Skip to content

Instantly share code, notes, and snippets.

@alothings
Created June 14, 2017 00:46
Show Gist options
  • Save alothings/b16b68f1ccf5a31618ce7ec9338a7597 to your computer and use it in GitHub Desktop.
Save alothings/b16b68f1ccf5a31618ce7ec9338a7597 to your computer and use it in GitHub Desktop.
python
"""
Pattern matching problem
Boyer Moore algorithm
First is my attempt, below is the code provided in the book
Idea:
Optimize brute force approach using 2 heuristics:
- Looking-Glass: start searches from last character of the
pattern and work backwards
- Character-Jump: During testing of a pattern P, a mismatch
in T[i] = c with corresponding pattern P[k] is handled:
a) if C is not contained in P, shift P completely past i.
b) if c is contained in P shift P until an occurrence of c
gets aligned with T[i]
"""
def find_boyer_moore(T, P):
""" return lowest index of T at which the substring P begins or -1"""
n, m = len(T), len(P)
if m == 0: return 0
last = {} # Using hash table for fast access
for k in range(m):
last[P[k]] = k
i = m - 1 # i index at T, k index at P
k = m - 1 # j index of last occurrence of T[i] in P
while i < n:
if T[i] == P[k]: # if chars are equal
i -= 1 # THIS
k -= 1 # THIS
if k == 0: # THIS
return i # check if Patter is complete THIS
else:
# if j < k (remember k index at P)
# shift i += m - (j+1)
# if j > k
# shift i += m - k
j = last.get(T[i], -1) # -1 if item not there
i += m - (min(k, j+1))
k = m - 1
return -1
def find_boyer_moore2(T, P):
""" return lowest index of T at which the substring P begins or -1"""
n, m = len(T), len(P)
if m == 0: return 0
last = {} # Using hash table for fast access
for k in range(m):
last[P[k]] = k
i = m - 1 # i index at T, k index at P
k = m - 1 # j index of last occurrence of T[i] in P
while i < n:
if T[i] == P[k]: # if chars are equal
if k == 0:
return i # check if Patter is complete
else:
i -= 1 # normal iteration
k -= 1
else:
# if j < k (remember k index at P)
# shift i += m - (j+1)
# if j > k
# shift i += m - k
j = last.get(T[i], -1) # -1 if item not there
i += m - (min(k, j+1))
k = m - 1
return -1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment