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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843
use std::{
cmp::Ordering,
fmt,
ops::{Range, RangeFrom, RangeFull, RangeInclusive, RangeTo, RangeToInclusive},
};
use ahash::AHashMap;
use itertools::Itertools;
use thiserror::Error;
mod characters;
#[cfg(feature = "tokenizers")]
mod huggingface;
#[cfg(feature = "rust-tokenizers")]
mod rust_tokenizers;
#[cfg(feature = "tiktoken-rs")]
mod tiktoken;
use crate::trim::Trim;
pub use characters::Characters;
/// Indicates there was an error with the chunk capacity configuration.
/// The `Display` implementation will provide a human-readable error message to
/// help debug the issue that caused the error.
#[derive(Error, Debug)]
#[error(transparent)]
pub struct ChunkCapacityError(#[from] ChunkCapacityErrorRepr);
/// Private error and free to change across minor version of the crate.
#[derive(Error, Debug)]
enum ChunkCapacityErrorRepr {
#[error("Max chunk size must be greater than or equal to the desired chunk size")]
MaxLessThanDesired,
}
/// Describes the valid chunk size(s) that can be generated.
///
/// The `desired` size is the target size for the chunk. In most cases, this
/// will also serve as the maximum size of the chunk. It is always possible
/// that a chunk may be returned that is less than the `desired` value, as
/// adding the next piece of text may have made it larger than the `desired`
/// capacity.
///
/// The `max` size is the maximum possible chunk size that can be generated.
/// By setting this to a larger value than `desired`, it means that the chunk
/// should be as close to `desired` as possible, but can be larger if it means
/// staying at a larger semantic level.
///
/// The splitter will consume text until at maxumum somewhere between `desired`
/// and `max`, if they differ, but never above `max`.
///
/// If you need to ensure a fixed size, set `desired` and `max` to the same
/// value. For example, if you are trying to maximize the context window for an
/// embedding.
///
/// If you are loosely targeting a size, but have some extra room, for example
/// in a RAG use case where you roughly want a certain part of a document, you
/// can set `max` to your absolute maxumum, and the splitter can stay at a
/// higher semantic level when determining the chunk.
#[derive(Copy, Clone, Debug, PartialEq)]
pub struct ChunkCapacity {
desired: usize,
max: usize,
}
impl ChunkCapacity {
/// Create a new `ChunkCapacity` with the same `desired` and `max` size.
#[must_use]
pub fn new(size: usize) -> Self {
Self {
desired: size,
max: size,
}
}
/// The `desired` size is the target size for the chunk. In most cases, this
/// will also serve as the maximum size of the chunk. It is always possible
/// that a chunk may be returned that is less than the `desired` value, as
/// adding the next piece of text may have made it larger than the `desired`
/// capacity.
#[must_use]
pub fn desired(&self) -> usize {
self.desired
}
/// The `max` size is the maximum possible chunk size that can be generated.
/// By setting this to a larger value than `desired`, it means that the chunk
/// should be as close to `desired` as possible, but can be larger if it means
/// staying at a larger semantic level.
#[must_use]
pub fn max(&self) -> usize {
self.max
}
/// If you need to ensure a fixed size, set `desired` and `max` to the same
/// value. For example, if you are trying to maximize the context window for an
/// embedding.
///
/// If you are loosely targeting a size, but have some extra room, for example
/// in a RAG use case where you roughly want a certain part of a document, you
/// can set `max` to your absolute maxumum, and the splitter can stay at a
/// higher semantic level when determining the chunk.
