kreuzberg 4.0.0.pre.rc.13 → 4.0.0.pre.rc.14

data/vendor/kreuzberg/src/text/token_reduction/simd_text.rs

@@ -1,147 +1,147 @@
-use memchr::{memchr, memchr3};
-
-pub struct SimdTextProcessor;
-
-impl Default for SimdTextProcessor {
-    fn default() -> Self {
-        Self
-    }
-}
-
-impl SimdTextProcessor {
-    pub fn new() -> Self {
-        Self
-    }
-
-    pub fn clean_punctuation(&self, text: &str) -> String {
-        let bytes = text.as_bytes();
-        let mut result = Vec::with_capacity(text.len());
-        let mut i = 0;
-        let len = bytes.len();
-
-        while i < len {
-            match self.find_next_punctuation(bytes, i) {
-                Some(offset) => {
-                    if offset > 0 {
-                        result.extend_from_slice(&bytes[i..i + offset]);
-                        i += offset;
-                    }
-                }
-                None => {
-                    result.extend_from_slice(&bytes[i..]);
-                    break;
-                }
-            }
-
-            if i >= len {
-                break;
-            }
-
-            let sequence_end = self.find_complex_punctuation_sequence_end(bytes, i);
-            let byte = bytes[i];
-
-            result.push(match byte {
-                b'!' | b'?' => byte,
-                b'.' => b'.',
-                b',' => b',',
-                _ => byte,
-            });
-
-            i = sequence_end;
-        }
-
-        String::from_utf8(result).unwrap_or_else(|_| text.to_string())
-    }
-
-    #[inline]
-    fn is_repeated_punctuation(&self, byte: u8) -> bool {
-        matches!(byte, b'!' | b'?' | b'.' | b',')
-    }
-
-    fn find_complex_punctuation_sequence_end(&self, bytes: &[u8], start: usize) -> usize {
-        let mut i = start + 1;
-
-        while i < bytes.len() && self.is_repeated_punctuation(bytes[i]) {
-            i += 1;
-        }
-
-        i
-    }
-
-    #[inline]
-    fn find_next_punctuation(&self, bytes: &[u8], start: usize) -> Option<usize> {
-        let search = &bytes[start..];
-        let primary = memchr3(b'!', b'?', b'.', search);
-        let comma = memchr(b',', search);
-
-        match (primary, comma) {
-            (Some(p), Some(c)) => Some(p.min(c)),
-            (Some(p), None) => Some(p),
-            (None, Some(c)) => Some(c),
-            (None, None) => None,
-        }
-    }
-}
-
-pub fn chunk_text_for_parallel(text: &str, target_chunks: usize) -> Vec<&str> {
-    if text.len() < 1000 || target_chunks <= 1 {
-        return vec![text];
-    }
-
-    let approximate_chunk_size = text.len() / target_chunks;
-    let mut chunks = Vec::with_capacity(target_chunks);
-    let mut start = 0;
-
-    while start < text.len() {
-        let mut end = (start + approximate_chunk_size).min(text.len());
-
-        if end < text.len() {
-            let search_start = end.saturating_sub(200).max(start);
-            let search_end = (end + 200).min(text.len());
-
-            if let Some(boundary_pos) = memchr3(b'.', b'!', b'?', &text.as_bytes()[search_start..search_end]) {
-                let actual_pos = search_start + boundary_pos + 1;
-                if actual_pos > start && actual_pos < text.len() {
-                    end = actual_pos;
-                }
-            }
-        }
-
-        chunks.push(&text[start..end]);
-        start = end;
-    }
-
-    chunks
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_clean_punctuation() {
-        let processor = SimdTextProcessor::new();
-        let input = "What???!!! Really... Yes,,,";
-        let result = processor.clean_punctuation(input);
-        assert_eq!(result, "What? Really. Yes,");
-    }
-
-    #[test]
-    fn test_clean_punctuation_large_ascii_block() {
-        let processor = SimdTextProcessor::new();
-        let mut input = "a".repeat(1024);
-        input.push_str("!!!");
-        let expected = format!("{}!", "a".repeat(1024));
-
-        let result = processor.clean_punctuation(&input);
-        assert_eq!(result, expected);
-    }
-
-    #[test]
-    fn test_chunk_text() {
-        let text = "This is a test. Another sentence! And one more? Final statement.";
-        let chunks = chunk_text_for_parallel(text, 2);
-        assert!(chunks.len() <= 2);
