/// <summary>
/// Sets the current node's value and it's left child
/// </summary>
public void Insert(String value, Node leftNode)
{
CurrentNode.Value = value;
CurrentNode.Left = new Node();
CurrentNode.Right = new Node();
CurrentNode.Left = leftNode;
CurrentNode = CurrentNode.Right;
}
/// <summary>
/// Returns a cloned node with all the same properties
/// </summary>
public Node Clone()
{
Node clone = new Node();
clone.Left = Left;
clone.Right = Right;
clone.Value = Value;
clone.Attributes = Attributes;
return clone;
}
/// <summary>
/// Given a node, deletes all the children
/// </summary>
public static void DeleteTree(Node Root)
{
if (Root.Left != null)
DeleteTree(Root.Left);
if (Root.Right != null)
DeleteTree(Root.Right);
Root = null;
}
/// <summary>
/// Outputs the parse tree in a readable format
/// </summary>
public static void DisplayTree(Node node, String inner)
{
if (node == null)
return;
Console.Write(inner + ">" + node.Value);
String Attrib = "";
foreach (String Attribute in node.Attributes)
if (Attribute != "")
Attrib += " " + Attribute;
if (Attrib != "" && node.Value != "STATEMENTS")
Console.Write(" (Attributes: " + Attrib + ")");
Console.Write("\n");
DisplayTree(node.Left, inner + "-");
DisplayTree(node.Right, inner + "-");
}
/// <summary>
/// Traverses the tree in post-order (bottom-up) and optimises any conditions
/// </summary>
private void OptimiseConditionalsOfNode(Node currentNode)
{
if (currentNode.Left != null)
{
OptimiseConditionalsOfNode(currentNode.Left);
}
if (currentNode.Right != null)
{
OptimiseConditionalsOfNode(currentNode.Right);
}
if (Parser.IsConditionalOperator(currentNode.Attributes[0]))
{
//If left node and right node are both values
Int32 leftValue;
Int32 rightValue;
if (currentNode.Left != null && Int32.TryParse(currentNode.Left.Value, out leftValue) &&
currentNode.Right != null && Int32.TryParse(currentNode.Right.Value, out rightValue) &&
currentNode.Left.Attributes[1] == "VALUE" && currentNode.Right.Attributes[1] == "VALUE")
{
if (currentNode.Attributes[0] == "==")
currentNode.Value = (leftValue == rightValue ? 1 : 0).ToString();
else if (currentNode.Attributes[0] == "<=")
currentNode.Value = (leftValue <= rightValue ? 1 : 0).ToString();
else if (currentNode.Attributes[0] == ">=")
currentNode.Value = (leftValue >= rightValue ? 1 : 0).ToString();
else
currentNode.Value = (leftValue != rightValue ? 1 : 0).ToString();
currentNode.Attributes[1] = "VALUE";
currentNode.Attributes[0] = currentNode.Attributes[2] = "";
currentNode.Left = null;
currentNode.Right = null;
}
}
}
/// <summary>
/// Parse a new statement
/// </summary>
private Node ParseStatement(Node node, ref Node currentNodeMarker)
{
//Call appropriate parse function according to first Token in the statement
if (VariableTypes.Contains(GetCurrentToken().Value))
node = DeclareVariable(node);
else if (GetCurrentToken().Value.Equals("INPUT"))
node = ParseInputFunction(node);
else if (GetCurrentToken().Value.Equals("PRINT"))
node = ParsePrintFunction(node);
else if (GetCurrentToken().Type.Equals(TokenType.Identifier) && Tokens[TokenIndex+1].Value.Equals("="))
node = AssignExpression(node);
else if (GetCurrentToken().Value.Equals("IF"))
node = ParseIfStatement(node, ref currentNodeMarker);
else
{
Console.WriteLine("INVALID STATEMENT on line " + GetCurrentToken().Line + ": " + GetCurrentToken().Value);
node = null;
}
return node;
}
/// <summary>
/// <PRINT> ::= 'PRINT', ( <IDENTIFIER> | <CONSTANT> );
/// </summary>
private Node ParsePrintFunction(Node node)
{
node = new Node();
node.Value = "PRINT";
TokenIndex++;
CheckIfOutOfTokens();
node.Attributes[0] = GetCurrentToken().Value;
//Semantic part
//Check for validity - argument should be variable or constant