///
/// # Errors
///
/// If the `max` size is less than the `desired` size, an error is returned.
pub fn with_max(mut self, max: usize) -> Result<Self, ChunkCapacityError> {
if max < self.desired {
Err(ChunkCapacityError(
ChunkCapacityErrorRepr::MaxLessThanDesired,
))
} else {
self.max = max;
Ok(self)
}
}
/// Validate if a given chunk fits within the capacity
///
/// - `Ordering::Less` indicates more could be added
/// - `Ordering::Equal` indicates the chunk is within the capacity range
/// - `Ordering::Greater` indicates the chunk is larger than the capacity
#[must_use]
pub fn fits(&self, chunk_size: usize) -> Ordering {
if chunk_size < self.desired {
Ordering::Less
} else if chunk_size > self.max {
Ordering::Greater
} else {
Ordering::Equal
}
}
}
impl From<usize> for ChunkCapacity {
fn from(size: usize) -> Self {
ChunkCapacity::new(size)
}
}
impl From<Range<usize>> for ChunkCapacity {
fn from(range: Range<usize>) -> Self {
ChunkCapacity::new(range.start)
.with_max(range.end.saturating_sub(1).max(range.start))
.expect("invalid range")
}
}
impl From<RangeFrom<usize>> for ChunkCapacity {
fn from(range: RangeFrom<usize>) -> Self {
ChunkCapacity::new(range.start)
.with_max(usize::MAX)
.expect("invalid range")
}
}
impl From<RangeFull> for ChunkCapacity {
fn from(_: RangeFull) -> Self {
ChunkCapacity::new(usize::MIN)
.with_max(usize::MAX)
.expect("invalid range")
}
}
impl From<RangeInclusive<usize>> for ChunkCapacity {
fn from(range: RangeInclusive<usize>) -> Self {
ChunkCapacity::new(*range.start())
.with_max(*range.end())
.expect("invalid range")
}
}
impl From<RangeTo<usize>> for ChunkCapacity {
fn from(range: RangeTo<usize>) -> Self {
ChunkCapacity::new(usize::MIN)
.with_max(range.end.saturating_sub(1))
.expect("invalid range")
}
}
impl From<RangeToInclusive<usize>> for ChunkCapacity {
fn from(range: RangeToInclusive<usize>) -> Self {
ChunkCapacity::new(usize::MIN)
.with_max(range.end)
.expect("invalid range")
}
}
/// Result returned from a `ChunkSizer`. Includes the size of the chunk, in units
/// determined by the sizer, as well as the max byte offset of the text that
/// would fit within the given `ChunkCapacity`.
#[derive(Copy, Clone, Debug, PartialEq)]
pub struct ChunkSize {
/// Whether or not the entire chunk fits within the `ChunkCapacity`
fits: Ordering,
/// max byte offset of the text that fit within the given `ChunkCapacity`.
max_chunk_size_offset: Option<usize>,
/// Size of the chunk, in units used by the sizer.
size: usize,
}
impl ChunkSize {
/// Generate a chunk size from a given size. Will not be able to compute the
/// max byte offset that fits within the capacity.
#[must_use]
pub fn from_size(size: usize, capacity: &ChunkCapacity) -> Self {
Self {
fits: capacity.fits(size),
max_chunk_size_offset: None,
size,
}
}
/// Generate a chunk size from an iterator of byte ranges for each encoded
/// element in the chunk.
pub fn from_offsets(
offsets: impl Iterator<Item = Range<usize>>,
capacity: &ChunkCapacity,
) -> Self {
let mut chunk_size = offsets.fold(
Self {
fits: Ordering::Less,
max_chunk_size_offset: None,
size: 0,
},
|mut acc, range| {
acc.size += 1;
if acc.size <= capacity.max {
acc.max_chunk_size_offset = Some(range.end);
}
acc
},
);
chunk_size.fits = capacity.fits(chunk_size.size);
chunk_size
}
/// Determine whether the chunk size fits within the capacity or not
#[must_use]
pub fn fits(&self) -> Ordering {
self.fits
}
/// Size of the chunk, in units used by the sizer.
#[must_use]
pub fn size(&self) -> usize {
self.size
}
}
/// Determines the size of a given chunk.
pub trait ChunkSizer {
/// Determine the size of a given chunk to use for validation
fn chunk_size(&self, chunk: &str, capacity: &ChunkCapacity) -> ChunkSize;
}
/// Indicates there was an error with the chunk configuration.
/// The `Display` implementation will provide a human-readable error message to
/// help debug the issue that caused the error.
#[derive(Error, Debug)]
#[error(transparent)]
pub struct ChunkConfigError(#[from] ChunkConfigErrorRepr);
/// Private error and free to change across minor version of the crate.
