Only add emojis to dict

Riey · Jul 30, 2024 · ab8e05b · ab8e05b
1 parent cce2b6c
commit ab8e05b
Show file tree

Hide file tree

Showing 4 changed files with 321 additions and 3 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/src/engine/core/tests/emoji.rs b/src/engine/core/tests/emoji.rs
@@ -12,7 +12,7 @@ fn thinking() {
     test_input(&[
         (EMOJI, "🏻(light skin tone)🏼(medium-light skin tone)🏽(medium skin tone)🏾(medium-dark skin tone)🏿(dark skin tone)", ""),
         (Key::normal(T), "t🏻(light skin tone)🏼(medium-light skin tone)🏽(medium skin tone)🏾(medium-dark skin tone)🏿(dark skin tone)", ""),
-        (Key::normal(H), "th😁(beaming face with smiling eyes)😂(face with tears of joy)🤣(rolling on the floor laughing)😃(grinning face with big eyes)😄(grinning face with smiling eyes)", ""),
+        (Key::normal(H), "th😃(grinning face with big eyes)😄(grinning face with smiling eyes)😁(beaming face with smiling eyes)😅(grinning face with sweat)🤣(rolling on the floor laughing)", ""),
         (Key::normal(I), "thi🤔(thinking face)🕧(twelve-thirty)🕜(one-thirty)🕝(two-thirty)🕞(three-thirty)", ""),
         (Key::normal(N), "thin🤔(thinking face)", ""),
         (Key::normal(K), "think🤔(thinking face)", ""),

diff --git a/src/engine/dict/Cargo.toml b/src/engine/dict/Cargo.toml
@@ -8,5 +8,6 @@ license = "GPL-3.0-or-later"
 [build-dependencies]
 serde = {version = "1.0.118", features = ["derive"]}
 serde_json = "1.0"
-itertools = "0.10.0"
+itertools = "0.13.0"
 quick-xml = { version = "0.27.1", features = ["encoding"] }
+unic = "0.9.0"
diff --git a/src/engine/dict/build.rs b/src/engine/dict/build.rs
@@ -11,6 +11,7 @@ use std::{
     mem,
     path::PathBuf,
 };
+use unic::emoji::char::is_emoji;
 
 #[derive(Default, Debug, Clone, Copy)]
 struct HanjaEntry {
@@ -203,6 +204,10 @@ fn main() {
     )
     .unwrap();
     for entry in load_unicode_annotations().unwrap() {
+        if !entry.cp.chars().any(|c| is_emoji(c)) {
+            continue;
+        }
+
         writeln!(
             out,
             "UnicodeAnnotation {{ codepoint: \"{}\", tts: \"{}\" }},",