Posted By

magicrebirth on 01/09/11


Tagged

language spelling nlp corrector


Versions (?)

How to Write a Spelling Corrector (in 21 lines)


 / Published in: Python
 

URL: http://norvig.com/spell-correct.html

Requires an external file for the statistical algorithm... (big.txt)

  1. import re, collections
  2.  
  3. def words(text): return re.findall('[a-z]+', text.lower())
  4.  
  5. def train(features):
  6. model = collections.defaultdict(lambda: 1)
  7. for f in features:
  8. model[f] += 1
  9. return model
  10.  
  11. NWORDS = train(words(file('big.txt').read()))
  12.  
  13. alphabet = 'abcdefghijklmnopqrstuvwxyz'
  14.  
  15. def edits1(word):
  16. splits = [(word[:i], word[i:]) for i in range(len(word) + 1)]
  17. deletes = [a + b[1:] for a, b in splits if b]
  18. transposes = [a + b[1] + b[0] + b[2:] for a, b in splits if len(b)>1]
  19. replaces = [a + c + b[1:] for a, b in splits for c in alphabet if b]
  20. inserts = [a + c + b for a, b in splits for c in alphabet]
  21. return set(deletes + transposes + replaces + inserts)
  22.  
  23. def known_edits2(word):
  24. return set(e2 for e1 in edits1(word) for e2 in edits1(e1) if e2 in NWORDS)
  25.  
  26. def known(words): return set(w for w in words if w in NWORDS)
  27.  
  28. def correct(word):
  29. candidates = known([word]) or known(edits1(word)) or known_edits2(word) or [word]
  30. return max(candidates, key=NWORDS.get)

Report this snippet  

You need to login to post a comment.