#[derive(Error, Debug)]
enum ChunkConfigErrorRepr {
#[error("The overlap is larger than or equal to the desired chunk capacity")]
OverlapLargerThanCapacity,
}
/// Configuration for how chunks should be created
#[derive(Debug)]
pub struct ChunkConfig<Sizer>
where
Sizer: ChunkSizer,
{
/// The chunk capacity to use for filling chunks
capacity: ChunkCapacity,
/// The amount of overlap between chunks. Defaults to 0.
overlap: usize,
/// The chunk sizer to use for determining the size of each chunk
sizer: Sizer,
/// Whether whitespace will be trimmed from the beginning and end of each chunk
trim: bool,
}
impl ChunkConfig<Characters> {
/// Create a basic configuration for chunking with only the required value a chunk capacity.
///
/// By default, chunk sizes will be calculated based on the number of characters in each chunk.
/// You can set a custom chunk sizer by calling [`Self::with_sizer`].
///
/// By default, chunks will be trimmed. If you want to preserve whitespace,
/// call [`Self::with_trim`] and set it to `false`.
#[must_use]
pub fn new(capacity: impl Into<ChunkCapacity>) -> Self {
Self {
capacity: capacity.into(),
overlap: 0,
sizer: Characters,
trim: true,
}
}
}
impl<Sizer> ChunkConfig<Sizer>
where
Sizer: ChunkSizer,
{
/// Retrieve a reference to the chunk capacity for this configuration.
pub fn capacity(&self) -> &ChunkCapacity {
&self.capacity
}
/// Retrieve the amount of overlap between chunks.
pub fn overlap(&self) -> usize {
self.overlap
}
/// Set the amount of overlap between chunks.
///
/// # Errors
///
/// Will return an error if the overlap is larger than or equal to the chunk capacity.
pub fn with_overlap(mut self, overlap: usize) -> Result<Self, ChunkConfigError> {
if overlap >= self.capacity.desired {
Err(ChunkConfigError(
ChunkConfigErrorRepr::OverlapLargerThanCapacity,
))
} else {
self.overlap = overlap;
Ok(self)
}
}
/// Retrieve a reference to the chunk sizer for this configuration.
pub fn sizer(&self) -> &Sizer {
&self.sizer
}
/// Set a custom chunk sizer to use for determining the size of each chunk
///
/// ```
/// use text_splitter::{Characters, ChunkConfig};
///
/// let config = ChunkConfig::new(512).with_sizer(Characters);
/// ```
#[must_use]
pub fn with_sizer<S: ChunkSizer>(self, sizer: S) -> ChunkConfig<S> {
ChunkConfig {
capacity: self.capacity,
overlap: self.overlap,
sizer,
trim: self.trim,
}
}
/// Whether chunkd should have whitespace trimmed from the beginning and end or not.
pub fn trim(&self) -> bool {
self.trim
}
/// Specify whether chunks should have whitespace trimmed from the
/// beginning and end or not.
///
/// If `false` (default), joining all chunks should return the original
/// string.
/// If `true`, all chunks will have whitespace removed from beginning and end.
///
/// ```
/// use text_splitter::ChunkConfig;
///
/// let config = ChunkConfig::new(512).with_trim(false);
/// ```
#[must_use]
pub fn with_trim(mut self, trim: bool) -> Self {
self.trim = trim;
self
}
}
impl<T> From<T> for ChunkConfig<Characters>
where
T: Into<ChunkCapacity>,
{
fn from(capacity: T) -> Self {
Self::new(capacity)
}
}
/// A memoized chunk sizer that caches the size of chunks.
/// Very helpful when the same chunk is being validated multiple times, which
/// happens often, and can be expensive to compute, such as with tokenizers.
