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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489
// This file is part of ICU4X. For terms of use, please see the file
// called LICENSE at the top level of the ICU4X source tree
// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
use zerofrom::ZeroFrom;
use zerovec::{ZeroSlice, ZeroVec};
// Match-node lead unit values, after masking off intermediate-value bits:
// 00..0f: Branch node. If node!=0 then the length is node+1, otherwise
// the length is one more than the next byte.
// For a branch sub-node with at most this many entries, we drop down
// to a linear search.
const MAX_BRANCH_LINEAR_SUB_NODE_LENGTH: usize = 5;
// 0030..003f: Linear-match node, match 1..16 units and continue reading the next node.
const MIN_LINEAR_MATCH: u16 = 0x30;
const MAX_LINEAR_MATCH_LENGTH: u16 = 0x10;
// Match-node lead unit bits 14..6 for the optional intermediate value.
// If these bits are 0, then there is no intermediate value.
// Otherwise, see the *NodeValue* constants below.
const MIN_VALUE_LEAD: u16 = MIN_LINEAR_MATCH + MAX_LINEAR_MATCH_LENGTH; // 0x40
const NODE_TYPE_MASK: u16 = MIN_VALUE_LEAD - 1; // 0x003f
// A final-value node has bit 15 set.
const VALUE_IS_FINAL: u16 = 0x8000;
// Compact value: After testing bit 0, shift right by 15 and then use the following thresholds.
const MAX_ONE_UNIT_VALUE: u16 = 0x3fff;
const MIN_TWO_UNIT_VALUE_LEAD: u16 = MAX_ONE_UNIT_VALUE + 1; // 0x4000
const MAX_ONE_UNIT_NODE_VALUE: u16 = 0xff;
const MIN_TWO_UNIT_NODE_VALUE_LEAD: u16 = MIN_VALUE_LEAD + ((MAX_ONE_UNIT_NODE_VALUE + 1) << 6); // 0x4040
const THREE_UNIT_NODE_VALUE_LEAD: u16 = 0x7fc0;
const THREE_UNIT_VALUE_LEAD: u16 = 0x7fff;
// Compact delta integers.
const MAX_ONE_UNIT_DELTA: u16 = 0xfbff;
const MIN_TWO_UNIT_DELTA_LEAD: u16 = MAX_ONE_UNIT_DELTA + 1; // 0xfc00
const THREE_UNIT_DELTA_LEAD: u16 = 0xffff;
fn skip_value(pos: usize, lead: u16) -> usize {
if lead < MIN_TWO_UNIT_VALUE_LEAD {
pos
} else if lead < THREE_UNIT_VALUE_LEAD {
pos + 1
} else {
pos + 2
}
}
fn skip_node_value(pos: usize, lead: u16) -> usize {
if lead < MIN_TWO_UNIT_NODE_VALUE_LEAD {
pos
} else if lead < THREE_UNIT_NODE_VALUE_LEAD {
pos + 1
} else {
pos + 2
}
}
/// This struct represents a de-serialized `Char16Trie` that was exported from
/// ICU binary data.
///
/// Light-weight, non-const reader class for a `CharsTrie`. Traverses a
/// char-serialized data structure with minimal state, for mapping 16-bit-unit
/// sequences to non-negative integer values.
///
/// For more information:
/// - [ICU4C UCharsTrie](https://unicode-org.github.io/icu-docs/apidoc/released/icu4c/classicu_1_1UCharsTrie.html)
/// - [ICU4J CharsTrie](https://unicode-org.github.io/icu-docs/apidoc/released/icu4j/com/ibm/icu/util/CharsTrie.html) API.
#[cfg_attr(feature = "serde", derive(serde::Deserialize, serde::Serialize))]
#[cfg_attr(feature = "databake", derive(databake::Bake), databake(path = icu_collections::char16trie))]
#[derive(Clone, Debug, PartialEq, Eq, ZeroFrom)]
pub struct Char16Trie<'data> {
/// An array of u16 containing the trie data.
#[cfg_attr(feature = "serde", serde(borrow))]
#[doc(hidden)] // #2417
pub data: ZeroVec<'data, u16>,
}
impl<'data> Char16Trie<'data> {
/// Returns a new [`Char16Trie`] with ownership of the provided data.
pub fn new(data: ZeroVec<'data, u16>) -> Self {
Self { data }
}
/// Returns a new [`Char16TrieIterator`] backed by borrowed data from the `trie` data
pub fn iter(&self) -> Char16TrieIterator {
Char16TrieIterator::new(&self.data)
}
}
/// This struct represents an iterator over a [`Char16Trie`].
