Initial prototype completed. Ready for feature extensions.

Author: jakseluz

README.md

@@ -23,7 +23,7 @@ GreekBase
 │   ├── ast_builder.py                ← used for generating AST tree from ANTLR tree
 │   ├── astGreek.py                   ← AST tree node classes
 │   ├── semantic_checker.py           ← semantic errors handling etc.
-│   └──
+│   └── codegen.py                    ← C (from AST) code generator
 ├── examples/                         ← example source files
 ├── output/                           ← generated C source files
 ├── run.sh                            ← bash script for generating ANTLR files
@@ -41,7 +41,8 @@ GreekBase
 	./[run.sh](./run.sh)
 in order to generate the parser and its tools needed later.
 
-For now, printing the AST tree is possible, as well as receiving semantic errors and warnings.
+For now, printing the AST tree is possible, as well as receiving semantic errors and warnings in the console.
+Output C source file is saved to ./[output/](./output/) directory if there are now [Error]s. 
 
 2. To do so, please run [main.py](./main.py).
 
@@ -51,7 +52,6 @@ For now, printing the AST tree is possible, as well as receiving semantic errors
 x : int := 5;
 y : int := 10;
 x : int;
-z := 5.5;
 
 if x < y then
     x := x + 1;
@@ -63,8 +63,26 @@ print y;
 ```
 2. AST tree:
 ```python
-Program(line=1, column=0, statements=[VariableDeclaration(line=1, column=0, varType=<class 'int'>, id='x', varValue=IntLiteral(line=1, column=11, value=5)), VariableDeclaration(line=2, column=0, varType=<class 'int'>, id='y', varValue=IntLiteral(line=2, column=11, value=10)), VariableDeclaration(line=3, column=0, varType=<class 'int'>, id='x', varValue=None), Assignment(line=4, column=0, id='z', value=FloatLiteral(line=4, column=5, value=5.5)), IfStatement(line=6, column=0, condition=Condition(line=6, column=3, left=Identifier(line=6, column=3, value='x'), operator='<', right=Identifier(line=6, column=7, value='y')), then_branch=[Assignment(line=7, column=4, id='x', value=AdditionOperator(line=7, column=9, left=Identifier(line=7, column=9, value='x'), operator='+', right=IntLiteral(line=7, column=13, value=1)))], else_branch=[Assignment(line=8, column=5, id='y', value=AdditionOperator(line=8, column=10, left=Identifier(line=8, column=10, value='y'), operator='-', right=IntLiteral(line=8, column=14, value=1)))]), PrintStatement(line=11, column=0, value=Identifier(line=11, column=6, value='x')), PrintStatement(line=12, column=0, value=Identifier(line=12, column=6, value='y'))])
+Program(line=1, column=0, statements=[VariableDeclaration(line=1, column=0, varType=<class 'int'>, id='x', varValue=IntLiteral(line=1, column=11, value=5)), VariableDeclaration(line=2, column=0, varType=<class 'int'>, id='y', varValue=IntLiteral(line=2, column=11, value=10)), VariableDeclaration(line=3, column=0, varType=<class 'int'>, id='x', varValue=None), IfStatement(line=5, column=0, condition=Condition(line=5, column=3, left=Identifier(line=5, column=3, value='x', type=None), operator='<', right=Identifier(line=5, column=7, value='y', type=None)), then_branch=[Assignment(line=6, column=4, id='x', value=AdditionOperator(line=6, column=9, left=Identifier(line=6, column=9, value='x', type=None), operator='+', right=IntLiteral(line=6, column=13, value=1)))], else_branch=[Assignment(line=7, column=5, id='y', value=AdditionOperator(line=7, column=10, left=Identifier(line=7, column=10, value='y', type=None), operator='-', right=IntLiteral(line=7, column=14, value=1)))]), PrintStatement(line=10, column=0, value=Identifier(line=10, column=6, value='x', type=None)), PrintStatement(line=11, column=0, value=Identifier(line=11, column=6, value='y', type=None))])
 ```
 3. Semantic check:
 
-![](./img/example1_semantic.png)
+![](./img/example1_semantic1.png)
+
+4. Code in C:
+```C
+#include <stdio.h>
+int main(){
+int x = 5;
+int y = 10;