#[derive(Debug)]
pub struct MemoizedChunkSizer<'sizer, Sizer>
where
Sizer: ChunkSizer,
{
/// Cache of chunk sizes per byte offset range for base capacity
capacity_cache: AHashMap<Range<usize>, ChunkSize>,
/// The configuration for the chunk sizer
chunk_config: &'sizer ChunkConfig<Sizer>,
/// Cache of chunk sizes per byte offset range for overlap ranges
overlap_cache: AHashMap<Range<usize>, ChunkSize>,
/// Semantic level, used for determining trimming behavior
trim: Trim,
}
impl<'sizer, Sizer> MemoizedChunkSizer<'sizer, Sizer>
where
Sizer: ChunkSizer,
{
/// Wrap any chunk sizer for memoization
pub fn new(chunk_config: &'sizer ChunkConfig<Sizer>, trim: Trim) -> Self {
Self {
capacity_cache: AHashMap::new(),
chunk_config,
overlap_cache: AHashMap::new(),
trim,
}
}
/// Determine the size of a given chunk to use for validation,
/// returning a cached value if it exists, and storing the result if not.
fn chunk_size(&mut self, offset: usize, chunk: &str, is_overlap: bool) -> ChunkSize {
let cache = if is_overlap {
&mut self.overlap_cache
} else {
&mut self.capacity_cache
};
let capacity = if is_overlap {
self.chunk_config.overlap.into()
} else {
self.chunk_config.capacity
};
*cache
.entry(offset..(offset + chunk.len()))
.or_insert_with(|| self.chunk_config.sizer.chunk_size(chunk, &capacity))
}
/// Check if the chunk is within the capacity. Chunk should be trimmed if necessary beforehand.
pub fn check_capacity(&mut self, offset: usize, chunk: &str, is_overlap: bool) -> ChunkSize {
let (offset, chunk) = self.trim_chunk(offset, chunk);
let mut chunk_size = self.chunk_size(offset, chunk, is_overlap);
if let Some(max_chunk_size_offset) = chunk_size.max_chunk_size_offset.as_mut() {
*max_chunk_size_offset += offset;
}
chunk_size
}
/// If trim chunks is on, trim the str and adjust the offset
pub fn trim_chunk<'text>(&self, offset: usize, chunk: &'text str) -> (usize, &'text str) {
if self.chunk_config.trim {
self.trim.trim(offset, chunk)
} else {
(offset, chunk)
}
}
/// Find the best level to start splitting the text
pub fn find_correct_level<'text, L: fmt::Debug>(
&mut self,
offset: usize,
levels_with_first_chunk: impl Iterator<Item = (L, &'text str)>,
) -> (Option<L>, Option<usize>) {
// If we aren't at the highest semantic level, stop iterating sections that go beyond the range of the next level.
let mut max_encoded_offset = None;
let mut semantic_level = None;
// We assume that larger levels are also longer. We can skip lower levels if going to a higher level would result in a shorter text
let levels_with_first_chunk =
levels_with_first_chunk.coalesce(|(a_level, a_str), (b_level, b_str)| {
if a_str.len() >= b_str.len() {
Ok((b_level, b_str))
} else {
Err(((a_level, a_str), (b_level, b_str)))
}
});
for (level, str) in levels_with_first_chunk {
let chunk_size = self.check_capacity(offset, str, false);
// If this no longer fits, we use the level we are at.
if chunk_size.fits.is_gt() {
max_encoded_offset = chunk_size.max_chunk_size_offset;
break;
}
// Otherwise break up the text with the next level
semantic_level = Some(level);
}
(semantic_level, max_encoded_offset)
}
/// Clear the cached values. Once we've moved the cursor,
/// we don't need to keep the old values around.
