torrents.csv/scripts/scan_torrent_files.js

const fs = require('fs'),
  path = require('path'),
  readTorrent = require('read-torrent'),
  argv = require('minimist')(process.argv.slice(2));

var torrentFiles = {};
var torrentCsvHashes = new Set();

var jsonFile = '../torrent_files.json';
var torrentsCsvFile = '../torrents.csv';
console.log(`Scanning torrent files from ${argv.dir} into ${jsonFile} ...`);
main();

async function main() {
  await fillTorrentFiles();
  await fillTorrentCsvHashes();
  await scanFolder();
  writeFile();
}


async function fillTorrentFiles() {
  if (fs.existsSync(jsonFile)) {
    var fileContents = await fs.promises.readFile(jsonFile, 'utf8');
    torrentFiles = JSON.parse(fileContents);
  }
}

async function fillTorrentCsvHashes() {
  var fileContents = await fs.promises.readFile(torrentsCsvFile, 'utf8');
  var lines = fileContents.split('\n');
  for (const line of lines) {
    var hash = line.split(';')[0];
    torrentCsvHashes.add(hash);
  }
  torrentCsvHashes.delete('infohash');
}

async function scanFolder() {
  console.log('Scanning dir: ' + argv.dir + '...');
  var fileHashes = new Set(Object.keys(torrentFiles));

  var files = fs.readdirSync(argv.dir).filter(f => {
    var sp = f.split('.');
    var ext = sp[1];
    var hash = sp[0];
    var fullPath = argv.dir + '/' + f;
    // It must be a torrent file,
    // NOT in the torrent_files.json
    // must be in the CSV file
    // must have a file size
    return (ext == 'torrent' &&
      !fileHashes.has(hash) &&
      torrentCsvHashes.has(hash) &&
      getFilesizeInBytes(fullPath) > 0);
  });
  for (const file of files) {
    var fullPath = argv.dir + '/' + file;
    console.log(`Scanning File ${fullPath}`);
    var torrent = await read(fullPath).catch(e => console.log('Read error'));
    torrentFiles = { ...torrentFiles, ...torrent }; // concat them
  };
  console.log('Done scanning.');
}

function getFilesizeInBytes(filename) {
    var stats = fs.statSync(filename);
    var fileSizeInBytes = stats["size"];
    return fileSizeInBytes;
}

function writeFile() {
  torrentFiles = Object.keys(torrentFiles)
    .sort()
    .filter(hash => torrentCsvHashes.has(hash))
    .reduce((r, k) => (r[k] = torrentFiles[k], r), {});
  fs.writeFileSync(jsonFile, "{\n");
  var first = true;
  for (let [key, value] of Object.entries(torrentFiles)) {
    if(first) first = false;
    else fs.appendFileSync(jsonFile, ",\n");
    fs.appendFileSync(jsonFile, `${JSON.stringify(key)}:${JSON.stringify(value)}`);
  }
  fs.appendFileSync(jsonFile, "\n}");
  console.log(`${jsonFile} written.`);
  process.exit();
}

function read(uri, options) {
  return new Promise((resolve, reject) => {
    readTorrent(uri, (err, info) => {
      if (!err) {
        // Removing some extra fields from files
        if (info.files) {
          info.files.forEach((f, i) => {
            f.i = i;
            f.p = f.path;
            f.l = f.length;
            delete f.name;
            delete f.offset;
            delete f.path;
            delete f.length;
          });
        }

        resolve({ [info.infoHash]: info.files });
      } else {
        console.error('Error in read-torrent: ' + err.message + ' for torrent uri: ' + uri);
        reject(err);
      }
    });
  });
}