Add 'special' checks for DNA

crates/codebook/src/dictionaries/mod.rs

@@ -1,3 +1,4 @@
 pub mod dictionary;
 pub mod manager;
 pub mod repo;
+pub mod special;

crates/codebook/src/dictionaries/special.rs

@@ -0,0 +1,33 @@
+pub fn check_special(word: &str) -> bool {
+    is_dna_sequence(word)
+}
+
+fn is_dna_sequence(s: &str) -> bool {
+    if s.len() < 4 {
+        return false;
+    }
+    for c in s.chars() {
+        match c {
+            'A' | 'T' | 'C' | 'G' | 'a' | 't' | 'c' | 'g' => {
+                continue;
+            }
+            _ => return false,
+        }
+    }
+    true
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_dna_sequence() {
+        assert!(is_dna_sequence("ATCGATCG"));
+        assert!(is_dna_sequence("ATCG"));
+        assert!(is_dna_sequence("atcgatcg"));
+        assert!(!is_dna_sequence("xyzATCGAbc"));
+        assert!(!is_dna_sequence("Hello"));
+        assert!(!is_dna_sequence("ATC"));
+    }
+}

crates/codebook/src/lib.rs

@@ -7,7 +7,7 @@ mod splitter;
 use std::sync::Arc;
 
 use codebook_config::CodebookConfig;
-use dictionaries::{dictionary, manager::DictionaryManager};
+use dictionaries::{dictionary, manager::DictionaryManager, special::check_special};
 use dictionary::Dictionary;
 use parser::WordLocation;
 
@@ -43,6 +43,9 @@ impl Codebook {
             if self.config.is_allowed_word(word) {
                 return true;
             }
+            if check_special(word) {
+                return true;
+            }
             for dictionary in &dictionaries {
                 if dictionary.check(word) {
                     return true;