diff --git a/trieregex.py b/trieregex.py
index 280515f..90b606f 100644
--- a/trieregex.py
+++ b/trieregex.py
@@ -5,6 +5,7 @@
 
 
 class Memoizer:
+    __slots__ = ['func', 'cache']
 
     def __init__(self, func):
         self.func = func
@@ -26,6 +27,7 @@ def _clear_cache(self):
 
 
 class TrieRegEx():
+    __slots__ = ['_trie', '_initials', '_finals']
 
     def __init__(self, *words: str) -> None:
         self._trie = {}  # type: Dict[str: dict]
@@ -33,27 +35,26 @@ def __init__(self, *words: str) -> None:
         self._finals = defaultdict(int)
         self.add(*words)
     
+    def _adjust_initials_finals(self, word, increase=True):
+        if increase:
+            self._initials[word[0]] += 1
+            self._finals[word[-1]] += 1
+        else:
+            self._initials[word[0]] -= 1
+            self._finals[word[-1]] -= 1
+
     @Memoizer
     def add(self, *words: str) -> None:
         self.regex.clear_cache()  # better performance to clear just once
         for word in words:
             if word != '' and not self.has(word):
-                self.adjust_initials_finals(word)
+                self._adjust_initials_finals(word)
                 trie = self._trie
                 for char in word:
                     if char not in trie:
                         trie[char] = {}
                     trie = trie[char]
                 trie['**'] = {}
-        # self.add.reset()
-
-    def adjust_initials_finals(self, word, increase=True):
-        if increase:
-            self._initials[word[0]] += 1
-            self._finals[word[-1]] += 1
-        else:
-            self._initials[word[0]] -= 1
-            self._finals[word[-1]] -= 1
 
     def remove(self, *words: str) -> None:
         self.add.clear_cache()    # better performance to clear just once
@@ -64,7 +65,7 @@ def remove(self, *words: str) -> None:
                 is_end = i == len(word)
                 if is_end and self.has(word[:i]):
                     remove_word = True
-                    self.adjust_initials_finals(word, increase=False)
+                    self._adjust_initials_finals(word, increase=False)
                 if remove_word:
                     node = self._trie
                     for j in range(i-1):
@@ -113,22 +114,22 @@ def regex(self, trie: dict = None, reset: bool = True) -> str:
             key = list(trie.keys())[0]
             if key == '**':
                 return ''
-            return escape(key) + self.regex(trie[key], False)
+            return f'{escape(key)}{self.regex(trie[key], False)}'
 
         else:
-            sequences = [escape(key) + self.regex(trie[key], False)
+            sequences = [f'{escape(key)}{self.regex(trie[key], False)}'
                          for key in trie if key != '**']
             sequences.sort(key=lambda x: (-len(x), x))  # for easier inspection
 
             if len(sequences) == 1:
                 result = sequences[0]
                 if len(sequences[0]) > 1:
-                    result = '(?:{})'.format(result)
-            elif len(sequences) == len(''.join(sequences)):
-                result = '[{}]'.format(''.join(sequences))
+                    result = f'(?:{result})'
+            elif len(sequences) == len("".join(sequences)):
+                result = f'[{"".join(sequences)}]'
             else:
-                result = '(?:{})'.format('|'.join(sequences))
+                result = f'(?:{"|".join(sequences)})'
 
             if '**' in trie:
-                result = result + '?'
+                return f'{result}?'
             return result