+int x;
+if(x < y){
+x = x + 1;
+}else{ 
+y = y - 1;
+}
+printf("%d", x);
+printf("%d", y);
+return 0;
+}
+```

examples/example1.adan

@@ -1,7 +1,6 @@
 x : int := 5;
 y : int := 10;
 x : int;
-z := 5.5;
 
 if x < y then
     x := x + 1;

img/example1_semantic1.png

@@ -8,6 +8,7 @@
 from GreekBaseParser import GreekBaseParser
 from ast_builder import GreekASTBuilder
 from semantic_checker import SemanticChecker
+from codegen import CGenerator
 
 
 def main():
@@ -25,9 +26,15 @@ def main():
     print(ast)
 
     checker = SemanticChecker()
-    checker.analyze(ast)
+    tab = checker.analyze(ast)
     checker.finalise()
 
+    codegen = CGenerator(tab)
+    code = codegen.generate(ast)
+    
+    with open("output/example1.c", "w") as filehandler:
+        filehandler.write(code)
+
 
 if __name__ == '__main__':
     main()

main.py

@@ -0,0 +1,14 @@
+#include <stdio.h>
+int main(){
+int x = 5;
+int y = 10;
+int x;
+if(x < y){
+x = x + 1;
+}else{ 
+y = y - 1;
+}
+printf("%d", x);
+printf("%d", y);
+return 0;
+}

output/example1.c

@@ -76,14 +76,15 @@ class ParenthesisExpression(Expression):
 @dataclass
 class Identifier(Expression):
     value: str
+    type: str
 
 
 # NON-DECLARATIVE STATEMENTS
 @dataclass
 class IfStatement(NonDeclarativeStatement):
     condition: Condition
     then_branch : List[NonDeclarativeStatement]
-    else_branch: List[NonDeclarativeStatement]
+    else_branch: List[NonDeclarativeStatement] | None
 
 @dataclass
 class LoopStatement(NonDeclarativeStatement):
@@ -98,6 +99,7 @@ class Assignment(NonDeclarativeStatement):
 @dataclass
 class PrintStatement(NonDeclarativeStatement):
     value: Expression
+    # id: str | None # TODO
 
 
 # RELATION OPERATORS
@@ -153,4 +155,4 @@ class Procedure(Statement):
 class VariableDeclaration(NonDeclarativeStatement):
     varType: VariableType
     id: Identifier
-    varValue : Literal
+    varValue : Literal | None

src/astGreek.py

@@ -45,7 +45,7 @@ def visitExpression(self, ctx: GreekBaseParser.ExpressionContext):
 
     def visitIdExpr(self, ctx: GreekBaseParser.IdExprContext):
         if ctx.IDENTIFIER():
-            return ast.Identifier(ctx.start.line, ctx.start.column, ctx.IDENTIFIER().getText())
+            return ast.Identifier(ctx.start.line, ctx.start.column, ctx.IDENTIFIER().getText(), None)
 
     def visitLiteral(self, ctx: GreekBaseParser.LiteralContext):
         if ctx.LIT_INT():

src/ast_builder.py

@@ -0,0 +1,103 @@
+import astGreek as ast
+
+class CGenerator:
+    def __init__(self, table: dict):
+        self.symbol_table = table
+
+    def generate(self, node: ast.ASTNode):
+        method = f'gen_{type(node).__name__}'
+        # print("\nteraz: ", method) # for debugging
+        if method:
+            return getattr(self, method)(node)
+    
+    def gen_Program(self, node : ast.ASTNode):
+        body = "\n".join([self.generate(statement) for statement in node.statements])
+        return f'#include <stdio.h>\nint main(){{\n{body}\nreturn 0;\n}}'
+
+    def gen_IntLiteral(self, node: ast.IntLiteral):
+        return str(node.value)
+    
+    def gen_FloatLiteral(self, node: ast.FloatLiteral):
+        return str(node.value)
+    
+    def gen_CharLiteral(self, node: ast.CharLiteral):
+        return '\'' + node.value + '\''
+    
+    def gen_StringLiteral(self, node: ast.StringLiteral):
+        return "\"" + node.value + "\""
+    
+    def gen_Condition(self, node: ast.Condition):
+        return " ".join(
+            [
+                self.generate(node.left),
+                "==" if node.operator == '=' else ("!=" if node.operator == "/=" else node.operator),
+                self.generate(node.right)
+            ]
+        )
+    
+    def gen_MultiplicationOperator(self, node: ast.MultiplicationOperator):
+        return " ".join(
+            [
+                self.generate(node.left),