pub fn clear_cache(&mut self) {
self.capacity_cache.clear();
self.overlap_cache.clear();
}
}
#[cfg(test)]
mod tests {
use std::sync::atomic::{self, AtomicUsize};
use crate::trim::Trim;
use super::*;
#[test]
fn check_chunk_capacity() {
let chunk = "12345";
assert_eq!(
Characters.chunk_size(chunk, &4.into()).fits,
Ordering::Greater
);
assert_eq!(
Characters.chunk_size(chunk, &5.into()).fits,
Ordering::Equal
);
assert_eq!(Characters.chunk_size(chunk, &6.into()).fits, Ordering::Less);
}
#[test]
fn check_chunk_capacity_for_range() {
let chunk = "12345";
assert_eq!(
Characters.chunk_size(chunk, &(0..0).into()).fits,
Ordering::Greater
);
assert_eq!(
Characters.chunk_size(chunk, &(0..5).into()).fits,
Ordering::Greater
);
assert_eq!(
Characters.chunk_size(chunk, &(5..6).into()).fits,
Ordering::Equal
);
assert_eq!(
Characters.chunk_size(chunk, &(6..100).into()).fits,
Ordering::Less
);
}
#[test]
fn check_chunk_capacity_for_range_from() {
let chunk = "12345";
assert_eq!(
Characters.chunk_size(chunk, &(0..).into()).fits,
Ordering::Equal
);
assert_eq!(
Characters.chunk_size(chunk, &(5..).into()).fits,
Ordering::Equal
);
assert_eq!(
Characters.chunk_size(chunk, &(6..).into()).fits,
Ordering::Less
);
}
#[test]
fn check_chunk_capacity_for_range_full() {
let chunk = "12345";
assert_eq!(
Characters.chunk_size(chunk, &(..).into()).fits,
Ordering::Equal
);
}
#[test]
fn check_chunk_capacity_for_range_inclusive() {
let chunk = "12345";
assert_eq!(
Characters.chunk_size(chunk, &(0..=4).into()).fits,
Ordering::Greater
);
assert_eq!(
Characters.chunk_size(chunk, &(5..=6).into()).fits,
Ordering::Equal
);
assert_eq!(
Characters.chunk_size(chunk, &(4..=5).into()).fits,
Ordering::Equal
);
assert_eq!(
Characters.chunk_size(chunk, &(6..=100).into()).fits,
Ordering::Less
);
}
#[test]
fn check_chunk_capacity_for_range_to() {
let chunk = "12345";
assert_eq!(
Characters.chunk_size(chunk, &(..0).into()).fits,
Ordering::Greater
);
assert_eq!(
Characters.chunk_size(chunk, &(..5).into()).fits,
Ordering::Greater
);
assert_eq!(
Characters.chunk_size(chunk, &(..6).into()).fits,
Ordering::Equal
);
}
#[test]
fn check_chunk_capacity_for_range_to_inclusive() {
let chunk = "12345";
assert_eq!(
Characters.chunk_size(chunk, &(..=4).into()).fits,
Ordering::Greater
);
assert_eq!(
Characters.chunk_size(chunk, &(..=5).into()).fits,
Ordering::Equal
);
assert_eq!(
Characters.chunk_size(chunk, &(..=6).into()).fits,
Ordering::Equal
);
}
#[test]
fn chunk_size_from_offsets() {
let offsets = [0..1, 1..2, 2..3];
let chunk_size = ChunkSize::from_offsets(offsets.clone().into_iter(), &1.into());
assert_eq!(
ChunkSize {
fits: Ordering::Greater,
size: offsets.len(),
max_chunk_size_offset: Some(1)
},
chunk_size
);
}
#[test]
fn chunk_size_from_empty_offsets() {
let offsets = [];
let chunk_size = ChunkSize::from_offsets(offsets.clone().into_iter(), &1.into());
assert_eq!(
ChunkSize {
fits: Ordering::Less,
size: offsets.len(),
max_chunk_size_offset: None
},
chunk_size
);
}
#[test]
fn chunk_size_from_small_offsets() {
let offsets = [0..1, 1..2, 2..3];
let chunk_size = ChunkSize::from_offsets(offsets.clone().into_iter(), &4.into());
assert_eq!(
ChunkSize {
fits: Ordering::Less,
size: offsets.len(),
max_chunk_size_offset: Some(3)
},
chunk_size
);
}
#[derive(Default)]
struct CountingSizer {
calls: AtomicUsize,
}
impl ChunkSizer for CountingSizer {
// Return character version, but count calls
fn chunk_size(&self, chunk: &str, capacity: &ChunkCapacity) -> ChunkSize {