if (!GetCurrentToken().Type.Equals(TokenType.Identifier) && !GetCurrentToken().Type.Equals(TokenType.String)
&& !GetCurrentToken().Type.Equals(TokenType.Number))
Console.Write("Invalid argument for PRINT statement");
//And in case of variable, it must exist
if (GetCurrentToken().Type.Equals(TokenType.Identifier) && !VariableExists(GetCurrentToken().Value))
Console.Write("Undeclared variable name used as argument in PRINT");
//Store some attributes to aid in code gen
if (GetCurrentToken().Type.Equals(TokenType.Number) ||
GetVariableType(GetCurrentToken().Value).Equals(VariableType.Integer))
node.Attributes[1] = "INT";
else
node.Attributes[1] = "STRING";
if (GetCurrentToken().Type.Equals(TokenType.Identifier))
node.Attributes[2] = "VARIABLE";
else
node.Attributes[2] = "VALUE";
TokenIndex++;
return node;
}
/// <summary>
/// <INPUT> ::= 'INPUT', <IDENTIFIER>
/// </summary>
private Node ParseInputFunction(Node node)
{
node = new Node();
node.Value = "INPUT";
TokenIndex++;
CheckIfOutOfTokens();
node.Attributes[0] = GetCurrentToken().Value;
//Semantic part
if (!GetCurrentToken().Type.Equals(TokenType.Identifier))
Console.WriteLine("Invalid argument for input statement");
if (!VariableExists(GetCurrentToken().Value))
Console.WriteLine("Undeclared variable name used as argument in INPUT");
if (GetVariableType(GetCurrentToken().Value).Equals(VariableType.Integer))
node.Attributes[1] = "INT";
else
node.Attributes[1] = "STRING";
TokenIndex++;
return node;
}
/// <summary>
/// <IFSTATEMENT> := 'IF (' <CONDITION> ')' <STATEMENTS> 'END'
/// If statements within if statements do not work
/// </summary>
private Node ParseIfStatement(Node node, ref Node currentNodeMarker)
{
node = new Node();
node.Value = "IF";
TokenIndex++;
//Semantics
if (!GetCurrentToken().Value.Equals("("))
Console.Write("Enclose the condition in (...)");
TokenIndex++;
//Parse the condition in the left child of the IF node
node.Left = ParseCondition(node.Left);
//Semantics
if (!GetCurrentToken().Value.Equals(")"))
Console.Write("Enclose the condition in (...)");
//Skip the right bracket and new line
TokenIndex++;
TokenIndex++;
//Create a root class node for the top of the IF tree
Root root = new Root();
//Insert the IF node in the root
root.Insert(node.Value, node.Left);
//Mark the current node and set it
currentNodeMarker = node.Right;
Node currentNode = node.Right;
//Parse statements until END is found
while (!GetCurrentToken().Value.Equals("END"))
{
if (GetCurrentToken().Type.Equals(TokenType.EOL))
{
TokenIndex++;
}
//Insert the statements in
currentNode = ParseStatement(currentNode, ref currentNodeMarker);
if (currentNode != null)
{
root.Insert("STATEMENTS", currentNode);
currentNode = currentNode.Right;
currentNodeMarker = currentNode;
}
TokenIndex++;
}
//The final right node will always be empty, so remove it
root.NullifyEmptyChildren();
return (Node)root;
}
/// <summary>
/// Parse a facor in an expression
/// EBNF: <factor> ::= <IDENTIFIER> | <CONSTANT> | ( ‘(’, <expression>, ‘)’ ) | ( '-', <factor> );
/// We need to have the node where the factor is to be held
/// A the type so we know whether we are expecting an int or a string
/// </summary>
private Node ParseFactor(Node node, VariableType type)
{
//Check if we have any tokens left
CheckIfOutOfTokens();
//First check for a parenthesised expression
//( ‘(’, <expression>, ‘)’ )
if (GetCurrentToken().Value.Equals("("))
{
//Skip the parenthesis
TokenIndex++;
//Get the expression
node = ParseExpression(node, type);
if (!GetCurrentToken().Value.Equals(")"))
Console.WriteLine("Expected closing bracket");
//Skip the closing paren
TokenIndex++;
return node;
}
//If it's not a parenthesised expression
//Check for a negative number
