Refactor Rust identifier constants

typedb · Feb 14, 2024 · 7237e05 · 7237e05
1 parent 7c8eb18
commit 7237e05
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 32 deletions.
diff --git a/rust/common/identifier.rs b/rust/common/identifier.rs
@@ -23,10 +23,7 @@ use std::sync::OnceLock;
 
 use regex::{Regex, RegexBuilder};
 
-pub fn is_valid_label_identifier(identifier: &str) -> bool {
-    static REGEX: OnceLock<Regex> = OnceLock::new();
-    let regex = REGEX.get_or_init(|| {
-        let identifier_start = "A-Za-z\
+const IDENTIFIER_CHAR: &str = "A-Za-z\
             \\u00C0-\\u00D6\
             \\u00D8-\\u00F6\
             \\u00F8-\\u02FF\
@@ -38,17 +35,23 @@ pub fn is_valid_label_identifier(identifier: &str) -> bool {
             \\u3001-\\uD7FF\
             \\uF900-\\uFDCF\
             \\uFDF0-\\uFFFD";
-        let identifier_tail = format!(
-            "{}\
-            0-9\
-            _\
+const IDENTIFIER_CONNECTOR: &str = "_\
             \\-\
             \\u00B7\
             \\u0300-\\u036F\
-            \\u203F-\\u2040",
-            identifier_start
+            \\u203F-\\u2040";
+const IDENTIFIER_DIGIT: &str = "0-9";
+
+pub fn is_valid_label_identifier(identifier: &str) -> bool {
+    static REGEX: OnceLock<Regex> = OnceLock::new();
+    let regex = REGEX.get_or_init(|| {
+        let identifier_tail = format!(
+            "{}{}{}",
+            IDENTIFIER_CHAR,
+            IDENTIFIER_CONNECTOR,
+            IDENTIFIER_DIGIT
         );
-        let identifier_pattern = format!("^[{}][{}]*$", identifier_start, identifier_tail);
+        let identifier_pattern = format!("^[{}][{}]*$", IDENTIFIER_CHAR, identifier_tail);
         RegexBuilder::new(&identifier_pattern).build().unwrap()
     });
     regex.is_match(identifier)
@@ -58,29 +61,19 @@ pub fn is_valid_label_identifier(identifier: &str) -> bool {
 pub fn is_valid_var_identifier(identifier: &str) -> bool {
     static REGEX: OnceLock<Regex> = OnceLock::new();
     let regex = REGEX.get_or_init(|| {
-        let identifier_start = "A-Za-z0-9\
-            \\u00C0-\\u00D6\
-            \\u00D8-\\u00F6\
-            \\u00F8-\\u02FF\
-            \\u0370-\\u037D\
-            \\u037F-\\u1FFF\
-            \\u200C-\\u200D\
-            \\u2070-\\u218F\
-            \\u2C00-\\u2FEF\
-            \\u3001-\\uD7FF\
-            \\uF900-\\uFDCF\
-            \\uFDF0-\\uFFFD";
+        let identifier_start = format!(
+            "{}{}",
+            IDENTIFIER_CHAR,
+            IDENTIFIER_DIGIT
+        );
         let identifier_tail = format!(
-            "{}\
-            _\
-            \\-\
-            \\u00B7\
-            \\u0300-\\u036F\
-            \\u203F-\\u2040",
-            identifier_start
+            "{}{}{}",
+            IDENTIFIER_CHAR,
+            IDENTIFIER_DIGIT,
+            IDENTIFIER_CONNECTOR
         );
         let identifier_pattern = format!("^[{}][{}]*$", identifier_start, identifier_tail);
         RegexBuilder::new(&identifier_pattern).build().unwrap()
     });
     regex.is_match(identifier)
-}
+}
diff --git a/rust/parser/typeql.pest b/rust/parser/typeql.pest
@@ -357,8 +357,8 @@ VAR_CONCEPT_ANONYMOUS_ = @{ "$_" ~ WB }
 VAR_CONCEPT_NAMED_ = @{ "$" ~ IDENTIFIER_VAR_H_ ~ IDENTIFIER_VAR_T_*  ~ WB }
 VAR_VALUE_ = @{ "?" ~ (IDENTIFIER_LABEL_H_ ~ IDENTIFIER_LABEL_T_* ~ WB) }
 IID_ = @{ "0x" ~ ASCII_HEX_DIGIT+ ~ WB }
-LABEL_SCOPED_ = @{ (IDENTIFIER_LABEL_H_ ~ IDENTIFIER_LABEL_T_* ) ~ ":" ~ (IDENTIFIER_LABEL_H_ ~ IDENTIFIER_LABEL_T_* ~ WB) }
 LABEL_ = @{ (IDENTIFIER_LABEL_H_ ~ IDENTIFIER_LABEL_T_* ~ WB) }
+LABEL_SCOPED_ = @{ (IDENTIFIER_LABEL_H_ ~ IDENTIFIER_LABEL_T_* ) ~ ":" ~ (IDENTIFIER_LABEL_H_ ~ IDENTIFIER_LABEL_T_* ~ WB) }
 
 
 // FRAGMENTS OF KEYWORDS =======================================================