#[derive(Clone)]
pub struct Char16TrieIterator<'a> {
/// A reference to the Char16Trie data to iterate over.
trie: &'a ZeroSlice<u16>,
/// Index of next trie unit to read, or `None` if there are no more matches.
pos: Option<usize>,
/// Remaining length of a linear-match node, minus 1, or `None` if not in
/// such a node.
remaining_match_length: Option<usize>,
}
/// An enum representing the return value from a lookup in [`Char16Trie`].
#[derive(Clone, Copy, Debug, PartialEq)]
pub enum TrieResult {
/// The input unit(s) did not continue a matching string.
/// Once `next()` returns `TrieResult::NoMatch`, all further calls to `next()`
/// will also return `TrieResult::NoMatch`.
NoMatch,
/// The input unit(s) matched a string but there is no value for the string
/// so far. (It is a prefix of a longer string.)
NoValue,
/// The input unit(s) continued a matching string and there is a value for
/// the string so far. No further input byte/unit can continue a matching
/// string.
FinalValue(i32),
/// The input unit(s) continued a matching string and there is a value for
/// the string so far. Another input byte/unit can continue a matching
/// string.
Intermediate(i32),
}
// Get the lead surrogate (0xd800..0xdbff) for a
// supplementary code point (0x10000..0x10ffff).
// @param supplementary 32-bit code point (U+10000..U+10ffff)
// @return lead surrogate (U+d800..U+dbff) for supplementary
fn u16_lead(supplementary: i32) -> u16 {
(((supplementary) >> 10) + 0xd7c0) as u16
}
// Get the trail surrogate (0xdc00..0xdfff) for a
// supplementary code point (0x10000..0x10ffff).
// @param supplementary 32-bit code point (U+10000..U+10ffff)
// @return trail surrogate (U+dc00..U+dfff) for supplementary
fn u16_tail(supplementary: i32) -> u16 {
(((supplementary) & 0x3ff) | 0xdc00) as u16
}
/// A macro that takes an `Option` argument and either unwraps it if it has a value or
/// causes the function to return `TrieResult::NoMatch` if there is no value.
/// This could perhaps be done with `std::ops::Try` once stabilized.
macro_rules! trie_unwrap {
($option:expr) => {
match $option {
Some(x) => x,
None => {
// Unexpected
debug_assert!(false);
return TrieResult::NoMatch;
}
}
};
}
impl<'a> Char16TrieIterator<'a> {
/// Returns a new [`Char16TrieIterator`] backed by borrowed data for the `trie` array
pub fn new(trie: &'a ZeroSlice<u16>) -> Self {
Self {
trie,
pos: Some(0),
remaining_match_length: None,
}
}
/// Traverses the trie from the current state for this input char.
///
/// # Examples
///
/// ```
/// use icu::collections::char16trie::{Char16Trie, TrieResult};
/// use zerovec::ZeroVec;
///
/// // A Char16Trie containing the ASCII characters 'a' and 'b'.
/// let trie_data = [48, 97, 176, 98, 32868];
/// let trie = Char16Trie::new(ZeroVec::from_slice_or_alloc(&trie_data));
///
/// let mut iter = trie.iter();
/// let res = iter.next('a');
/// assert_eq!(res, TrieResult::Intermediate(1));
/// let res = iter.next('b');
/// assert_eq!(res, TrieResult::FinalValue(100));
/// let res = iter.next('c');
/// assert_eq!(res, TrieResult::NoMatch);
/// ```
pub fn next(&mut self, c: char) -> TrieResult {
if (c as u32) <= 0xffff {
self.next16(c as u16)
} else {
match self.next16(u16_lead(c as i32)) {
TrieResult::NoValue | TrieResult::Intermediate(_) => {
self.next16(u16_tail(c as i32))
}
_ => TrieResult::NoMatch,
}
}
}
/// Traverses the trie from the current state for this input char.
///
/// # Examples
///
/// ```
/// use icu::collections::char16trie::{Char16Trie, TrieResult};
/// use zerovec::ZeroVec;
///
/// // A Char16Trie containing the ASCII characters 'a' and 'b'.
