feat: add parser visitor for Tamil keywords

src/tampy/transpiler.py

@@ -0,0 +1,82 @@
+"""Parse Tamil code into Python AST."""
+
+from typing import Any
+import ast
+
+
+class TamilParser(ast.NodeVisitor):
+    """Parse Tamil code and map keywords to Python."""
+
+    def __init__(self, keywords: dict[str, str]):
+        self.keywords = keywords
+        self.errors: list[str] = []
+    
+    def parse(self, code: str) -> ast.AST:
+        """Parse Tamil code and return Python AST."""
+        try:
+            tree = ast.parse(code)
+            self._transform(tree)
+            return tree
+        except SyntaxError as e:
+            self.errors.append(f"SyntaxError at line {e.lineno}: {e.msg}")
+            raise
+    
+    def _transform(self, node: ast.AST) -> None:
+        """Transform node to handle Tamil keywords."""
+        node = self._visit(node)
+        if isinstance(node, ast.AST):
+            for field, value in ast.iter_fields(node):
+                if isinstance(value, list):
+                    for i, item in enumerate(value):
+                        if isinstance(item, ast.AST):
+                            value[i] = self._visit(item)
+                elif isinstance(value, ast.AST):
+                    setattr(node, field, self._visit(value))
+    
+    def _visit(self, node: ast.AST) -> Any:
+        """Visit and transform a single node."""
+        if not hasattr(node, "lineno"):
+            return node
+        
+        line = node.lineno
+        col = getattr(node, "col_offset", 0)
+        
+        keyword = self._get_keyword_at_position(node, line, col)
+        if keyword:
+            node.keywords = self.keywords.get(keyword, keyword)
+        
+        visitor = self._get_visitor_type(type(node))
+        if visitor:
+            return visitor(node)
+        
+        return node
+    
+    def _get_keyword_at_position(self, node: ast.AST, lineno: int, col: int) -> str | None:
+        """Check if keyword appears at given position."""
+        # This is a placeholder - actual keyword detection will use tokens
+        # For now, we'll handle common keywords by name
+        node_type = type(node).__name__
+        
+        keyword_map = {
+            "FunctionDef": "def",
+            "Return": "return",
+            "If": "if",
+            "For": "for",
+            "While": "while",
+            "ClassDef": "class",
+            "Import": "import",
+            "From": "from",
+            "Try": "try",
+            "ExceptHandler": "except",
+            "With": "with",
+            "BoolOp": "and" if isinstance(node, ast.BoolOp) and node.op.__class__.__name__ == "And" else "or",
+            "Compare": "in",
+            "NamedExpr": "if",
+        }
+        
+        return keyword_map.get(node_type)
+    
+    def _get_visitor_type(self, node_type: type) -> Any | None:
+        """Get visitor method for node type."""
+        visitor_name = f"visit_{node_type.__name__}"
+        return getattr(self, visitor_name, None)

tests/test_parser.py

@@ -0,0 +1,44 @@
+"""Test parser functionality."""
+
+import pytest
+from src.tampy.transpiler import TamilParser
+
+
+def test_parse_simple_python():
+    """Test parsing simple Python code."""
+    code = "print('hello')"
+    keywords = {"print": "இருப்பு"}
+    
+    parser = TamilParser(keywords)
+    tree = parser.parse(code)
+    
+    assert isinstance(tree, ast.Module)
+
+
+def test_parse_import_statement():
+    """Test parsing import statement."""
+    code = "import sys"
+    keywords = {"import": "மேற்கோள்கள்"}
+    
+    parser = TamilParser(keywords)
+    tree = parser.parse(code)
+    
+    assert isinstance(tree, ast.Module)
+
+
+def test_parse_function_definition():
+    """Test parsing function definition."""
+    code = "def foo(): pass"
+    keywords = {"def": "வரையறை"}
+    
+    parser = TamilParser(keywords)
+    tree = parser.parse(code)
+    
+    assert isinstance(tree, ast.Module)
+
+
+if __name__ == "__main__":
+    test_parse_simple_python()
+    test_parse_import_statement()
+    test_parse_function_definition()
+    print("All tests passed")