build-a-squad/scripts/parse-assets.ts

/**
 * Asset Catalog Parser
 *
 * Parses `documents/Latest Asset List.md` (pipe-delimited markdown table)
 * and outputs three JSON files into `data/`:
 *   - assets.json        — full Asset[] catalog
 *   - staffingRoutes.json — Record<uniques, Record<role, hours>>
 *   - roleDisciplineMap.json — Record<role, discipline>
 *
 * Run:   npm run parse-catalog
 * When:  After updating `documents/Latest Asset List.md` or the source CSV
 * Check: Review console output for asset count (~334), staffing route count,
 *        and any roles flagged as UNKNOWN discipline.
 */

import * as fs from 'fs';
import * as path from 'path';
import { fileURLToPath } from 'url';

const __filename = fileURLToPath(import.meta.url);
const __dirname = path.dirname(__filename);

// ---------------------------------------------------------------------------
// Types
// ---------------------------------------------------------------------------

interface Asset {
  uniques: string;
  catalogId: string;
  category: string;
  subCategory: string;
  assetName: string;
  complexityLevel: string;
  description: string;
  complexityDescription: string;
  studioCaveats: string;
  isMaster: boolean;
}

// ---------------------------------------------------------------------------
// Base role → discipline map (from existing app data)
// ---------------------------------------------------------------------------

const BASE_ROLE_DISCIPLINE_MAP: Record<string, string> = {
  "Business Director": "Account Management",
  "Account Director / Content Team Leader": "Account Management",
  "Account Director": "Account Management",
  "Account Manager": "Account Management",
  "Senior Account Manager": "Account Management",
  "Group Account Director": "Account Management",
  "Programme Director": "Delivery",
  "Senior Project Manager": "Delivery",
  "Project Manager": "Delivery",
  "Project Manager ": "Delivery",
  "Strategy Director": "Strategy",
  "Planning Director / Strategy Director": "Strategy",
  "Senior Planner / Strategist": "Strategy",
  "Planner / Strategist": "Strategy",
  "Strategist": "Strategy",
  "Strategist ": "Strategy",
  "Executive Creative Director": "Creative",
  "Senior Creative Director": "Creative",
  "Creative Director": "Creative",
  "Creative Director ": "Creative",
  "Associate Creative Director": "Creative",
  "Senior Designer / Lead Creator": "Creative",
  "Senior Designer / Lead Creator ": "Creative",
  "Designer / Creator": "Creative",
  "Designer / Creator ": "Creative",
  "Senior Art Director": "Creative",
  "Art Director": "Creative",
  "Senior Copywriter": "Editorial",
  "Conceptual Copywriter": "Editorial",
  "Copywriter / Brand Journalist": "Editorial",
  "Copywriter / Brand Journalist ": "Editorial",
  "Executive Producer": "Production",
  "Producer": "Production",
  "Producer ": "Production",
  "Senior Producer": "Production",
  "Senior Motion Design / Editor": "Production",
  "Motion Designer / Editor": "Production",
  "Motion Designer / Editor ": "Production",
  "Web (Front-End) Developer": "Tech",
  "Web (Front-End) Developer ": "Tech",
  "Senior Web (Front-End) Developer": "Tech",
  "Content Manager": "Tech",
  "Content Manager ": "Tech",
  "Web Designer": "Tech",
  "Web Designer ": "Tech",
  "CGI Operator (Medium)": "Tech",
  "QA Manager": "QA",
  "QA Manager ": "QA",
};

// ---------------------------------------------------------------------------
// Constants
// ---------------------------------------------------------------------------

const COMPLEXITY_MAP: Record<string, string> = {
  '1': 'Simple',
  '2': 'Medium',
  '3': 'Complex',
};

// ---------------------------------------------------------------------------
// Helpers
// ---------------------------------------------------------------------------

/** Remove markdown escapes: `\_` → `_`, `\&` → `&`, `\[` → `[`, `\]` → `]` */
function unescapeMarkdown(text: string): string {
  return text
    .replace(/\\_/g, '_')
    .replace(/\\&/g, '&')
    .replace(/\\\[/g, '[')
    .replace(/\\\]/g, ']');
}

/** Replace `<br/>` with newlines and unescape markdown. */
function cleanText(text: string): string {
  return unescapeMarkdown(text.replace(/<br\/>/g, '\n').trim());
}

/**
 * Determine whether an asset is a "master" (origination) or not.
 * Returns `false` for adaptations/reedits/edits.
 */
function determineIsMaster(subCategory: string, assetName: string): boolean {
  if (/adaptation|reedits/i.test(subCategory)) return false;
  if (/\b(?:Adapt|Re-?Edit|Re-?purpose|Edit)\b/i.test(assetName)) return false;
  return true;
}

/**
 * Auto-classify a role name into a discipline by keyword matching.
 * Used only as a fallback for roles not in the base map.
 */
function classifyRole(roleName: string): string {
  const name = roleName.trim().toLowerCase();

  // Tech (check before Creative to avoid "web designer" → Creative)
  if (/web designer/i.test(name)) return 'Tech';
  if (/content manager/i.test(name)) return 'Tech';
  if (/developer/i.test(name)) return 'Tech';
  if (/\bcgi\b/i.test(name)) return 'Tech';

  // QA
  if (/\bqa\b|quality assurance/i.test(name)) return 'QA';

  // Production (check before Editorial to prioritise "motion" / "editor" correctly)
  if (/producer/i.test(name)) return 'Production';
  if (/motion/i.test(name)) return 'Production';

  // Editorial
  if (/copywriter|brand journalist/i.test(name)) return 'Editorial';

