|
| 1 | +// In computer science, a suffix tree (also called PAT tree or, in an earlier form, position tree) |
| 2 | +// is a compressed trie containing all the suffixes of the given text as their keys and positions |
| 3 | +// in the text as their values. Suffix trees allow particularly fast implementations of many |
| 4 | +// important string operations. Source: https://en.wikipedia.org/wiki/Suffix_tree |
| 5 | + |
| 6 | +#[derive(Debug, PartialEq, Eq, Clone)] |
| 7 | +pub struct Node { |
| 8 | + pub sub: String, // substring of input string |
| 9 | + pub ch: Vec<usize>, // vector of child nodes |
| 10 | +} |
| 11 | + |
| 12 | +impl Node { |
| 13 | + fn new(sub: String, children: Vec<usize>) -> Self { |
| 14 | + Node { |
| 15 | + sub, |
| 16 | + ch: children.to_vec(), |
| 17 | + } |
| 18 | + } |
| 19 | + pub fn empty() -> Self { |
| 20 | + Node { |
| 21 | + sub: "".to_string(), |
| 22 | + ch: vec![], |
| 23 | + } |
| 24 | + } |
| 25 | +} |
| 26 | + |
| 27 | +pub struct SuffixTree { |
| 28 | + pub nodes: Vec<Node>, |
| 29 | +} |
| 30 | + |
| 31 | +impl SuffixTree { |
| 32 | + pub fn new(s: String) -> Self { |
| 33 | + let mut suf_tree = SuffixTree { |
| 34 | + nodes: vec![Node::empty()], |
| 35 | + }; |
| 36 | + for i in 0..s.len() { |
| 37 | + let (_, substr) = s.split_at(i); |
| 38 | + suf_tree.add_suffix(substr); |
| 39 | + } |
| 40 | + suf_tree |
| 41 | + } |
| 42 | + fn add_suffix(&mut self, suf: &str) { |
| 43 | + let mut n = 0; |
| 44 | + let mut i = 0; |
| 45 | + while i < suf.len() { |
| 46 | + let b = suf.chars().nth(i); |
| 47 | + let mut x2 = 0; |
| 48 | + let mut n2: usize; |
| 49 | + loop { |
| 50 | + let children = &self.nodes[n].ch; |
| 51 | + if children.len() == x2 { |
| 52 | + n2 = self.nodes.len(); |
| 53 | + self.nodes.push(Node::new( |
| 54 | + { |
| 55 | + let (_, sub) = suf.split_at(i); |
| 56 | + sub.to_string() |
| 57 | + }, |
| 58 | + vec![], |
| 59 | + )); |
| 60 | + self.nodes[n].ch.push(n2); |
| 61 | + return; |
| 62 | + } |
| 63 | + n2 = children[x2]; |
| 64 | + if self.nodes[n2].sub.chars().next() == b { |
| 65 | + break; |
| 66 | + } |
| 67 | + x2 += 1; |
| 68 | + } |
| 69 | + let sub2 = self.nodes[n2].sub.clone(); |
| 70 | + let mut j = 0; |
| 71 | + while j < sub2.len() { |
| 72 | + if suf.chars().nth(i + j) != sub2.chars().nth(j) { |
| 73 | + let n3 = n2; |
| 74 | + n2 = self.nodes.len(); |
| 75 | + self.nodes.push(Node::new( |
| 76 | + { |
| 77 | + let (sub, _) = sub2.split_at(j); |
| 78 | + sub.to_string() |
| 79 | + }, |
| 80 | + vec![n3], |
| 81 | + )); |
| 82 | + let (_, temp_sub) = sub2.split_at(j); |
| 83 | + self.nodes[n3].sub = temp_sub.to_string(); |
| 84 | + self.nodes[n].ch[x2] = n2; |
| 85 | + break; |
| 86 | + } |
| 87 | + j += 1; |
| 88 | + } |
| 89 | + i += j; |
| 90 | + n = n2; |
| 91 | + } |
| 92 | + } |
| 93 | +} |
| 94 | + |
| 95 | +#[cfg(test)] |
| 96 | +mod tests { |
| 97 | + use super::*; |
| 98 | + |
| 99 | + #[test] |
| 100 | + fn test_suffix_tree() { |
| 101 | + let suf_tree = SuffixTree::new("banana$".to_string()); |
| 102 | + assert_eq!( |
| 103 | + suf_tree.nodes, |
| 104 | + vec![ |
| 105 | + Node { |
| 106 | + sub: "".to_string(), |
| 107 | + ch: vec![1, 8, 6, 10] |
| 108 | + }, |
| 109 | + Node { |
| 110 | + sub: "banana$".to_string(), |
| 111 | + ch: vec![] |
| 112 | + }, |
| 113 | + Node { |
| 114 | + sub: "na$".to_string(), |
| 115 | + ch: vec![] |
| 116 | + }, |
| 117 | + Node { |
| 118 | + sub: "na$".to_string(), |
| 119 | + ch: vec![] |
| 120 | + }, |
| 121 | + Node { |
| 122 | + sub: "na".to_string(), |
| 123 | + ch: vec![2, 5] |
| 124 | + }, |
| 125 | + Node { |
| 126 | + sub: "$".to_string(), |
| 127 | + ch: vec![] |
| 128 | + }, |
| 129 | + Node { |
| 130 | + sub: "na".to_string(), |
| 131 | + ch: vec![3, 7] |
| 132 | + }, |
| 133 | + Node { |
| 134 | + sub: "$".to_string(), |
| 135 | + ch: vec![] |
| 136 | + }, |
| 137 | + Node { |
| 138 | + sub: "a".to_string(), |
| 139 | + ch: vec![4, 9] |
| 140 | + }, |
| 141 | + Node { |
| 142 | + sub: "$".to_string(), |
| 143 | + ch: vec![] |
| 144 | + }, |
| 145 | + Node { |
| 146 | + sub: "$".to_string(), |
| 147 | + ch: vec![] |
| 148 | + } |
| 149 | + ] |
| 150 | + ); |
| 151 | + } |
| 152 | +} |
0 commit comments