//( '-', <factor> )
if (type.Equals(VariableType.Integer))
{
if (GetCurrentToken().Value.Equals("-"))
{
TokenIndex++;
node = ParseFactor(node, type);
//Apply the negative sign to the parsed factor
node.Value = "-" + node.Value;
return node;
}
}
//We have narrowed it down to either a <IDENTIFIER> or <CONSTANT>
node = new Node();
//Store the value in the node
node.Value = GetCurrentToken().Value;
//Semantics - check if variable exists
if (GetCurrentToken().Type.Equals(TokenType.Identifier) && !VariableExists(GetCurrentToken().Value))
Console.WriteLine("Undeclared variable");
//Give the node extra info to aid in code gen
if (GetCurrentToken().Type.Equals(TokenType.Identifier))
node.Attributes[0] = "VARIABLE";
else
node.Attributes[1] = "VALUE";
TokenIndex++;
return node;
}
/// <summary>
/// To declare a variable
/// In EBNF:
/// <DECLARE> ::= ( 'INT' | 'STRING' | 'BOOL'), <IDENTIFIER>;
/// </summary>
private Node DeclareVariable(Node node)
{
node = new Node();
node.Value = "DECLARE";
node.Attributes[1] = GetCurrentToken().Value;
TokenIndex++;
CheckIfOutOfTokens();
node.Attributes[0] = GetCurrentToken().Value;
//Semantics - check variable name is valid
if (!GetCurrentToken().Type.Equals(TokenType.Identifier) || IsNewVariableNameInvalid(GetCurrentToken().Value))
Console.WriteLine("Invalid variable name used in declaration");
//Add the variable to the list for future reference
var varInfo = new VariableInfo();
varInfo.Name = GetCurrentToken().Value;
varInfo.Type = GetVariableFromValue(node.Attributes[1]);
Variables.Add(varInfo);
TokenIndex++;
return node;
}
/// <summary>
/// To assign expression to a varibale
/// In EBNF:
/// <ASSIGN> ::= <IDENTIFIER>, '=', <EXPRESSION>;
/// </summary>
private Node AssignExpression(Node node)
{
node.Value = "ASSIGN";
node.Left = new Node();
node.Left.Value = GetCurrentToken().Value;
//Skip through '=' symbol
TokenIndex += 2;
node.Right = ParseExpression(node.Right, GetVariableType(node.Left.Value));
//Semantics - checks variable exists
if (!VariableExists(node.Left.Value))
Console.WriteLine("Undeclared var name used in assignment");
//Additional node information to aid code gen
if (GetVariableType(node.Left.Value).Equals(VariableType.Integer))
node.Attributes[0] = "INT";
else if (GetVariableType(node.Left.Value).Equals(VariableType.Boolean))
node.Attributes[0] = "BOOL";
else
node.Attributes[0] = "STRING";
return node;
}
/// <summary>
/// Inserts the final node in the parse tree
/// </summary>
public void InsertFinal(Node node)
{
CurrentNode.Value = "STATEMENTS";
CurrentNode.Left = node;
}
public Root()
{
CurrentNode = this;
}
/// <summary>
/// Recursively searches through the child nodes
/// </summary>
public void NullifyEmptyChildren()
{
if (Left != null)
{
Left.NullifyEmptyChildren();
if (Left.IsEmpty())
Left = null;
}
if (Right != null)
{
Right.NullifyEmptyChildren();
if (Right.IsEmpty())
Right = null;
}
}
/// <summary>
/// To parse a term in the epression
/// In EBNF: <term> ::= <factor>, { ‘*’|’/’, <factor> };
/// </summary>
private Node ParseTerm(Node node)
{
CheckIfOutOfTokens();
//Parse the first factor
node = ParseFactor(node, VariableType.Integer);
//Then the {...}
//For each * or / parse another factor
while (GetCurrentToken().Value.Equals("*") || GetCurrentToken().Value.Equals("/"))
{
Node clonedNode = node.Clone();
node.Left = clonedNode;
node.Value = GetCurrentToken().Value;
TokenIndex++;
node.Right = ParseFactor(node.Right, VariableType.Integer);
}
return node;
}
/// <summary>
/// Parses all the tokens
/// </summary>
public void ParseProgram()
{
//Last Token must be EOL so add one in case
Token eol = new Token();
eol.Line = Tokens[Tokens.Count - 1].Line + 1;
eol.Type = TokenType.EOL;
