@lusc/truth-table can parse truth statements and generate truth tables. Try it out.

A statement such as A AND B passed to generateTable will generate a table as seen below.

Though generating tables is what this package was made for, it can easily be used to only parse a truth statement and generate an AST representation of the statement. See parseOperation.

• AND
• OR
• XOR
• NOT
• IFF
• IFTHEN

See operator-aliases.md for all the aliases.

A variable is made up of at least one character. It must contain only latin letters (a-z). Variables are case-insensitive. and, or, and so on are reserved and will be treated as operators.

This extends Error and has from and to indicating were in the string there is an error. from is inclusive, to is exclusive. The message describes what the issue is and from and to can be used, for example, to highlight the problematic characters.

parseOperation('(a & b');
/*
IndexedError {
	message: "Unmatched opening parens at position 0.",
	from: 0,
	to: 1
}
*/

function parseOperation(input: string): AST;

If there are invalid characters, unmatched parens, two operators following each other or similar it throws IndexedError. For all other errors Error is thrown.

It returns AST.

operationToString takes an AST as input and returns a string representation of the AST. It doesn't return the original input but instead returns a standardised form.

Example:

const ast = parseOperation('a & b');
operationToString(ast); // === "(A ∧ B)"

function generateTable(input: string, includeSteps = true): ParsedTable;

type ParsedTable = {
	columns: readonly string[];
	rows: readonly boolean[][];
	ast: AST;
};

input is the operation that will be parsed.

By default, it will show each step to get to the result. For example with a statement like (a & b) | c, it will first contain the columns for the variables, then the column for A ∧ B, then the column for (A ∧ B) ∨ C.

If includeSteps is set to false, it will omit all steps and only return columns of the variables and the full statement.

The input is passed to parseOperation so it will throw the same errors.

Example:

generateTable('a != b');
// returns
{
	columns: ['A', 'B', 'A ↮ B'],
	rows: [
		[ true,  true,  false ],
		[ true,  false, true  ],
		[ false, true,  true  ],
		[ false, false, false ]
	],
	ast: {...}
}

This is an enum used to identify the type of operation.

It's implementation is pretty straightforward:

enum Operator {
	iff = 'iff',
	ifthen = 'ifthen',
	not = 'not',
	and = 'and',
	xor = 'xor',
	or = 'or',
}

This is a representation of any given truth statement. The shape of the AST is as follows.

A variable is made of any latin letters. Variables are case insensitive. Additionally, _ is permitted.

A variable matches the regex /^[a-z_]+$/i.

type AST = {
	type: 'variable';
	variable: string;
};

type AST = {
	type: 'operator';
	operator: Operator.not;
	values: [AST];
};

type AST = {
	type: 'operator';
	operator: Operator.and | Operator.or | ...; // All operators except Operator.not
	values: [AST, AST];
};