Using native string methods to see if it is faster

candidate_cache.py

@@ -7,17 +7,11 @@ def make_key(dictionary_name, letters):
 
 
 def candidates(letters, dictionary: Dictionary, min=2, permutations=None):
-    letter_counter = Counter(letters)
+    letters = "".join(letters)
     possibilities = []
-    for word in dictionary.word_frequency.keys():
-        word_counter = Counter(word)
-        try:
-            add = all(
-                word_counter[key] <= letter_counter[key] for key in word_counter.keys()
-            )
-        except KeyError:
-            add = False
-        if add:
+    for word in filter(lambda word: len(word) > min, dictionary.word_frequency.keys()):
+        letter_set = set(word)
+        if all(word.count(letter) < letters.count(letter) for letter in letter_set):
             possibilities.append(word)
     return possibilities