feat(tree-sitter): Add support for Java 22 (#309)

Co-authored-by: Timon Vonk <[email protected]>

.gitignore

@@ -22,3 +22,4 @@ tmp
 .env*.local
 
 **/.fastembed_cache
+.idea/

swiftide-integrations/Cargo.toml

@@ -46,6 +46,7 @@ tree-sitter-python = { version = "0.23", optional = true }
 tree-sitter-ruby = { version = "0.23", optional = true }
 tree-sitter-typescript = { version = "0.23", optional = true }
 tree-sitter-javascript = { version = "0.23", optional = true }
+tree-sitter-java = { version = "0.23", optional = true }
 fastembed = { version = "4.0", optional = true }
 spider = { workspace = true, optional = true }
 htmd = { version = "0.1", optional = true }
@@ -110,6 +111,7 @@ tree-sitter = [
   "dep:tree-sitter-ruby",
   "dep:tree-sitter-typescript",
   "dep:tree-sitter-javascript",
+  "dep:tree-sitter-java",
 ]
 # OpenAI for embedding and prompting
 openai = ["dep:async-openai"]

swiftide-integrations/src/treesitter/code_tree.rs

@@ -8,7 +8,7 @@ use tree_sitter::{Parser, Query, QueryCursor, Tree};
 use anyhow::{Context as _, Result};
 use std::collections::HashSet;
 
-use crate::treesitter::queries::{javascript, python, ruby, rust, typescript};
+use crate::treesitter::queries::{java, javascript, python, ruby, rust, typescript};
 
 use super::SupportedLanguages;
 
@@ -107,7 +107,7 @@ impl CodeTree<'_> {
 }
 
 fn ts_queries_for_language(language: SupportedLanguages) -> (&'static str, &'static str) {
-    use SupportedLanguages::{Javascript, Python, Ruby, Rust, Typescript};
+    use SupportedLanguages::{Java, Javascript, Python, Ruby, Rust, Typescript};
 
     match language {
         Rust => (rust::DEFS, rust::REFS),
@@ -116,6 +116,7 @@ fn ts_queries_for_language(language: SupportedLanguages) -> (&'static str, &'sta
         Typescript => (typescript::DEFS, typescript::REFS),
         Javascript => (javascript::DEFS, javascript::REFS),
         Ruby => (ruby::DEFS, ruby::REFS),
+        Java => (java::DEFS, java::REFS),
     }
 }
 
@@ -231,4 +232,52 @@ mod tests {
         assert_eq!(result.definitions, vec!["MyClass", "Test", "myMethod"]);
         assert_eq!(result.references, vec!["log", "otherThing"]);
     }
+
+    #[test]
+    fn test_parsing_on_java() {
+        let parser = CodeParser::from_language(SupportedLanguages::Java);
+        let code = r#"
+        public class Hello {
+            public static void main(String[] args) {
+                System.out.printf("Hello %s!%n", args[0]);
+            }
+        }
+        "#;
+        let tree = parser.parse(code).unwrap();
+        let result = tree.references_and_definitions().unwrap();
+        assert_eq!(result.definitions, vec!["Hello", "main"]);
+        assert_eq!(result.references, vec!["printf"]);
+    }
+
+    #[test]
+    fn test_parsing_on_java_enum() {
+        let parser = CodeParser::from_language(SupportedLanguages::Java);
+        let code = r#"
+        enum Material {
+            DENIM,
+            CANVAS,
+            SPANDEX_3_PERCENT
+        }
+
+        class Person {
+
+
+          Person(string name) {
+            this.name = name;
+
+            this.pants = new Pants<Pocket>();
+          }
+
+          String getName() {
+            a = this.name;
+            b = new one.two.Three();
+            c = Material.DENIM;
+          }
+        }
+        "#;
+        let tree = parser.parse(code).unwrap();
+        let result = tree.references_and_definitions().unwrap();
+        assert_eq!(result.definitions, vec!["Material", "Person", "getName"]);
+        assert!(result.references.is_empty());
+    }
 }

swiftide-integrations/src/treesitter/metadata_refs_defs_code.rs

@@ -105,6 +105,7 @@ mod test {
     #[test_case("python", "def main(): print('Hello, World!')", "print", "main"; "python")]
     #[test_case("javascript", "function main() { console.log('Hello, World!'); }", "log", "main"; "javascript")]
     #[test_case("typescript", "function main() { console.log('Hello, World!'); }", "log", "main"; "typescript")]
+    #[test_case("java", "public class Main { public static void main(String[] args) { System.out.println(\"Hello, World!\"); } }", "println", "Main,main"; "java")]
     #[tokio::test]
     async fn assert_refs_defs_from_code(
         lang: &str,

