Prepare for addition of cipher types outside of current cipher groups

This commit future proofs the code for the addition of new cipher types outside of the currently supported cipher groups (e.g. 'aca' and 'rotor'). It therefore:

- Improves detection of rotor ciphers in provided `--ciphers` argument
- Ensures `CipherStatisticsDataset` works with missing cipher types
- Future proofs selection of cipher group 'all' in `expand_cipher_groups`

(Squashed merge of PR #14.)

Author: MaikBastian

cipherTypeDetection/cipherStatisticsDataset.py

@@ -362,7 +362,7 @@ def _rearrange_ciphertexts(self, ciphertext_label_pairs):
     @property
     def is_initialized(self):
         """Indicates whether the dataset was initialized with input data."""
-        return self._number_of_cipher_classes != 0
+        return self._number_of_cipher_classes != 0 and self._batch_size > 0
 
     def _are_files_closed(self):
         """Returns True if at least one ciphertext file is closed."""
@@ -452,7 +452,7 @@ def __init__(self, plaintext_paths, dataset_params, logger):
 
         key_lengths_count = 0
         for cipher_t in self._cipher_types:
-            index = self._cipher_types.index(cipher_t)
+            index = config.CIPHER_TYPES.index(cipher_t)
             if isinstance(config.KEY_LENGTHS[index], list):
                 key_lengths_count += len(config.KEY_LENGTHS[index])
             else:
@@ -472,7 +472,7 @@ def key_lengths_count(self):
     @property
     def is_initialized(self):
         """Indicates whether the dataset was initialized with input data."""
-        return len(self._plaintext_paths) != 0 and len(self._cipher_types) != 0
+        return len(self._plaintext_paths) != 0 and len(self._cipher_types) != 0 and self._batch_size > 0
 
     def __iter__(self):
         return self
@@ -610,15 +610,14 @@ def perform(self, plaintexts):
 
         for line in plaintexts:
             for cipher_type in self._cipher_types:
-                index = config.cipher_types.index(cipher_type)
-                label = self._cipher_types.index(cipher_type)
+                label = config.cipher_types.index(cipher_type)
                 if isinstance(config.key_lengths[label], list):
                     key_lengths = config.key_lengths[label]
                 else:
                     key_lengths = [config.key_lengths[label]]
                 for key_length in key_lengths:
                     try:
-                        ciphertext = encrypt(line, index, key_length, self._keep_unknown_symbols)
+                        ciphertext = encrypt(line, label, key_length, self._keep_unknown_symbols)
                     except:
                         multiprocessing_logger.error(f"Could not encrypt line with cipher "
                                                      f"'{cipher_type}'. and key length {key_length}. "

cipherTypeDetection/train.py

@@ -485,8 +485,7 @@ def load_plaintext_datasets_from_disk(args, requested_cipher_types, batch_size):
         `CipherStatisticsDataset`s for training and testing.
     """
     # Filter cipher_types to exclude non-plaintext ciphers
-    aca_cipher_types = [config.CIPHER_TYPES[i] for i in range(56)]
-    cipher_types = [type for type in requested_cipher_types if type in aca_cipher_types]
+    cipher_types = [type for type in requested_cipher_types if type not in config.ROTOR_CIPHER_TYPES]
 
     # Get all paths to plaintext files in the `plaintext_input_directory`
     plaintext_files = []
@@ -1031,8 +1030,7 @@ def expand_cipher_groups(cipher_types):
             expanded.append(config.CIPHER_TYPES[i])
     elif config.ALL in expanded:
         del expanded[expanded.index(config.ALL)]
-        for i in range(61):
-            expanded.append(config.CIPHER_TYPES[i])
+        expanded.extend(config.CIPHER_TYPES)
     return expanded
 
 def main():