dmc_transform/builtin/
emoji.rs1use crate::pipeline::Transformer;
4use crate::visit::{NodeAction, Visitor, walk_root};
5use dmc_diagnostic::Code;
6use dmc_diagnostic::metadata::SourceMeta;
7use dmc_parser::ast::*;
8use duck_diagnostic::DiagnosticEngine;
9
10#[derive(Default, Debug)]
11pub struct Emoji;
12
13impl Transformer for Emoji {
14 fn name(&self) -> &str {
15 "emoji"
16 }
17 fn transform(&self, doc: &mut Document, _meta: &SourceMeta, _engine: &mut DiagnosticEngine<Code>) {
18 let mut v = Apply;
19 walk_root(&mut doc.children, &mut v);
20 }
21}
22
23struct Apply;
24
25impl Visitor for Apply {
26 fn visit_node(&mut self, node: &mut Node) -> NodeAction {
27 let Node::Text(t) = node else { return NodeAction::Keep };
28 let Some(replaced) = expand_emoji(&t.value) else { return NodeAction::Keep };
29 t.value = replaced;
30 NodeAction::Keep
31 }
32}
33
34fn expand_emoji(text: &str) -> Option<String> {
38 if !text.contains(':') {
39 return None;
40 }
41 let bytes = text.as_bytes();
42 let mut out = String::with_capacity(text.len());
43 let mut i = 0;
44 let mut found_any = false;
45 while i < bytes.len() {
46 if bytes[i] != b':' {
47 let ch_len = utf8_char_len(bytes[i]);
48 out.push_str(&text[i..i + ch_len]);
49 i += ch_len;
50 continue;
51 }
52 let max_end = (i + 1 + 64).min(bytes.len());
56 let close = (i + 1..max_end).find(|&j| bytes[j] == b':');
57 let Some(close) = close else {
58 out.push(':');
59 i += 1;
60 continue;
61 };
62 let shortcode = &text[i + 1..close];
63 if !is_shortcode(shortcode) {
64 out.push(':');
65 i += 1;
66 continue;
67 }
68 if let Some(emo) = emojis::get_by_shortcode(shortcode) {
69 out.push_str(emo.as_str());
70 i = close + 1;
71 found_any = true;
72 } else {
73 out.push(':');
74 i += 1;
75 }
76 }
77 if found_any { Some(out) } else { None }
78}
79
80fn is_shortcode(s: &str) -> bool {
81 !s.is_empty() && s.bytes().all(|b| b.is_ascii_alphanumeric() || matches!(b, b'_' | b'+' | b'-'))
82}
83
84fn utf8_char_len(b: u8) -> usize {
85 if b < 0x80 {
86 1
87 } else if b < 0xE0 {
88 2
89 } else if b < 0xF0 {
90 3
91 } else {
92 4
93 }
94}
95
96#[cfg(test)]
97mod tests {
98 use super::*;
99
100 #[test]
101 fn passthrough_when_no_colon() {
102 assert!(expand_emoji("hello world").is_none());
103 }
104
105 #[test]
106 fn known_shortcode_expands() {
107 let out = expand_emoji("hi :smile: there").unwrap();
108 assert!(out.contains("hi "));
109 assert!(out.contains(" there"));
110 assert!(!out.contains(":smile:"), "shortcode survived: {out}");
111 }
112
113 #[test]
114 fn unknown_shortcode_is_kept() {
115 assert!(expand_emoji("see :nonexistent_emoji_token: here").is_none());
116 }
117
118 #[test]
119 fn ratio_text_passes_through() {
120 assert!(expand_emoji("ratio 1:2").is_none());
122 }
123
124 #[test]
125 fn multiple_shortcodes() {
126 let out = expand_emoji(":heart: and :star:").unwrap();
127 assert!(!out.contains(':'), "leftover colon: {out}");
128 }
129}