eol.Value = "\n";
Tokens.Add(eol);
//Start at the first Token
TokenIndex = 0;
for (int line = 0; line < LineCount && GetCurrentToken() != null; line++)
{
Node currentNode = new Node();
Node currentNodeMarker;
if (!GetCurrentToken().Type.Equals(TokenType.EOL))
{
currentNodeMarker = null;
if (line == LineCount - 1)
{
currentNode = ParseStatement(currentNode, ref currentNodeMarker);
RootNode.InsertFinal(currentNode);
if (currentNodeMarker != null)
RootNode.CurrentNode = currentNodeMarker;
}
else
{
RootNode.Insert("STATEMENTS", ParseStatement(currentNode, ref currentNodeMarker));
if (currentNodeMarker != null)
RootNode.CurrentNode = currentNodeMarker;
}
}
TokenIndex++;
}
RootNode.NullifyEmptyChildren();
}
/// <summary>
/// Parses a condition
/// EBNF: <condition> ::= <expression> | <expression> '==' <expression>
/// </summary>
private Node ParseCondition(Node node)
{
node = new Node();
node.Value = "CONDITION";
//LHS of condition
node.Left = ParseExpression(node, VariableType.Integer);
//Check for anything that isn't a variable or value
//If we don't find either, the condition has been parsed
if (!GetCurrentToken().Type.Equals(TokenType.Identifier) && !GetCurrentToken().Type.Equals(TokenType.Number)
&& !IsConditionalOperator(GetCurrentToken().Value))
return node;
//Get the equality comparators
node.Attributes[0] = GetCurrentToken().Value;
TokenIndex++;
//Get the RHS of the condition
node.Right = ParseExpression(node, VariableType.Integer);
return node;
}
/// <summary>
/// Traverses the tree in post-order (bottom-up) and optimises any arithmetic
/// </summary>
private void OptimiseOperatorsAndValuesOfNode(Node currentNode)
{
if (currentNode.Left != null)
{
OptimiseOperatorsAndValuesOfNode(currentNode.Left);
}
if (currentNode.Right != null)
{
OptimiseOperatorsAndValuesOfNode(currentNode.Right);
}
if ("+-*/".Contains(currentNode.Value))
{
//If left node and right node are both values
Int32 leftValue;
Int32 rightValue;
if (currentNode.Left != null && Int32.TryParse(currentNode.Left.Value, out leftValue) &&
currentNode.Right != null && Int32.TryParse(currentNode.Right.Value, out rightValue) &&
currentNode.Left.Attributes[1] == "VALUE" && currentNode.Right.Attributes[1] == "VALUE")
{
if (currentNode.Value == "+")
currentNode.Value = (leftValue + rightValue).ToString();
else if (currentNode.Value == "-")
currentNode.Value = (leftValue - rightValue).ToString();
else if (currentNode.Value == "*")
currentNode.Value = (leftValue * rightValue).ToString();
else
currentNode.Value = (leftValue / rightValue).ToString();
currentNode.Attributes[1] = "VALUE";
currentNode.Attributes[0] = currentNode.Attributes[2] = "";
currentNode.Left = null;
currentNode.Right = null;
}
}
}
/// <summary>
/// To parse the expression
/// In EBNF: <expression> ::= <term>, { ‘+’|’-’, <term> };
/// </summary>
private Node ParseExpression(Node node, VariableType type)
{
CheckIfOutOfTokens();
//A numeric expr can be made up of terms, but a string expr is made up of string factors only
//(we can't multilply, divide or subtract strings)
if (type.Equals(VariableType.Integer))
node = ParseTerm(node);
else if (type.Equals(VariableType.Boolean))
node = ParseCondition(node);
else
node = ParseFactor(node, VariableType.String);
//While a valid operator is found
while (GetCurrentToken().Value.Equals("+") || (type.Equals(VariableType.Integer) &&
GetCurrentToken().Value.Equals("-")))
{
if (node == null)
node = new Node();
Node clonedNode = node.Clone();
node.Left = clonedNode;
node.Value = GetCurrentToken().Value;
TokenIndex++;
if (type.Equals(VariableType.Integer))
node.Right = ParseTerm(node.Right);
else
node.Right = ParseFactor(node.Right, VariableType.String);
}
return node;
}