1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
#[derive(Debug, Clone)]
pub struct Token{
pub value: String,
pub synonyms: Vec<String>
}
pub struct Node {
pub chars: [u32; 37],
val: Option<u32>,
}
impl Node{
pub fn new() -> Node {
Node{chars: [u32::MAX;37], val: None}
}
}
pub struct Trie {
pub nodes: Vec<Node>,
pub synonym_dict: Vec<Vec<String>>
}
fn char_2_index(c: u8) -> usize{
let index: usize = match c{
0..=31 => panic!("{} contains non alphanumeric characters", c as char),
32 => 0,
33..=47 => panic!("{} contains non alphanumeric characters", c as char),
48..=57 => c as usize - 47,
58..=96 => panic!("{} contains non alphanumeric characters", c as char),
97..=122 => c as usize - 86,
123..=255 => panic!("{} contains non alphanumeric characters", c as char),
};
return index;
}
impl Trie {
pub fn new() -> Trie {
Trie { nodes: vec![Node::new()], synonym_dict: Vec::new() }
}
fn transition(&self, node_pos: usize, c: &u8) -> usize{
return self.nodes[node_pos].chars[char_2_index(*c)] as usize;
}
pub fn insert(&mut self, string: &str, val: Option<u32>) {
let mut node_pos: usize = 0;
let mut temp_node_pos: usize;
for c in string.as_bytes(){
temp_node_pos = self.transition(node_pos, c);
if temp_node_pos == 4294967295{
self.nodes.push(Node::new());
self.nodes[node_pos].chars[char_2_index(*c)] = (self.nodes.len() - 1) as u32;
node_pos = self.nodes.len() - 1;
}
else {node_pos = temp_node_pos;}
}
if val == Some(4294967295){
self.nodes[node_pos].val = Some(1);
}
else{
self.nodes[node_pos].val = val;
}
}
pub fn insert_synonym(&mut self, string: &str, synonyms: Vec<String>) {
let mut node_pos: usize = 0;
let mut temp_node_pos: usize;
for c in string.as_bytes(){
temp_node_pos = self.transition(node_pos, c);
if temp_node_pos == 4294967295{
self.nodes.push(Node::new());
self.nodes[node_pos].chars[char_2_index(*c)] = (self.nodes.len() - 1) as u32;
node_pos = self.nodes.len() - 1;
}
else {node_pos = temp_node_pos;}
}
self.synonym_dict.push(synonyms);
self.nodes[node_pos].val = Some(4000000000+self.synonym_dict.len() as u32 - 1);
}
pub fn search(&self, string: &str) -> Option<u32> {
let mut node_pos: usize = 0;
for c in string.as_bytes(){
node_pos = self.transition(node_pos, c);
if node_pos == 4294967295 {return None;}
}
return self.nodes[node_pos].val;
}
pub fn longest_common_prefix_search(&self, string: &str) -> Option<(u32, usize)>{
let mut node_pos: usize = 0;
let mut value: Option<u32> = None;
let mut offset: usize = 0;
for i in 0..string.len(){
node_pos = self.transition(node_pos, &string.as_bytes()[i]);
if node_pos != 4294967295 && i < string.len()-1{
if string.as_bytes()[i+1] == ' ' as u8{
value = self.nodes[node_pos].val;
offset = i;
}
}
if i == string.len()-1 && node_pos != 4294967295{
return Some((self.nodes[node_pos].val.unwrap_or(1), i));
}
if node_pos == 4294967295 {
match value {
None => {
return None;
},
Some(val) => {
return Some((val, offset));
}
}
}
}
return Some((value.unwrap_or(0), offset));
}
pub fn common_prefix_search(&self, string: &str) -> Vec<(u32, usize)> {
let mut node_pos: usize = 0;
let mut results: Vec<(u32, usize)> = Vec::new();
for i in 0..string.len(){
node_pos = self.transition(node_pos, &string.as_bytes()[i]);
if node_pos != 4294967295 && self.nodes[node_pos].val != None{
if i == string.len()-1{
results.push((self.nodes[node_pos].val.unwrap(), i));
}
else {
if string.as_bytes()[i+1] == ' ' as u8{
results.push((self.nodes[node_pos].val.unwrap(), i));
}
}
}
if node_pos == 4294967295 {
return results;
}
}
return results;
}
pub fn get_all_tokens(&self, input: &str)-> Vec<Token>{
if !input.contains(' '){return vec![Token {value: input.to_string(), synonyms: match self.search(input){
None => Vec::new(),
Some(val) => {
if val >= 4000000000{
self.synonym_dict[val as usize - 4000000000].clone()
}
else{Vec::new()}
}
}
}
]}
let mut candidates: Vec<(String, usize, usize, u32)> = Vec::new();
let mut offsets: Vec<usize> = input.match_indices(' ').map(|s| s.0+1).collect();
offsets.push(0);
offsets.rotate_right(1);
for offset in offsets{
