進階字串演算法

def manacher_algorithm(s):
    """Manacher's algorithm to find all palindromes in O(n)"""

    def preprocess(s):
        """Preprocess string to handle even-length palindromes"""
        # Transform "abba" -> "^#a#b#b#a#$"
        result = "^"
        for c in s:
            result += "#" + c
        result += "#$"
        return result

    processed = preprocess(s)
    n = len(processed)
    p = [0] * n  # p[i] = radius of palindrome centered at i
    center = right = 0

    for i in range(1, n - 1):
        # Mirror of i with respect to center
        mirror = 2 * center - i

        # If i is within the right boundary, use previously computed values
        if i < right:
            p[i] = min(right - i, p[mirror])

        # Try to expand palindrome centered at i
        while processed[i + p[i] + 1] == processed[i - p[i] - 1]:
            p[i] += 1

        # If palindrome centered at i extends past right, adjust center and right
        if i + p[i] > right:
            center, right = i, i + p[i]

    return p, processed

def manacher_applications():
    """Various applications of Manacher's algorithm"""

    def longest_palindromic_substring(s):
        """Find longest palindromic substring (LC 5)"""
        if not s:
            return ""

        p, processed = manacher_algorithm(s)
        max_len = 0
        center_index = 0

        for i in range(1, len(p) - 1):
            if p[i] > max_len:
                max_len = p[i]
                center_index = i

        # Convert back to original string coordinates
        start = (center_index - max_len) // 2
        return s[start:start + max_len]

    def count_palindromic_substrings(s):
        """Count all palindromic substrings (LC 647)"""
        if not s:
            return 0

        p, processed = manacher_algorithm(s)
        count = 0

        for i in range(1, len(p) - 1):
            # Each palindrome of radius r contributes (r+1)/2 palindromes
            count += (p[i] + 1) // 2

        return count

    def is_palindrome_range(s, left, right):
        """Check if substring s[left:right+1] is palindrome"""
        p, processed = manacher_algorithm(s)

        # Convert to processed string coordinates
        center = left + right + 2
        radius = right - left

        return p[center] >= radius

    def all_palindromic_substrings(s):
        """Get all palindromic substrings with their positions"""
        p, processed = manacher_algorithm(s)
        palindromes = []

        for i in range(1, len(p) - 1):
            for r in range(p[i] + 1):
                # Convert back to original coordinates
                start = (i - r - 1) // 2
                end = (i + r - 1) // 2
                if start <= end:
                    palindromes.append((start, end, s[start:end + 1]))

        return palindromes

    return {
        'longest': longest_palindromic_substring,
        'count': count_palindromic_substrings,
        'is_palindrome': is_palindrome_range,
        'all_palindromes': all_palindromic_substrings
    }

def z_algorithm(s):
    """Z-algorithm: compute Z array where Z[i] = length of longest substring
    starting from s[i] which is also a prefix of s"""

    n = len(s)
    z = [0] * n
    left = right = 0

    for i in range(1, n):
        if i <= right:
            # Use previously computed values
            z[i] = min(right - i + 1, z[i - left])

        # Try to extend match
        while i + z[i] < n and s[z[i]] == s[i + z[i]]:
            z[i] += 1

        # Update window if we extended past right boundary
        if i + z[i] - 1 > right:
            left, right = i, i + z[i] - 1

    return z

def z_algorithm_applications():
    """Applications of Z-algorithm"""

    def pattern_search_z(text, pattern):
        """Pattern searching using Z-algorithm"""
        combined = pattern + "$" + text
        z = z_algorithm(combined)
        pattern_len = len(pattern)
        matches = []

        for i in range(pattern_len + 1, len(combined)):
            if z[i] == pattern_len:
                matches.append(i - pattern_len - 1)

        return matches

    def find_all_occurrences(s, pattern):
        """Find all occurrences of pattern in string"""
        return pattern_search_z(s, pattern)

    def longest_prefix_suffix(s):
        """Find longest prefix which is also suffix"""
        z = z_algorithm(s)
        n = len(s)

        for i in range(n - 1, 0, -1):
            if z[i] == n - i:
                return s[:n - i]

        return ""

    def period_detection(s):
        """Find the period of string using Z-algorithm"""
        z = z_algorithm(s)
        n = len(s)

