rag-ultimate

// Maximum number of expansion results to keep in memory. Old entries are
// evicted by insertion order (FIFO approximation) once this cap is reached.
const MAX_EXPANSION_CACHE = 512;
const expansionCache = new Map<string, string[]>();

const CLAUSE_SPLIT_RE = /[,;\/]|\b(?:and|or|vs|versus)\b/gi;

function normalizeQuery(query: string): string {
  return query.trim().replace(/\s+/g, " ");
}

function splitClauses(query: string): string[] {
  return query
    .split(CLAUSE_SPLIT_RE)
    .map(part => part.trim())
    .filter(Boolean);
}

function unique(values: string[]): string[] {
  const seen = new Set<string>();
  const result: string[] = [];
  for (let i = 0; i < values.length; i++) {
    const v = values[i];
    const key = v.toLowerCase();
    if (seen.has(key)) continue;
    seen.add(key);
    result.push(v);
  }
  return result;
}

export function expandQueries(query: string, maxCount: number): string[] {
  const base = normalizeQuery(query);
  const cacheKey = base + "\u0000" + maxCount;

  const cached = expansionCache.get(cacheKey);
  if (cached) return cached;

  const words = base.split(/\W+/).filter(Boolean);
  const focus = words.slice(0, 6).join(" ");
  const clauseVariants = splitClauses(base).slice(0, 2);

  const v0 = words.length > 5 ? words.slice(0, 5).join(" ") : "";
  const v1 = focus && focus !== base ? focus : "";
  const v2 = clauseVariants[0] ? "details about " + clauseVariants[0] : "";
  const v3 = clauseVariants[1] ? "details about " + clauseVariants[1] : "";
  const v4 = words.length > 3 ? "technical details about " + base : "";

  const limited = unique([base, v0, v1, v2, v3, v4]).slice(0, Math.max(1, maxCount));

  if (limited.length === 0) {
    expansionCache.set(cacheKey, [base]);
    return [base];
  }

  if (expansionCache.size >= MAX_EXPANSION_CACHE) {
    const oldest = expansionCache.keys().next().value;
    if (oldest !== undefined) expansionCache.delete(oldest);
  }
  expansionCache.set(cacheKey, limited);

  return limited;
}