+                node.operator,
+                self.generate(node.right)
+            ]
+        )
+    
+    def gen_AdditionOperator(self, node: ast.AdditionOperator):
+        return " ".join(
+            [
+                self.generate(node.left),
+                node.operator,
+                self.generate(node.right)
+            ]
+        )
+    
+    def gen_ParenthesisExpression(self, node: ast.ParenthesisExpression):
+        return f"( {self.generate(node.value)} )"
+    
+    def gen_Identifier(self, node: ast.Identifier):
+        return node.value
+    
+    def gen_IfStatement(self, node: ast.IfStatement):
+        return "\n".join(
+            [
+                f"if({self.generate(node.condition)})" + '{',
+
+                *[self.generate(statement) for statement in node.then_branch],
+                
+                "}else{ " if len(node.else_branch) > 0 else ("}else{" if len(node.else_branch) == 1 else None), # TODO
+
+                *[self.generate(statement) for statement in node.else_branch if len(node.else_branch) > 0],
+                "}"
+            ]
+        )
+    
+    def gen_LoopStatement(self, node: ast.LoopStatement):
+        return "\n".join(
+            [
+                f"while({self.generate(node.condition)})" + '{',
+                *[self.generate(statement) for statement in node.then],
+                '}'
+            ]
+        )
+    
+    def gen_Assignment(self, node: ast.Assignment):
+        # self.symbol_table[node.id] = self.generate(node.value)
+        return f"{node.id} = {self.generate(node.value)};"
+    
+    def gen_PrintStatement(self, node: ast.PrintStatement): # TODO
+        value_type = self.symbol_table[node.value.value][0]
+        if(value_type == int):
+            value = f"\"%d\", {node.value.value}"
+        elif(value_type == float):
+            value = f"\"%lf\", {node.value.value}"
+        elif(value_type == str):
+            value = f"\"%s\", {node.value.value}"
+
+        return f"printf({value});"
+    
+    def gen_VariableDeclaration(self, node: ast.VariableDeclaration):
+        #if(node.varValue is not None):
+        #    self.symbol_table[node.id] = self.generate(node.varValue)
+        return f"{node.varType.__name__} {node.id}" + ((f" = {node.varValue.value}") if node.varValue is not None else '') + ';'

src/codegen.py

@@ -35,6 +35,7 @@ def analyze_Program(self, node: ast.Program):
         # Undoubtedly the first node. Here the fun starts.
         for statement in node.statements:
             self.analyze(statement)
+        return self.symbol_table
 
     def analyze_VariableDeclaration(self, node: ast.VariableDeclaration):
         if(node.id in self.symbol_table):
@@ -44,6 +45,8 @@ def analyze_VariableDeclaration(self, node: ast.VariableDeclaration):
                 self.errors[f"{self.RED}[Error]: {self.WHITE}Variable {node.id} redeclared with a different type than earlier ({node.varType} instead of {self.symbol_table[node.id][0]}) {self.PURPLE}in line {node.line}, column {node.column}{self.WHITE}!"] = 1
         else:
             self.symbol_table[node.id] = (node.varType, node.varValue)
+        if(node.varValue):
+            node.varValue.type = node.varType.__name__
 
     def analyze_Assignment(self, node: ast.Assignment):
         # node_type = self.analyze(node.value)
@@ -71,6 +74,7 @@ def analyze_MultiplicationOperator(self, node: ast.MultiplicationOperator):
     def analyze_Identifier(self, node: ast.Identifier):
         if(node.value in self.symbol_table):
             ident_val_type = self.symbol_table[node.value][0]
+            node.type = ident_val_type.__name__
             return ident_val_type
         else:
             self.errors[f"{self.RED}[ERROR]: {self.WHITE}Unknown identifier reference: {node.value} {self.PURPLE}in line {node.line}, column {node.column}{self.WHITE}!"] = 1