fixed failure function and cleaned up code in kmp + added rabin-karp

This commit is contained in:
damelLP 2018-01-07 12:49:51 +00:00
parent 495fdc1ff9
commit 0d36dc60c5
2 changed files with 88 additions and 15 deletions

View File

@ -1,4 +1,4 @@
def kmp(pattern, text, len_p=None, len_t=None):
def kmp(pattern, text):
"""
The Knuth-Morris-Pratt Algorithm for finding a pattern within a piece of text
with complexity O(n + m)
@ -14,14 +14,7 @@ def kmp(pattern, text, len_p=None, len_t=None):
"""
# 1) Construct the failure array
failure = [0]
i = 0
for index, char in enumerate(pattern[1:]):
if pattern[i] == char:
i += 1
else:
i = 0
failure.append(i)
failure = get_failure_array(pattern)
# 2) Step through text searching for pattern
i, j = 0, 0 # index into text, pattern
@ -29,20 +22,38 @@ def kmp(pattern, text, len_p=None, len_t=None):
if pattern[j] == text[i]:
if j == (len(pattern) - 1):
return True
i += 1
j += 1
# if this is a prefix in our pattern
# just go back far enough to continue
elif failure[j] > 0:
j = failure[j] - 1
else:
i += 1
elif j > 0:
j = failure[j - 1]
continue
i += 1
return False
if __name__ == '__main__':
def get_failure_array(pattern):
"""
Calculates the new index we should go to if we fail a comparison
:param pattern:
:return:
"""
failure = [0]
i = 0
j = 1
while j < len(pattern):
if pattern[i] == pattern[j]:
i += 1
elif i > 0:
i = failure[i-1]
continue
j += 1
failure.append(i)
return failure
if __name__ == '__main__':
# Test 1)
pattern = "abc1abc12"
text1 = "alskfjaldsabc1abc1abc12k23adsfabcabc"
@ -54,4 +65,16 @@ if __name__ == '__main__':
text = "ABABZABABYABABX"
assert kmp(pattern, text)
# Test 3)
pattern = "AAAB"
text = "ABAAAAAB"
assert kmp(pattern, text)
# Test 4)
pattern = "abcdabcy"
text = "abcxabcdabxabcdabcdabcy"
assert kmp(pattern, text)
# Test 5)
pattern = "aabaabaaa"
assert get_failure_array(pattern) == [0, 1, 0, 1, 2, 3, 4, 5, 2]

50
strings/rabin-karp.py Normal file
View File

@ -0,0 +1,50 @@
def rabin_karp(pattern, text):
"""
The Rabin-Karp Algorithm for finding a pattern within a piece of text
with complexity O(nm), most efficient when it is used with multiple patterns
as it is able to check if any of a set of patterns match a section of text in o(1) given the precomputed hashes.
This will be the simple version which only assumes one pattern is being searched for but it's not hard to modify
1) Calculate pattern hash
2) Step through the text one character at a time passing a window with the same length as the pattern
calculating the hash of the text within the window compare it with the hash of the pattern. Only testing
equality if the hashes match
"""
p_len = len(pattern)
p_hash = hash(pattern)
for i in range(0, len(text) - (p_len - 1)):
# written like this t
text_hash = hash(text[i:i + p_len])
if text_hash == p_hash and \
text[i:i + p_len] == pattern:
return True
return False
if __name__ == '__main__':
# Test 1)
pattern = "abc1abc12"
text1 = "alskfjaldsabc1abc1abc12k23adsfabcabc"
text2 = "alskfjaldsk23adsfabcabc"
assert rabin_karp(pattern, text1) and not rabin_karp(pattern, text2)
# Test 2)
pattern = "ABABX"
text = "ABABZABABYABABX"
assert rabin_karp(pattern, text)
# Test 3)
pattern = "AAAB"
text = "ABAAAAAB"
assert rabin_karp(pattern, text)
# Test 4)
pattern = "abcdabcy"
text = "abcxabcdabxabcdabcdabcy"
assert rabin_karp(pattern, text)