self.calls.fetch_add(1, atomic::Ordering::SeqCst);
Characters.chunk_size(chunk, capacity)
}
}
#[test]
fn memoized_sizer_only_calculates_once_per_text() {
let chunk_config = ChunkConfig::new(10).with_sizer(CountingSizer::default());
let mut memoized_sizer = MemoizedChunkSizer::new(&chunk_config, Trim::All);
let text = "1234567890";
for _ in 0..10 {
memoized_sizer.chunk_size(0, text, false);
}
assert_eq!(
memoized_sizer
.chunk_config
.sizer()
.calls
.load(atomic::Ordering::SeqCst),
1
);
}
#[test]
fn memoized_sizer_calculates_once_per_different_text() {
let chunk_config = ChunkConfig::new(10).with_sizer(CountingSizer::default());
let mut memoized_sizer = MemoizedChunkSizer::new(&chunk_config, Trim::All);
let text = "1234567890";
for i in 0..10 {
memoized_sizer.chunk_size(0, text.get(0..i).unwrap(), false);
}
assert_eq!(
memoized_sizer
.chunk_config
.sizer()
.calls
.load(atomic::Ordering::SeqCst),
10
);
}
#[test]
fn can_clear_cache_on_memoized_sizer() {
let chunk_config = ChunkConfig::new(10).with_sizer(CountingSizer::default());
let mut memoized_sizer = MemoizedChunkSizer::new(&chunk_config, Trim::All);
let text = "1234567890";
for _ in 0..10 {
memoized_sizer.chunk_size(0, text, false);
memoized_sizer.clear_cache();
}
assert_eq!(
memoized_sizer
.chunk_config
.sizer()
.calls
.load(atomic::Ordering::SeqCst),
10
);
}
#[test]
fn test_chunk_size_from_size() {
let chunk_size = ChunkSize::from_size(10, &10.into());
assert_eq!(
ChunkSize {
fits: Ordering::Equal,
size: 10,
max_chunk_size_offset: None
},
chunk_size
);
}
#[test]
fn basic_chunk_config() {
let config = ChunkConfig::new(10);
assert_eq!(config.capacity, 10.into());
assert_eq!(config.sizer, Characters);
assert!(config.trim());
}
#[test]
fn disable_trimming() {
let config = ChunkConfig::new(10).with_trim(false);
assert!(!config.trim());
}
#[test]
fn new_sizer() {
#[derive(Debug, PartialEq)]
struct BasicSizer;
impl ChunkSizer for BasicSizer {
fn chunk_size(&self, _chunk: &str, _capacity: &ChunkCapacity) -> ChunkSize {
unimplemented!()
}
}
let config = ChunkConfig::new(10).with_sizer(BasicSizer);
assert_eq!(config.capacity, 10.into());
assert_eq!(config.sizer, BasicSizer);
assert!(config.trim());
}
#[test]
fn chunk_capacity_max_and_desired_equal() {
let capacity = ChunkCapacity::new(10);
assert_eq!(capacity.desired(), 10);
assert_eq!(capacity.max(), 10);
}
#[test]
fn chunk_capacity_can_adjust_max() {
let capacity = ChunkCapacity::new(10).with_max(20).unwrap();
assert_eq!(capacity.desired(), 10);
assert_eq!(capacity.max(), 20);
}
#[test]
fn chunk_capacity_max_cant_be_less_than_desired() {
let capacity = ChunkCapacity::new(10);
let err = capacity.with_max(5).unwrap_err();
assert_eq!(
err.to_string(),
"Max chunk size must be greater than or equal to the desired chunk size"
);
assert_eq!(capacity.desired(), 10);
assert_eq!(capacity.max(), 10);
}
#[test]
fn set_chunk_overlap() {
let config = ChunkConfig::new(10).with_overlap(5).unwrap();
assert_eq!(config.overlap(), 5);
}
#[test]
fn cant_set_overlap_larger_than_capacity() {
let chunk_config = ChunkConfig::new(5);
let err = chunk_config.with_overlap(10).unwrap_err();
assert_eq!(
err.to_string(),
"The overlap is larger than or equal to the desired chunk capacity"
);
}
#[test]
fn cant_set_overlap_larger_than_desired() {
let chunk_config = ChunkConfig::new(5..15);
let err = chunk_config.with_overlap(10).unwrap_err();
assert_eq!(
err.to_string(),
"The overlap is larger than or equal to the desired chunk capacity"
);
}
}