (Part C.3)
  const syntheticCode = codeForUncoded(m.machine);
  return {
    machine: m.machine.replace(/^SITE CONTANER/, "Site Container").trim(),
    code: syntheticCode || "",
    data: m.data,
  };
});
writeFileSync(resolve(__dirname, "..", "data", "source", "machines.json"), JSON.stringify(machines, null, 2));
console.log(`  ✓ ${machines.length} machines (sequential codes: ${Object.keys(uncodedCounters).length} prefixes)`);

// === MRC from Monthly Running Cost sheet ===
console.log("\n→ Extracting monthly running costs");
const wsMRC = wb.Sheets["Monthly Running Cost"];
const monthCols = { 1: 3, 2: 5, 3: 7, 4: 9, 5: 11, 6: 13, 7: 15, 8: 17 };
const mrc = [];
const mrcRange = XLSX.utils.decode_range(wsMRC["!ref"]);
for (let r = 3; r <= 46; r++) {
  const name = wsMRC[XLSX.utils.encode_cell({ r, c: 1 })]?.v;
  const code = wsMRC[XLSX.utils.encode_cell({ r, c: 2 })]?.v;
  if (!code) continue;
  for (const [month, col] of Object.entries(monthCols)) {
    const v = wsMRC[XLSX.utils.encode_cell({ r, c: col })]?.v;
    if (typeof v === "number" && v > 0) {
      mrc.push({
        name: name ? String(name).trim() : String(code).trim(),
        code: String(code).trim(),
        month: Number(month),
        amount: v,
      });
    }
  }
}
writeFileSync(resolve(__dirname, "..", "data", "source", "mrc.json"), JSON.stringify(mrc, null, 2));
console.log(`  ✓ ${mrc.length} running-cost rows`);

// === OVERHEADS from Cost sheet ===
console.log("\n→ Extracting overheads from Cost sheet");
const wsCost = wb.Sheets["Cost"];
const sections = [
  { name: "Plant Overheads", rows: [5, 40] },
  { name: "Vehicle Costs", rows: [45, 48] },
  { name: "Bastion Finance", rows: [53, 65] },
  { name: "Personnel", rows: [70, 74] },
  { name: "Subscriptions", rows: [80, 82] },
  { name: "Rent & Utilities", rows: [87, 88] },
];
const overheads = [];
for (const { name, rows } of sections) {
  for (let r = rows[0]; r <= rows[1]; r++) {
    const desc = wsCost[XLSX.utils.encode_cell({ r, c: 1 })]?.v;
    if (!desc) continue;
    const gl = wsCost[XLSX.utils.encode_cell({ r, c: 2 })]?.v;
    const sup = wsCost[XLSX.utils.encode_cell({ r, c: 3 })]?.v;
    const qty = wsCost[XLSX.utils.encode_cell({ r, c: 4 })]?.v;
    const pu = wsCost[XLSX.utils.encode_cell({ r, c: 5 })]?.v;
    const amt = wsCost[XLSX.utils.encode_cell({ r, c: 6 })]?.v;
    const vat = wsCost[XLSX.utils.encode_cell({ r, c: 7 })]?.v;
    const tot = wsCost[XLSX.utils.encode_cell({ r, c: 8 })]?.v;
    overheads.push({
      section: name,
      description: String(desc).trim(),
      gl_account: gl ? String(gl).trim() : "",
      supplier: sup ? String(sup).trim() : "",
      qty: typeof qty === "number" ? qty : 0,
      unit_price: typeof pu === "number" ? pu : 0,
      amount_excl: typeof amt === "number" ? amt : 0,
      vat: typeof vat === "number" ? vat : 0,
      amount_incl: typeof tot === "number" ? tot : 0,
    });
  }
}
writeFileSync(resolve(__dirname, "..", "data", "source", "overheads.json"), JSON.stringify(overheads, null, 2));
console.log(`  ✓ ${overheads.length} overhead rows`);

console.log("\n✅ Extraction complete → ./data/source/");