        for period in range(1, n):
            if period + z[period] == n:
                return period

        return n  # String has no period shorter than itself

    return {
        'search': pattern_search_z,
        'find_all': find_all_occurrences,
        'prefix_suffix': longest_prefix_suffix,
        'period': period_detection
    }

def suffix_array_construction(s):
    """Construct suffix array using counting sort (O(n log n))"""

    def counting_sort(arr, key_func, max_val):
        """Stable counting sort"""
        count = [0] * (max_val + 1)
        for item in arr:
            count[key_func(item)] += 1

        for i in range(1, len(count)):
            count[i] += count[i - 1]

        result = [0] * len(arr)
        for i in range(len(arr) - 1, -1, -1):
            key = key_func(arr[i])
            count[key] -= 1
            result[count[key]] = arr[i]

        return result

    n = len(s)
    if n == 0:
        return []

    # Initial ranking based on first character
    suffixes = list(range(n))
    rank = [ord(c) for c in s]

    k = 1
    while k < n:
        # Sort by (rank[i], rank[i + k])
        def sort_key(i):
            return (rank[i], rank[i + k] if i + k < n else -1)

        # Find maximum rank for counting sort
        max_rank = max(rank) + 1

        # Sort by second key first, then by first key
        suffixes = counting_sort(suffixes, lambda i: rank[i + k] if i + k < n else -1, max_rank)
        suffixes = counting_sort(suffixes, lambda i: rank[i], max_rank)

        # Update ranks
        new_rank = [0] * n
        for i in range(1, n):
            if (rank[suffixes[i]], rank[suffixes[i] + k] if suffixes[i] + k < n else -1) == \
               (rank[suffixes[i - 1]], rank[suffixes[i - 1] + k] if suffixes[i - 1] + k < n else -1):
                new_rank[suffixes[i]] = new_rank[suffixes[i - 1]]
            else:
                new_rank[suffixes[i]] = i

        rank = new_rank
        k *= 2

    return suffixes

def lcp_array(s, suffix_array):
    """Compute LCP (Longest Common Prefix) array"""
    n = len(s)
    if n == 0:
        return []

    # Inverse of suffix array
    rank = [0] * n
    for i in range(n):
        rank[suffix_array[i]] = i

    lcp = [0] * (n - 1)
    h = 0

    for i in range(n):
        if rank[i] > 0:
            j = suffix_array[rank[i] - 1]
            while i + h < n and j + h < n and s[i + h] == s[j + h]:
                h += 1
            lcp[rank[i] - 1] = h
            if h > 0:
                h -= 1

    return lcp

// java
// LC 65 - Valid Number
// IDEA: hand-built DFA — classify char, follow transition table, accept only in a terminal state
// time = O(N), space = O(1)   (table is a fixed 8x4 constant)
public class ValidNumber {
    // rows = states, cols = char class {0:digit, 1:sign, 2:dot, 3:exp}, -1 = dead
    private static final int[][] DFA = {
        //  d   s   .   e
        {   2,  1,  3, -1 },  // 0 start
        {   2, -1,  3, -1 },  // 1 after leading sign
        {   2, -1,  4,  5 },  // 2 integer digits          (ACCEPT)
        {   4, -1, -1, -1 },  // 3 dot with no digit yet
        {   4, -1, -1,  5 },  // 4 fraction digits         (ACCEPT)
        {   7,  6, -1, -1 },  // 5 just saw 'e' / 'E'
        {   7, -1, -1, -1 },  // 6 sign after 'e'
        {   7, -1, -1, -1 },  // 7 exponent digits         (ACCEPT)
    };

    public boolean isNumber(String s) {
        int state = 0;
        for (char c : s.toCharArray()) {
            int cls;
            if (c >= '0' && c <= '9') cls = 0;
            else if (c == '+' || c == '-') cls = 1;
            else if (c == '.') cls = 2;
            else if (c == 'e' || c == 'E') cls = 3;
            else return false;                 // illegal character
            state = DFA[state][cls];
            if (state == -1) return false;     // illegal transition
        }
        return state == 2 || state == 4 || state == 7;   // accepting states
    }
}