  // Creative
  if (/creative director|designer|art director|creator/i.test(name)) return 'Creative';

  // Strategy
  if (/strateg|planner/i.test(name)) return 'Strategy';

  // Delivery
  if (/project manager|programme director/i.test(name)) return 'Delivery';

  // Account Management
  if (/business director|account director|account manager|group account/i.test(name)) return 'Account Management';

  return 'UNKNOWN';
}

// ---------------------------------------------------------------------------
// Main
// ---------------------------------------------------------------------------

const mdPath = path.resolve(__dirname, '../documents/Latest Asset List.md');
const content = fs.readFileSync(mdPath, 'utf-8');
const lines = content.split('\n');

const assets: Asset[] = [];
const staffingRoutes: Record<string, Record<string, number>> = {};
const allRoles = new Set<string>();
const seenUniques = new Set<string>();

// Data rows start at line 5 (1-indexed) → index 4 (0-indexed).
// Lines 1-4 are: title, blank, header row, separator row.
for (let i = 4; i < lines.length; i++) {
  const line = lines[i];
  if (!line || !line.startsWith('|')) continue;

  const cols = line.split('|');

  // Column mapping (after split on |, index 0 is empty before first pipe)
  const rawCatalogId = cols[1]?.trim() ?? '';
  if (!rawCatalogId) continue;

  const catalogId = unescapeMarkdown(rawCatalogId);
  const category = cols[2]?.trim() ?? '';
  const subCategory = cols[3]?.trim() ?? '';
  // cols[4] = sub-category description (not in Asset interface)
  const assetName = unescapeMarkdown(cols[5]?.trim() ?? '');
  const complexityRaw = cols[6]?.trim() ?? '';
  const complexityLevel = COMPLEXITY_MAP[complexityRaw] ?? complexityRaw;
  const description = cleanText(cols[7] ?? '');
  const complexityDescription = cleanText(cols[8] ?? '');
  let studioCaveats = cleanText(cols[9] ?? '');
  if (studioCaveats === 'None') studioCaveats = '';

  const uniques = `${catalogId} - ${assetName} - ${complexityLevel}`;

  if (seenUniques.has(uniques)) {
    console.warn(`  DUPLICATE skipped: ${uniques}`);
    continue;
  }
  seenUniques.add(uniques);

  const isMaster = determineIsMaster(subCategory, assetName);

  assets.push({
    uniques,
    catalogId,
    category,
    subCategory,
    assetName,
    complexityLevel,
    description,
    complexityDescription,
    studioCaveats,
    isMaster,
  });

  // --- Staffing routes ---
  // Role groups start at col index 13, repeating in blocks of 4:
  //   [Role, Location, Hours, <empty separator>]
  const roles: Record<string, number> = {};
  for (let j = 13; j + 2 < cols.length; j += 4) {
    const roleName = cols[j]?.trim() ?? '';
    const location = cols[j + 1]?.trim() ?? '';
    const hoursStr = cols[j + 2]?.trim() ?? '';

    if (!roleName || !hoursStr) continue;
    const hours = parseFloat(hoursStr);
    if (isNaN(hours) || hours === 0) continue;

    // Append trailing space for Oliver+ location to distinguish from Local
    const roleKey = location === 'Oliver+' ? `${roleName} ` : roleName;

    roles[roleKey] = (roles[roleKey] ?? 0) + hours;
    allRoles.add(roleKey);
  }

  if (Object.keys(roles).length > 0) {
    staffingRoutes[uniques] = roles;
  }
}

// ---------------------------------------------------------------------------
// Build merged role → discipline map
// ---------------------------------------------------------------------------

const roleDisciplineMap: Record<string, string> = { ...BASE_ROLE_DISCIPLINE_MAP };
const unknownRoles: string[] = [];

for (const role of allRoles) {
  if (!roleDisciplineMap[role]) {
    const discipline = classifyRole(role);
    roleDisciplineMap[role] = discipline;
    if (discipline === 'UNKNOWN') {
      unknownRoles.push(role);
    } else {
      console.log(`  New role classified: "${role}" → ${discipline}`);
    }
  }
}

// ---------------------------------------------------------------------------
// Write output
// ---------------------------------------------------------------------------

const dataDir = path.resolve(__dirname, '../data');
fs.mkdirSync(dataDir, { recursive: true });

fs.writeFileSync(
  path.join(dataDir, 'assets.json'),
  JSON.stringify(assets, null, 2),
);
fs.writeFileSync(
  path.join(dataDir, 'staffingRoutes.json'),
  JSON.stringify(staffingRoutes, null, 2),
);
fs.writeFileSync(
  path.join(dataDir, 'roleDisciplineMap.json'),
  JSON.stringify(roleDisciplineMap, null, 2),
);

// ---------------------------------------------------------------------------
// Summary
// ---------------------------------------------------------------------------

console.log('\n--- Parse Complete ---');
console.log(`  Assets:          ${assets.length}`);
console.log(`  Staffing routes: ${Object.keys(staffingRoutes).length}`);
console.log(`  Unique roles:    ${allRoles.size}`);
console.log(`  Disciplines:     ${[...new Set(Object.values(roleDisciplineMap))].join(', ')}`);

if (unknownRoles.length > 0) {
  console.warn(`\n  ⚠ UNKNOWN discipline roles (need manual classification):`);
  for (const r of unknownRoles) {
    console.warn(`    - "${r}"`);
  }
}

const categories = [...new Set(assets.map(a => a.category))];
console.log(`\n  Categories (${categories.length}):`);
for (const c of categories) {
  const count = assets.filter(a => a.category === c).length;
  console.log(`    ${c}: ${count} assets`);
}