import { readFile } from "node:fs/promises";

export type ArchiveContents = {
  ids: Set<string>;
  byExtractor: Map<string, Set<string>>;
};

// yt-dlp archive lines have the form `<extractor> <id>` (e.g., `youtube abc123`,
// `lbry 9720133ae3a206487321f21eb5bdcc29f40828fc`).
export function parseArchive(text: string): ArchiveContents {
  const ids = new Set<string>();
  const byExtractor = new Map<string, Set<string>>();
  for (const rawLine of text.split("\n")) {
    const line = rawLine.trim();
    if (!line) continue;
    const spaceIdx = line.indexOf(" ");
    if (spaceIdx <= 0) continue;
    const extractor = line.slice(0, spaceIdx);
    const id = line.slice(spaceIdx + 1).trim();
    if (!id) continue;
    ids.add(id);
    let set = byExtractor.get(extractor);
    if (!set) {
      set = new Set<string>();
      byExtractor.set(extractor, set);
    }
    set.add(id);
  }
  return { ids, byExtractor };
}

export async function readArchive(filePath: string): Promise<ArchiveContents> {
  let text: string;
  try {
    text = await readFile(filePath, "utf8");
  } catch {
    return { ids: new Set(), byExtractor: new Map() };
  }
  return parseArchive(text);
}