Function split_paragraphs

nodedb-query/src/chunk_text.rs:216–248

Split text into paragraphs at double-newline (`\n\n`) boundaries.

(text: &str)

/// Split text into paragraphs at double-newline (`\n\n`) boundaries.
fn split_paragraphs(text: &str) -> Vec<(usize, String)> {
    let mut segments = Vec::new();
    let mut char_offset = 0usize;

    // Split on \n\n (or \r\n\r\n).
    let mut remaining = text;
    while let Some(pos) = find_paragraph_break(remaining) {
        let para = &remaining[..pos];
        let para_chars: Vec<char> = para.chars().collect();
        if !para_chars.is_empty() {
            segments.push((char_offset, para.to_string()));
        }
        char_offset += para.chars().count();

        // Skip the paragraph break characters.
        let break_str = &remaining[pos..];
        let break_len = if break_str.starts_with("\r\n\r\n") {
            4
        } else {
            2 // \n\n
        };
        let break_chars = remaining[pos..pos + break_len].chars().count();
        char_offset += break_chars;
        remaining = &remaining[pos + break_len..];
    }

    // Remaining text.
    if !remaining.is_empty() {
        segments.push((char_offset, remaining.to_string()));
    }

    segments
}

/// Find the byte position of the next paragraph break (\n\n or \r\n\r\n).
fn find_paragraph_break(text: &str) -> Option<usize> {