-        assert_eq!(chunks.join(""), text);
-    }
-}
+use memchr::{memchr, memchr3};
+
+pub struct SimdTextProcessor;
+
+impl Default for SimdTextProcessor {
+    fn default() -> Self {
+        Self
+    }
+}
+
+impl SimdTextProcessor {
+    pub fn new() -> Self {
+        Self
+    }
+
+    pub fn clean_punctuation(&self, text: &str) -> String {
+        let bytes = text.as_bytes();
+        let mut result = Vec::with_capacity(text.len());
+        let mut i = 0;
+        let len = bytes.len();
+
+        while i < len {
+            match self.find_next_punctuation(bytes, i) {
+                Some(offset) => {
+                    if offset > 0 {
+                        result.extend_from_slice(&bytes[i..i + offset]);
+                        i += offset;
+                    }
+                }
+                None => {
+                    result.extend_from_slice(&bytes[i..]);
+                    break;
+                }
+            }
+
+            if i >= len {
+                break;
+            }
+
+            let sequence_end = self.find_complex_punctuation_sequence_end(bytes, i);
+            let byte = bytes[i];
+
+            result.push(match byte {
+                b'!' | b'?' => byte,
+                b'.' => b'.',
+                b',' => b',',
+                _ => byte,
+            });
+
+            i = sequence_end;
+        }
+
+        String::from_utf8(result).unwrap_or_else(|_| text.to_string())
+    }
+
+    #[inline]
+    fn is_repeated_punctuation(&self, byte: u8) -> bool {
+        matches!(byte, b'!' | b'?' | b'.' | b',')
+    }
+
+    fn find_complex_punctuation_sequence_end(&self, bytes: &[u8], start: usize) -> usize {
+        let mut i = start + 1;
+
+        while i < bytes.len() && self.is_repeated_punctuation(bytes[i]) {
+            i += 1;
+        }
+
+        i
+    }
+
+    #[inline]
+    fn find_next_punctuation(&self, bytes: &[u8], start: usize) -> Option<usize> {
+        let search = &bytes[start..];
+        let primary = memchr3(b'!', b'?', b'.', search);
+        let comma = memchr(b',', search);
+
+        match (primary, comma) {
+            (Some(p), Some(c)) => Some(p.min(c)),
+            (Some(p), None) => Some(p),
+            (None, Some(c)) => Some(c),
+            (None, None) => None,
+        }
+    }
+}
+
+pub fn chunk_text_for_parallel(text: &str, target_chunks: usize) -> Vec<&str> {
+    if text.len() < 1000 || target_chunks <= 1 {
+        return vec![text];
+    }
+
+    let approximate_chunk_size = text.len() / target_chunks;
+    let mut chunks = Vec::with_capacity(target_chunks);
+    let mut start = 0;
+
+    while start < text.len() {
+        let mut end = (start + approximate_chunk_size).min(text.len());
+
+        if end < text.len() {
+            let search_start = end.saturating_sub(200).max(start);
+            let search_end = (end + 200).min(text.len());
+
+            if let Some(boundary_pos) = memchr3(b'.', b'!', b'?', &text.as_bytes()[search_start..search_end]) {
+                let actual_pos = search_start + boundary_pos + 1;
+                if actual_pos > start && actual_pos < text.len() {
+                    end = actual_pos;
+                }
+            }
+        }
+
+        chunks.push(&text[start..end]);
+        start = end;
+    }
+
+    chunks
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_clean_punctuation() {
+        let processor = SimdTextProcessor::new();
+        let input = "What???!!! Really... Yes,,,";
+        let result = processor.clean_punctuation(input);
+        assert_eq!(result, "What? Really. Yes,");
+    }
+
+    #[test]
+    fn test_clean_punctuation_large_ascii_block() {
+        let processor = SimdTextProcessor::new();
+        let mut input = "a".repeat(1024);
+        input.push_str("!!!");
+        let expected = format!("{}!", "a".repeat(1024));
+
+        let result = processor.clean_punctuation(&input);
+        assert_eq!(result, expected);
+    }
+
+    #[test]
+    fn test_chunk_text() {
+        let text = "This is a test. Another sentence! And one more? Final statement.";
+        let chunks = chunk_text_for_parallel(text, 2);
+        assert!(chunks.len() <= 2);
+        assert_eq!(chunks.join(""), text);
+    }
+}