search method

List<(int, double)> search(String query, { int limit = 10, })

Returns up to limit hits as (documentIndex, score), sorted by score descending. Each record is (index, score) with index the document index and score the TF similarity.

Implementation

List<(int, double)> search(String query, {int limit = 10}) {
  final Map<String, int> qTf = textToTf(query);
  final List<(int, double)> results = List.filled(_tfs.length, (0, 0.0));
  int count = 0;
  for (final MapEntry<int, Map<String, int>> entry in _tfs.asMap().entries) {
    final double score = cosineSimilarity(qTf, entry.value);
    if (score > 0) {
      results[count++] = (entry.key, score);
    }
  }
  final List<(int, double)> trimmed = results.sublist(0, count);
  trimmed.sort((a, b) {
    final (_, aScore) = a;
    final (_, bScore) = b;
    return bScore.compareTo(aScore);
  });
  return trimmed.take(limit).toList();
}