swiftide-integrations/src/treesitter/outliner.rs

@@ -98,7 +98,7 @@ impl CodeOutliner {
 
     fn is_unneeded_node(&self, node: Node) -> bool {
         match self.language {
-            SupportedLanguages::Rust => matches!(node.kind(), "block"),
+            SupportedLanguages::Rust | SupportedLanguages::Java => matches!(node.kind(), "block"),
             SupportedLanguages::Typescript | SupportedLanguages::Javascript => {
                 matches!(node.kind(), "statement_block")
             }
@@ -298,4 +298,28 @@ class Bla {
             "\nimport { Context as _, Result } from 'anyhow';\n// This is a comment\nfunction main(a, b) \n\nclass Bla {\n    constructor() \n\n    ok() \n}"
         );
     }
+
+    #[test]
+    fn test_outline_java() {
+        let code = r#"
+import java.io.PrintStream;
+import java.util.Scanner;
+
+public class HelloWorld {
+    // This is a comment
+    public static void main(String[] args) {
+        PrintStream out = System.out;
+
+        out.println("Hello, World!");
+    }
+}
+"#;
+        let outliner = CodeOutliner::new(SupportedLanguages::Java);
+        let summary = outliner.outline(code).unwrap();
+        println!("{}", summary);
+        assert_eq!(
+            summary,
+            "\nimport java.io.PrintStream;\nimport java.util.Scanner;\n\npublic class HelloWorld {\n    // This is a comment\n    public static void main(String[] args) \n}"
+        );
+    }
 }

swiftide-integrations/src/treesitter/queries.rs

@@ -294,3 +294,31 @@ pub mod rust {
                 macro: (identifier) @name)
         ";
 }
+
+// https://github.com/tree-sitter/tree-sitter-java/blob/master/queries/tags.scm
+pub mod java {
+    pub const DEFS: &str = "
+           (class_declaration
+                name: (identifier) @name)
+
+           (enum_declaration
+                name: (identifier) @name)
+
+            (method_declaration
+                name: (identifier) @name)
+
+            (interface_declaration
+                name: (identifier) @name)
+
+            (type_list
+                (type_identifier) @name)
+
+            (superclass (type_identifier) @name)";
+    pub const REFS: &str = "
+            (method_invocation
+                name: (identifier) @name
+                arguments: (argument_list))
+
+            (object_creation_expression
+                type: (type_identifier) @name)";
+}

swiftide-integrations/src/treesitter/supported_languages.rs

@@ -44,6 +44,8 @@ pub enum SupportedLanguages {
     Ruby,
     #[serde(alias = "javascript")]
     Javascript,
+    #[serde(alias = "java")]
+    Java,
 }
 
 /// Static array of file extensions for Rust files.
@@ -61,6 +63,9 @@ static RUBY_EXTENSIONS: &[&str] = &["rb"];
 /// Static array of file extensions for Javascript files.
 static JAVASCRIPT_EXTENSIONS: &[&str] = &["js", "jsx"];
 
+/// Static array of file extensions for Java files.
+static JAVA_EXTENSIONS: &[&str] = &["java"];
+
 impl SupportedLanguages {
     /// Returns the file extensions associated with the supported language.
     ///
@@ -73,6 +78,7 @@ impl SupportedLanguages {
             SupportedLanguages::Python => PYTHON_EXTENSIONS,
             SupportedLanguages::Ruby => RUBY_EXTENSIONS,
             SupportedLanguages::Javascript => JAVASCRIPT_EXTENSIONS,
+            SupportedLanguages::Java => JAVA_EXTENSIONS,
         }
     }
 }
@@ -95,6 +101,7 @@ impl From<SupportedLanguages> for tree_sitter::Language {
             SupportedLanguages::Typescript => tree_sitter_typescript::LANGUAGE_TYPESCRIPT,
             SupportedLanguages::Javascript => tree_sitter_javascript::LANGUAGE,
             SupportedLanguages::Ruby => tree_sitter_ruby::LANGUAGE,
+            SupportedLanguages::Java => tree_sitter_java::LANGUAGE,
         }
         .into()
     }
@@ -116,6 +123,10 @@ mod test {
             SupportedLanguages::from_str("typescript"),
             Ok(SupportedLanguages::Typescript)
         );
+        assert_eq!(
+            SupportedLanguages::from_str("java"),
+            Ok(SupportedLanguages::Java)
+        );
     }
 
     /// Tests the case-insensitive string conversion for `SupportedLanguages` with different casing.
@@ -129,6 +140,11 @@ mod test {
             SupportedLanguages::from_str("TypeScript"),
             Ok(SupportedLanguages::Typescript)
         );
+
+        assert_eq!(
+            SupportedLanguages::from_str("Java"),
+            Ok(SupportedLanguages::Java)
+        );
     }
 
     #[test]