Rabin-Karp string search done!

Not too bad to convert Java to Python. Converting the for loop
and playing indexing is interesting, good exercise.

1 files changed, 56 insertions, 0 deletions

diff --git a/sources/hash_substring.py b/sources/hash_substring.py
new file mode 100644
index 0000000..260084a
--- /dev/null
+++ b/sources/hash_substring.py
@@ -0,0 +1,56 @@
+# python3
+import random
+
+
+def read_input():
+    return (input().rstrip(), input().rstrip())
+
+
+def print_occurrences(output):
+    print(' '.join(map(str, output)))
+
+
+def get_occurrences(pattern, text):
+    return [
+        i
+        for i in range(len(text) - len(pattern) + 1)
+        if text[i:i + len(pattern)] == pattern
+    ]
+
+
+def poly_hash(S, p, x):
+    ans = 0
+    for c in reversed(S):
+        ans = (ans * x + ord(c)) % p
+
+    return ans
+
+
+def pre_compute_hashes(T, pattenLength, p, x):
+    H = [0] * (len(T) - pattenLength + 1)
+    S = T[len(T) - pattenLength:]
+    H[len(T) - pattenLength] = poly_hash(S, p, x)
+    y = 1
+    for _ in range(pattenLength):
+        y = (y * x) % p
+    for i in reversed(range(len(T) - pattenLength)):
+        H[i] = (x * H[i + 1] + ord(T[i]) - y * ord(T[i + pattenLength])) % p
+    return H
+
+
+def rabin_karp(P, T):
+    p = 1000000007
+    x = random.randint(1, 1000000006)
+    positions = []
+    pHash = poly_hash(P, p, x)
+    H = pre_compute_hashes(T, len(P), p, x)
+    for i in range(len(T) - len(P) + 1):
+        if pHash != H[i]:
+            continue
+        if T[i:i + len(P)] == P:
+            positions.append(i)
+    return positions
+
+
+if __name__ == '__main__':
+    print_occurrences(rabin_karp(*read_input()))