/// let trie_data = [48, 97, 176, 98, 32868];
/// let trie = Char16Trie::new(ZeroVec::from_slice_or_alloc(&trie_data));
///
/// let mut iter = trie.iter();
/// let res = iter.next('a');
/// assert_eq!(res, TrieResult::Intermediate(1));
/// let res = iter.next('b');
/// assert_eq!(res, TrieResult::FinalValue(100));
/// let res = iter.next('c');
/// assert_eq!(res, TrieResult::NoMatch);
/// ```
pub fn next32(&mut self, c: u32) -> TrieResult {
if c <= 0xffff {
self.next16(c as u16)
} else {
match self.next16(u16_lead(c as i32)) {
TrieResult::NoValue | TrieResult::Intermediate(_) => {
self.next16(u16_tail(c as i32))
}
_ => TrieResult::NoMatch,
}
}
}
/// Traverses the trie from the current state for this input char.
///
/// # Examples
///
/// ```
/// use icu::collections::char16trie::{Char16Trie, TrieResult};
/// use zerovec::ZeroVec;
///
/// // A Char16Trie containing the ASCII characters 'a' and 'b'.
/// let trie_data = [48, 97, 176, 98, 32868];
/// let trie = Char16Trie::new(ZeroVec::from_slice_or_alloc(&trie_data));
///
/// let mut iter = trie.iter();
/// let res = iter.next16('a' as u16);
/// assert_eq!(res, TrieResult::Intermediate(1));
/// let res = iter.next16('b' as u16);
/// assert_eq!(res, TrieResult::FinalValue(100));
/// let res = iter.next16('c' as u16);
/// assert_eq!(res, TrieResult::NoMatch);
/// ```
pub fn next16(&mut self, c: u16) -> TrieResult {
let mut pos = match self.pos {
Some(p) => p,
None => return TrieResult::NoMatch,
};
if let Some(length) = self.remaining_match_length {
// Remaining part of a linear-match node
if c == trie_unwrap!(self.trie.get(pos)) {
pos += 1;
self.pos = Some(pos);
if length == 0 {
self.remaining_match_length = None;
let node = trie_unwrap!(self.trie.get(pos));
if node >= MIN_VALUE_LEAD {
return self.value_result(pos);
}
} else {
self.remaining_match_length = Some(length - 1);
}
return TrieResult::NoValue;
}
self.stop();
TrieResult::NoMatch
} else {
self.next_impl(pos, c)
}
}
fn branch_next(&mut self, pos: usize, length: usize, in_unit: u16) -> TrieResult {
let mut pos = pos;
let mut length = length;
if length == 0 {
length = trie_unwrap!(self.trie.get(pos)) as usize;
pos += 1;
}
length += 1;
// The length of the branch is the number of units to select from.
// The data structure encodes a binary search.
while length > MAX_BRANCH_LINEAR_SUB_NODE_LENGTH {
if in_unit < trie_unwrap!(self.trie.get(pos)) {
length >>= 1;
pos = trie_unwrap!(self.jump_by_delta(pos + 1));
} else {
length = length - (length >> 1);
pos = trie_unwrap!(self.skip_delta(pos + 1));
}
}
// Drop down to linear search for the last few bytes.
// length>=2 because the loop body above sees length>kMaxBranchLinearSubNodeLength>=3
// and divides length by 2.
loop {
if in_unit == trie_unwrap!(self.trie.get(pos)) {
pos += 1;
let mut node = trie_unwrap!(self.trie.get(pos));
if node & VALUE_IS_FINAL != 0 {
self.pos = Some(pos);
return self.value_result(pos);
}
// Use the non-final value as the jump delta.
pos += 1;
if node < MIN_TWO_UNIT_VALUE_LEAD {
pos += node as usize;
} else if node < THREE_UNIT_VALUE_LEAD {
pos += (((node - MIN_TWO_UNIT_VALUE_LEAD) as u32) << 16) as usize
| trie_unwrap!(self.trie.get(pos)) as usize;
pos += 1;
} else {
pos += (trie_unwrap!(self.trie.get(pos)) as usize) << 16
| trie_unwrap!(self.trie.get(pos + 1)) as usize;
pos += 2;
}
node = trie_unwrap!(self.trie.get(pos));
self.pos = Some(pos);
if node >= MIN_VALUE_LEAD {
return self.value_result(pos);
}
return TrieResult::NoValue;
}
length -= 1;
pos = trie_unwrap!(self.skip_value(pos + 1));
if length <= 1 {
break;
}
}
if in_unit == trie_unwrap!(self.trie.get(pos)) {
pos += 1;
self.pos = Some(pos);
let node = trie_unwrap!(self.trie.get(pos));
if node >= MIN_VALUE_LEAD {
return self.value_result(pos);
}
TrieResult::NoValue
} else {
self.stop();
TrieResult::NoMatch
}
}
fn next_impl(&mut self, pos: usize, in_unit: u16) -> TrieResult {
let mut node = trie_unwrap!(self.trie.get(pos));
let mut pos = pos + 1;
loop {
if node < MIN_LINEAR_MATCH {
return self.branch_next(pos, node as usize, in_unit);
} else if node < MIN_VALUE_LEAD {
// Match the first of length+1 units.
