Minimalist implementation of Symmetric Delete Spelling Correction

symspell.py

import re
from collections import Counter, defaultdict

# create lexicon with word frequency from big text
def words(text): return re.findall(r'\w+', text.lower())
WORDS = Counter(words(open('big.txt').read()))

# generate all deletion edits, plus original
def edits(word):
  splits  = [(word[:i], word[i:])    for i in range(len(word) + 1)]
  deletes = [L + R[1:]               for L, R in splits if R]
  return deletes + [word]

# expand lexicon with edited words
EXPANDED = defaultdict(list)
for word in WORDS:
  for deletion in edits(word):
    EXPANDED[deletion].append(word)
 
# compute word frequency
def P(word, N=sum(WORDS.values())): 
    return WORDS[word] / N
 
# match all 1-deletion edits from word to all 1-deletion edits from lexicon
def correction(word):
  candidates = []

  for deletion in edits(word):
    candidates += EXPANDED[deletion]

  return max(set(candidates), key=P)

print('lenght =>', correction('lenght'))