From b05fce7bca0e0e69851dc609eb527956304defcf Mon Sep 17 00:00:00 2001 From: sigoden Date: Wed, 8 Mar 2023 10:40:19 +0800 Subject: refactor: split long paragraphs for smoother stream output (#33) --- src/render/cmd.rs | 223 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 223 insertions(+) (limited to 'src/render/cmd.rs') diff --git a/src/render/cmd.rs b/src/render/cmd.rs index f29fe7d..d349c79 100644 --- a/src/render/cmd.rs +++ b/src/render/cmd.rs @@ -23,6 +23,17 @@ pub fn cmd_render_stream(rx: Receiver, abort: SharedAbortSigna dump(markdown_render.render(&output), 1); } else { buffer = format!("{buffer}{text}"); + if !(markdown_render.is_code_block() + || buffer.len() < 60 + || buffer.starts_with('#') + || buffer.starts_with('>') + || buffer.starts_with('|')) + { + if let Some((output, remain)) = split_line(&buffer) { + dump(markdown_render.render_line_stateless(&output), 0); + buffer = remain + } + } } } ReplyStreamEvent::Done => { @@ -35,3 +46,215 @@ pub fn cmd_render_stream(rx: Receiver, abort: SharedAbortSigna } Ok(()) } + +fn split_line(line: &str) -> Option<(String, String)> { + let mut balance: Vec = Vec::new(); + let chars: Vec = line.chars().collect(); + let mut index = 0; + let len = chars.len(); + while index < len - 1 { + let ch = chars[index]; + if balance.is_empty() + && ((matches!(ch, ',' | '.' | ';') && chars[index + 1].is_whitespace()) + || matches!(ch, '，' | '。' | '；')) + { + let (output, remain) = chars.split_at(index + 1); + return Some((output.iter().collect(), remain.iter().collect())); + } + if index + 2 < len && do_balance(&mut balance, &chars[index..=index + 2]) { + index += 3; + continue; + } + if do_balance(&mut balance, &chars[index..=index + 1]) { + index += 2; + continue; + } + do_balance(&mut balance, &chars[index..index + 1]); + index += 1 + } + + None +} + +#[derive(Debug, Clone, Copy, Eq, PartialEq)] +enum Kind { + ParentheseStart, + ParentheseEnd, + BracketStart, + BracketEnd, + Asterisk, + Asterisk2, + SingleQuota, + DoubleQuota, + Tilde, + Tilde2, + Backtick, + Backtick3, +} + +impl Kind { + fn from_chars(chars: &[char]) -> Option { + let kind = match chars.len() { + 1 => match chars[0] { + '(' => Kind::ParentheseStart, + ')' => Kind::ParentheseEnd, + '[' => Kind::BracketStart, + ']' => Kind::BracketEnd, + '*' => Kind::Asterisk, + '\'' => Kind::SingleQuota, + '"' => Kind::DoubleQuota, + '~' => Kind::Tilde, + '`' => Kind::Backtick, + _ => return None, + }, + 2 if chars[0] == chars[1] => match chars[0] { + '*' => Kind::Asterisk2, + '~' => Kind::Tilde2, + _ => return None, + }, + 3 => { + if chars == ['`', '`', '`'] { + Kind::Backtick3 + } else { + return None; + } + } + _ => return None, + }; + Some(kind) + } +} + +fn do_balance(balance: &mut Vec, chars: &[char]) -> bool { + if let Some(kind) = Kind::from_chars(chars) { + let last = balance.last(); + match (kind, last) { + (Kind::ParentheseStart | Kind::BracketStart, _) => { + balance.push(kind); + true + } + (Kind::ParentheseEnd, Some(&Kind::ParentheseStart)) => { + balance.pop(); + true + } + (Kind::BracketEnd, Some(&Kind::BracketStart)) => { + balance.pop(); + true + } + (Kind::Asterisk, Some(&Kind::Asterisk)) + | (Kind::Asterisk2, Some(&Kind::Asterisk2)) + | (Kind::SingleQuota, Some(&Kind::SingleQuota)) + | (Kind::DoubleQuota, Some(&Kind::DoubleQuota)) + | (Kind::Tilde, Some(&Kind::Tilde)) + | (Kind::Tilde2, Some(&Kind::Tilde2)) + | (Kind::Backtick, Some(&Kind::Backtick)) + | (Kind::Backtick3, Some(&Kind::Backtick3)) => { + balance.pop(); + true + } + (Kind::Asterisk, _) + | (Kind::Asterisk2, _) + | (Kind::SingleQuota, _) + | (Kind::DoubleQuota, _) + | (Kind::Tilde, _) + | (Kind::Tilde2, _) + | (Kind::Backtick, _) + | (Kind::Backtick3, _) => { + balance.push(kind); + true + } + _ => false, + } + } else { + false + } +} + +#[cfg(test)] +mod tests { + use super::*; + + macro_rules! assert_split_line { + ($a:literal, $b:literal, true) => { + assert_eq!( + split_line(&format!("{}{}", $a, $b)), + Some(($a.into(), $b.into())) + ); + }; + ($a:literal, $b:literal, false) => { + assert_eq!(split_line(&format!("{}{}", $a, $b)), None); + }; + } + + #[test] + fn test_split_line() { + assert_split_line!( + "Lorem ipsum dolor sit amet,", + " consectetur adipiscing elit.", + true + ); + assert_split_line!( + "Lorem ipsum dolor sit amet.", + " consectetur adipiscing elit.", + true + ); + assert_split_line!("黃更室幼許刀知，", "波食小午足田世根候法。", true); + assert_split_line!("黃更室幼許刀知。", "波食小午足田世根候法。", true); + assert_split_line!("黃更室幼許刀知；", "波食小午足田世根候法。", true); + assert_split_line!( + "Lorem ipsum (dolor sit amet).", + " consectetur adipiscing elit.", + true + ); + assert_split_line!( + "Lorem ipsum dolor sit `amet,", + " consectetur` adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit ```amet,", + " consectetur``` adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit *amet,", + " consectetur* adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit **amet,", + " consectetur** adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit ~amet,", + " consectetur~ adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit ~~amet,", + " consectetur~~ adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit ``amet,", + " consectetur`` adipiscing elit.", + true + ); + assert_split_line!( + "Lorem ipsum dolor sit \"amet,", + " consectetur\" adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit 'amet,", + " consectetur' adipiscing elit.", + false + ); + assert_split_line!( + "Lorem ipsum dolor sit amet.", + "consectetur adipiscing elit.", + false + ); + } +} -- cgit v1.2.3