let length = node - MIN_LINEAR_MATCH;
if in_unit == trie_unwrap!(self.trie.get(pos)) {
pos += 1;
if length == 0 {
self.remaining_match_length = None;
self.pos = Some(pos);
node = trie_unwrap!(self.trie.get(pos));
if node >= MIN_VALUE_LEAD {
return self.value_result(pos);
}
return TrieResult::NoValue;
}
self.remaining_match_length = Some(length as usize - 1);
self.pos = Some(pos);
return TrieResult::NoValue;
}
// No match
break;
} else if (node & VALUE_IS_FINAL) != 0 {
// No further matching units.
break;
} else {
// Skip intermediate value.
pos = skip_node_value(pos, node);
node &= NODE_TYPE_MASK;
}
}
self.stop();
TrieResult::NoMatch
}
fn stop(&mut self) {
self.pos = None;
}
#[inline(always)] // 1 call site and we want the Option to go away
fn jump_by_delta(&self, pos: usize) -> Option<usize> {
let delta = self.trie.get(pos)?;
let v = if delta < MIN_TWO_UNIT_DELTA_LEAD {
// nothing to do
pos + 1 + delta as usize
} else if delta == THREE_UNIT_DELTA_LEAD {
let delta =
((self.trie.get(pos + 1)? as usize) << 16) | (self.trie.get(pos + 2)? as usize);
pos + delta + 3
} else {
let delta = ((delta - MIN_TWO_UNIT_DELTA_LEAD) as usize) << 16
| (self.trie.get(pos + 1)? as usize);
pos + delta + 2
};
Some(v)
}
#[inline(always)] // 1 call site and we want the Option to go away
fn skip_value(&self, pos: usize) -> Option<usize> {
let lead_unit = self.trie.get(pos)?;
Some(skip_value(pos + 1, lead_unit & 0x7fff))
}
#[inline(always)] // 1 call site and we want the Option to go away
fn skip_delta(&self, pos: usize) -> Option<usize> {
let delta = self.trie.get(pos)?;
let v = if delta < MIN_TWO_UNIT_DELTA_LEAD {
pos + 1
} else if delta == THREE_UNIT_DELTA_LEAD {
pos + 3
} else {
pos + 2
};
Some(v)
}
fn value_result(&self, pos: usize) -> TrieResult {
match self.get_value(pos) {
Some(result) => result,
None => {
// Unexpected
debug_assert!(false);
TrieResult::NoMatch
}
}
}
#[inline(always)] // 1 call site and we want the Option to go away
fn get_value(&self, pos: usize) -> Option<TrieResult> {
let lead_unit = self.trie.get(pos)?;
if lead_unit & VALUE_IS_FINAL == VALUE_IS_FINAL {
self.read_value(pos + 1, lead_unit & 0x7fff)
.map(TrieResult::FinalValue)
} else {
self.read_node_value(pos + 1, lead_unit)
.map(TrieResult::Intermediate)
}
}
#[inline(always)] // 1 call site and we want the Option to go away
fn read_value(&self, pos: usize, lead_unit: u16) -> Option<i32> {
let v = if lead_unit < MIN_TWO_UNIT_VALUE_LEAD {
lead_unit.into()
} else if lead_unit < THREE_UNIT_VALUE_LEAD {
((lead_unit - MIN_TWO_UNIT_VALUE_LEAD) as i32) << 16 | self.trie.get(pos)? as i32
} else {
(self.trie.get(pos)? as i32) << 16 | self.trie.get(pos + 1)? as i32
};
Some(v)
}
#[inline(always)] // 1 call site and we want the Option to go away
fn read_node_value(&self, pos: usize, lead_unit: u16) -> Option<i32> {
let v = if lead_unit < (MIN_TWO_UNIT_NODE_VALUE_LEAD) {
((lead_unit >> 6) - 1).into()
} else if lead_unit < THREE_UNIT_NODE_VALUE_LEAD {
(((lead_unit & 0x7fc0) - MIN_TWO_UNIT_NODE_VALUE_LEAD) as i32) << 10
| self.trie.get(pos)? as i32
} else {
(self.trie.get(pos)? as i32) << 16 | self.trie.get(pos + 1)? as i32
};
Some(v)
}
}