story_builder/lib/markdownTableParser.js

const fs = require('fs');

function isTableRowLine(line) {
  const t = line.trim();
  return t.length > 0 && t.includes('|');
}

function isSeparatorLine(line) {
  const t = line.trim();
  if (!t.includes('-')) return false;
  const cells = stripOuterPipes(t).split('|');
  return cells.length > 0 && cells.every((c) => /^:?-{1,}:?$/.test(c.trim()));
}

function stripOuterPipes(t) {
  let s = t;
  if (s.startsWith('|')) s = s.slice(1);
  if (s.endsWith('|')) s = s.slice(0, -1);
  return s;
}

function splitRow(line) {
  return stripOuterPipes(line.trim()).split('|').map((c) => c.trim());
}

function parseRollCell(raw) {
  if (raw == null) return null;
  const t = raw.trim().replace(/\*\*/g, '').trim();
  if (t === '') return null;
  let m = /^(\d+)\s*[-–—]\s*$/.exec(t);
  if (m) return { min: parseInt(m[1], 10), max: 9999 };
  m = /^(\d+)\s*\+\s*$/.exec(t);
  if (m) return { min: parseInt(m[1], 10), max: 9999 };
  m = /^(\d+)\s*or\s*(less|lower|fewer|under)$/i.exec(t);
  if (m) return { min: 0, max: parseInt(m[1], 10) };
  m = /^(\d+)\s*or\s*(more|higher|greater|above)$/i.exec(t);
  if (m) return { min: parseInt(m[1], 10), max: 9999 };
  m = /^(\d+)\s*[-–—]\s*(\d+)$/.exec(t);
  if (m) return { min: parseInt(m[1], 10), max: parseInt(m[2], 10) };
  m = /^(\d+)$/.exec(t);
  if (m) {
    const v = parseInt(m[1], 10);
    return { min: v, max: v };
  }
  return null;
}

function inferDiceFromMax(maxRoll) {
  const standard = [4, 6, 8, 10, 12, 20, 100];
  const found = standard.find((d) => maxRoll <= d);
  return `d${found || maxRoll}`;
}

function buildTableObject(columns, rowsRaw, headingText) {
  if (columns.length === 1) {
    const rows = rowsRaw.map((cells, idx) => ({
      roll_min: idx + 1,
      roll_max: idx + 1,
      cells,
    }));
    return {
      columns,
      rows,
      diceNotation: rows.length ? `d${rows.length}` : null,
      tableType: 'list',
    };
  }

  const headingDiceMatch = /\((\d*d\d+)\)/i.exec(headingText || '');
  const diceFromHeading = headingDiceMatch ? headingDiceMatch[1].toLowerCase() : null;
  const firstColHeader = columns[0];
  const diceFromColumn = /^\d*d\d+$/i.test(firstColHeader.replace(/\s/g, ''))
    ? firstColHeader.toLowerCase()
    : null;

  const firstColRanges = rowsRaw.map((cells) => parseRollCell(cells[0]));
  const allFirstColParsed = firstColRanges.length > 0 && firstColRanges.every((r) => r !== null);

  if (!allFirstColParsed) {
    // First column isn't actually a roll column (or some cells use unrecognized
    // phrasing) - keep it as real data rather than silently discarding it.
    if (firstColRanges.some((r) => r !== null)) {
      console.warn(`"${headingText}": first column has a mix of roll-like and non-roll-like cells; keeping it as a data column.`);
    }
    const rows = rowsRaw.map((cells, idx) => ({ roll_min: idx + 1, roll_max: idx + 1, cells }));
    return {
      columns,
      rows,
      diceNotation: rows.length ? `d${rows.length}` : null,
      tableType: 'list',
    };
  }

  let maxRoll = 0;
  const parsedRows = rowsRaw.map((cells, idx) => {
    const range = firstColRanges[idx];
    if (range.max < 9999) maxRoll = Math.max(maxRoll, range.max);
    return { roll_min: range.min, roll_max: range.max, cells: cells.slice(1) };
  });

  const resultColumns = columns.slice(1);
  const tableType = resultColumns.length > 1 ? 'multi' : 'single';
  let diceNotation = diceFromHeading || diceFromColumn || null;
  if (!diceNotation && maxRoll > 0) {
    diceNotation = inferDiceFromMax(maxRoll);
  }

  return { columns: resultColumns, rows: parsedRows, diceNotation, tableType };
}

function extractFirstTable(bodyLines, headingText) {
  for (let i = 0; i < bodyLines.length - 1; i++) {
    if (isTableRowLine(bodyLines[i]) && isSeparatorLine(bodyLines[i + 1])) {
      const columns = splitRow(bodyLines[i]);
      const rowsRaw = [];
      let j = i + 2;
      while (j < bodyLines.length && isTableRowLine(bodyLines[j]) && !isSeparatorLine(bodyLines[j])) {
        rowsRaw.push(splitRow(bodyLines[j]));
        j++;
      }
      return buildTableObject(columns, rowsRaw, headingText);
    }
  }
  return null;
}

function parseMarkdownFile(filePath) {
  const content = fs.readFileSync(filePath, 'utf8');
  const lines = content.split(/\r?\n/);
  const root = { level: 0, headingText: null, table: null, children: [] };
  const stack = [root];
  let bodyLines = [];
  let currentNode = null;

  function flushBody() {
    if (!currentNode) return;
    const table = extractFirstTable(bodyLines, currentNode.headingText);
    if (table) currentNode.table = table;
  }

  for (const line of lines) {
    const m = /^(#{1,6})\s+(.+?)\s*$/.exec(line);
    if (m) {
      flushBody();
      bodyLines = [];
      const level = m[1].length;
      const headingText = m[2].trim();
      while (stack.length && stack[stack.length - 1].level >= level) stack.pop();
      const parent = stack[stack.length - 1];
      const node = { level, headingText, table: null, children: [] };
      parent.children.push(node);
      stack.push(node);
      currentNode = node;
    } else {
      bodyLines.push(line);
    }
  }
  flushBody();

  return root.children;
}

module.exports = { parseMarkdownFile, parseRollCell };