# python
# LC 65 - Valid Number
# IDEA: hand-built DFA — classify char, follow transition table, accept only in a terminal state
# time = O(N), space = O(1)
DFA = [
    #  d   s   .   e
    [  2,  1,  3, -1],   # 0 start
    [  2, -1,  3, -1],   # 1 after leading sign
    [  2, -1,  4,  5],   # 2 integer digits          (ACCEPT)
    [  4, -1, -1, -1],   # 3 dot with no digit yet
    [  4, -1, -1,  5],   # 4 fraction digits         (ACCEPT)
    [  7,  6, -1, -1],   # 5 just saw 'e' / 'E'
    [  7, -1, -1, -1],   # 6 sign after 'e'
    [  7, -1, -1, -1],   # 7 exponent digits         (ACCEPT)
]

def isNumber(s):
    state = 0
    for c in s:
        if c.isdigit():
            cls = 0
        elif c in "+-":
            cls = 1
        elif c == ".":
            cls = 2
        elif c in "eE":
            cls = 3
        else:
            return False              # illegal character
        state = DFA[state][cls]
        if state == -1:               # illegal transition
            return False
    return state in (2, 4, 7)         # accepting states

"53"      0 -d-> 2                     accept (integer)
"53.5"    0 -d-> 2 -.-> 4 -d-> 4       accept (fraction)
"3."      0 -d-> 2 -.-> 4              accept ("digits then dot" is legal)
".9"      0 -.-> 3 -d-> 4              accept (dot first needs a digit after)
"."       0 -.-> 3                     REJECT (state 3 is not accepting)
"53e-9"   0 -d-> 2 -e-> 5 -s-> 6 -d-> 7  accept
"4e+"     0 -d-> 2 -e-> 5 -s-> 6       REJECT (exponent needs a digit)
"99e2.5"  ... state 7 -.-> -1          REJECT (no dot in the exponent)

class AhoCorasick:
    """Aho-Corasick algorithm for multiple pattern matching"""

    def __init__(self):
        self.trie = {}
        self.failure = {}
        self.output = {}

    def add_pattern(self, pattern, pattern_id):
        """Add pattern to trie"""
        node = self.trie
        for char in pattern:
            if char not in node:
                node[char] = {}
            node = node[char]
        node['$'] = pattern_id

    def build_failure_function(self):
        """Build failure function for Aho-Corasick"""
        from collections import deque

        queue = deque()
        self.failure[id(self.trie)] = self.trie

        # Initialize first level
        for char, child in self.trie.items():
            if char != '$':
                self.failure[id(child)] = self.trie
                queue.append((child, char))

        # Build failure links using BFS
        while queue:
            current, char = queue.popleft()
            current_id = id(current)

            for next_char, next_node in current.items():
                if next_char == '$':
                    continue

                queue.append((next_node, next_char))

                # Find failure link
                failure_node = self.failure[current_id]
                while failure_node != self.trie and next_char not in failure_node:
                    failure_node = self.failure[id(failure_node)]

                if next_char in failure_node:
                    self.failure[id(next_node)] = failure_node[next_char]
                else:
                    self.failure[id(next_node)] = self.trie

    def search_all(self, text):
        """Find all pattern occurrences in text"""
        current = self.trie
        results = []

        for i, char in enumerate(text):
            # Follow failure links
            while current != self.trie and char not in current:
                current = self.failure[id(current)]

            if char in current:
                current = current[char]

                # Check for matches
                temp = current
                while temp != self.trie:
                    if '$' in temp:
                        pattern_id = temp['$']
                        results.append((i, pattern_id))
                    temp = self.failure[id(temp)]

        return results

def choose_string_algorithm(problem_characteristics):
    """Guide for choosing optimal string algorithm"""

    if problem_characteristics['type'] == 'pattern_matching':
        if problem_characteristics['single_pattern']:
            return "KMP or Z-Algorithm"
        else:
            return "Aho-Corasick"

    elif problem_characteristics['type'] == 'palindromes':
        if problem_characteristics['all_palindromes']:
            return "Manacher's Algorithm"
        else:
            return "Expand around centers"

    elif problem_characteristics['type'] == 'substring_queries':
        if problem_characteristics['many_queries']:
            return "Rolling Hash or Suffix Array"
        else:
            return "Simple comparison"

    elif problem_characteristics['type'] == 'string_matching':
        if problem_characteristics['approximate']:
            return "Edit distance DP"
        else:
            return "KMP or Rolling Hash"