unobtanium-segmenter 0.5.2

// SPDX-FileCopyrightText: 2026 Slatian
//
// SPDX-License-Identifier: LGPL-3.0-only

use unicode_segmentation::UnicodeSegmentation;

use std::vec::IntoIter;

use crate::SegmentedToken;
use crate::UseOrSubdivide;
use crate::segmentation::Segmenter;

/// Split text into words according to the Unicode definition of what a word is. While not perfect, it should work well enough as an easy starting point.
///
/// Uses [UnicodeSegmentation::split_sentence_bounds] under the hood.
#[derive(Debug, Clone, Default)]
pub struct UnicodeWordSplitter {}

impl UnicodeWordSplitter {
	/// Create a new UnicodeWordSplitter instance
	pub fn new() -> Self {
		Default::default()
	}
}

impl Segmenter for UnicodeWordSplitter {
	type SubdivisionIter<'a> = IntoIter<SegmentedToken<'a>>;

	fn subdivide<'a>(
		&self,
		token: SegmentedToken<'a>,
	) -> UseOrSubdivide<SegmentedToken<'a>, IntoIter<SegmentedToken<'a>>> {
		let word_iterator = token.text.split_word_bounds();
		let mut collection = Vec::new();
		for word in word_iterator {
			collection.push(SegmentedToken::new_derived_from(word, &token));
		}
		if collection.len() > 1 {
			return UseOrSubdivide::Subdivide(collection.into_iter());
		} else {
			return UseOrSubdivide::Use(token);
		}
	}
}

#[cfg(test)]
mod test {

	use super::*;

	use crate::chain::ChainSegmenter;
	use crate::chain::StartSegmentationChain;

	#[test]
	fn test_unicode_word_split() {
		let test_text = "The quick (\"brown\") fox can't jump 32.3 feet, right?\nThe quick (\"brown\")  fox. The value of π in german is '3,141592…'.";

		let result: Vec<&str> = test_text
			.start_segmentation_chain()
			.chain_segmenter(&UnicodeWordSplitter::new())
			.map(|t| t.text)
			.collect();

		let expected_tokens = vec![
			"The", " ", "quick", " ", "(", "\"", "brown", "\"", ")", " ", "fox", " ", "can't", " ",
			"jump", " ", "32.3", " ", "feet", ",", " ", "right", "?", "\n", "The", " ", "quick",
			" ", "(", "\"", "brown", "\"", ")", "  ", "fox", ".", " ", "The", " ", "value", " ",
			"of", " ", "π", " ", "in", " ", "german", " ", "is", " ", "'", "3,141592", "…", "'",
			".",
		];

		assert_eq!(result, expected_tokens);
	}
}