let word = self.longest_common_prefix_search(&input[offset..]).unwrap_or((0, 0));
let candidate = (input[offset..=offset+word.1].to_string(), offset, offset+word.1,word.0);
if candidate.2 > candidates.last().unwrap_or(&("".to_string(), 0, 0, 0)).2 && word.1!=0{
candidates.push(candidate);
}
}
let mut offset: usize = 0;
let mut to_add: Vec<(String, usize, usize, u32)> = Vec::new();
if candidates.len() == 0 {
return self.split_candidate((input.to_string(), 0, 0 ,0));
}
if candidates.len() == 1{
for candidate in &candidates{
if &input[offset..candidate.1] != "" && &input[offset..candidate.1] != " "{
if offset == 0{
to_add.push((input[offset..candidate.1-1].to_string(), offset, candidate.1-2, 0))
}
else{
to_add.push((input[offset+1..candidate.1-1].to_string(), offset+1, candidate.1-2, 0))
}
}
offset = candidate.2+1;
}
if &input[offset..input.len()] != "" && &input[offset..input.len()] != " "{
to_add.push((input[offset+1..input.len()].to_string(), offset+1, input.len(), 0))
}
candidates.append(&mut to_add);
candidates.sort_by_key(|a| a.1);
let mut result: Vec<Token> = Vec::new();
for candidate in candidates{
if candidate.3 == 0{
let mut splits = self.split_candidate(candidate);
result.append(&mut splits);
}
else{
if candidate.3 >= 4000000000{
result.push(Token{value: candidate.0, synonyms: self.synonym_dict[candidate.3 as usize -4000000000].clone()})
}
else{
result.push(Token{value: candidate.0, synonyms: Vec::new()});
}
}
}
return result;
}
if candidates.len() == 2{
if candidates[0].3 >= candidates[1].3 && candidates[0].2 >= candidates[1].1{
candidates.remove(1);
}
else{
if candidates[0].3 < candidates[1].3 && candidates[0].2 >= candidates[1].1{
candidates.remove(0);
}
}
}
if candidates.len()>=3{
let mut tokens_to_remove: Vec<usize> = Vec::new();
let windows_iter = candidates.windows(3);
for (i, window) in windows_iter.enumerate(){
if window[1].3 < window[0].3 + window[2].3 && window[0].2 >= window[1].1 && window[1].2 >= window[2].1 {
tokens_to_remove.push(i+1);
}
else{
if window[0].3 >= window[1].3 && window[0].2 >= window[1].1{
tokens_to_remove.push(i+1);
}
if window[1].3 > window[0].3 && window[0].2 >= window[1].1{
tokens_to_remove.push(i);
}
if window[1].3 >= window[2].3 && window[1].2 >= window[2].1{
tokens_to_remove.push(i+2);
}
}
}
tokens_to_remove.dedup();
for index in tokens_to_remove.iter().rev(){
candidates.remove(*index);
}
if candidates.len()>=2{
if candidates[candidates.len()-1].1 < candidates[candidates.len()-2].2{
if candidates[candidates.len()-2].3 >= candidates[candidates.len()-1].3{
candidates.remove(candidates.len()-1);
}
else{
candidates.remove(candidates.len()-2);
}
}
}
}
for candidate in &candidates{
if &input[offset..candidate.1] != "" && &input[offset..candidate.1] != " "{
if offset == 0{
to_add.push((input[offset..candidate.1].to_string(), offset, candidate.1-2, 0))
}
else{
to_add.push((input[offset..candidate.1].to_string(), offset+1, candidate.1-2, 0))
}
}
offset = candidate.2+1;
}
if &input[offset..input.len()] != "" && &input[offset..input.len()] != " "{
to_add.push((input[offset+1..input.len()].to_string(), offset+1, input.len(), 0))
}
candidates.append(&mut to_add);
candidates.sort_by_key(|a| a.1);
let mut result: Vec<Token> = Vec::new();
for candidate in candidates{
if candidate.3 == 0{
let mut splits = self.split_candidate(candidate);
result.append(&mut splits);
}
else{
if candidate.3 >= 4000000000{
result.push(Token{value: candidate.0, synonyms: self.synonym_dict[candidate.3 as usize -4000000000].clone()})
}
else{
result.push(Token{value: candidate.0, synonyms: Vec::new()});
}
}
}
return result;
}
fn split_candidate(&self, candidate: (String, usize, usize, u32)) -> Vec<Token>{
match self.search(&candidate.0){
Some(val)=> {
if val >= 4000000000 {
return vec![Token{value: candidate.0, synonyms: self.synonym_dict[val as usize -4000000000].clone()}];
}
else{
return vec![Token{value: candidate.0, synonyms: Vec::new()}];
}
}
None => {
let splits: Vec<Token> = candidate.0.split_whitespace().map(|s| Token{value: s.to_owned(), synonyms: Vec::new()}).collect();
return splits;
}
}
}
}