Parser generator · Python 3.13+

Write the grammar.
Get the parser.

A small Python parser generator with no dependencies. Give it a grammar, and it builds LR0, SLR or LR1 parsing tables, and parsers that compute values on every reduction, with your callbacks or the grammar's primitives.

Press start GitHub

syntactes is a parser generator written in Python. The name comes from the Greek συντάκτης (/sin'daktis/), meaning editor or composer.

Grammars as text or as objects. Write one lhs -> symbols rule per line, or build the Tokens and Rules yourself. See Grammars.
Three kinds of tables. LR0, SLR and LR1 generators, with a readable printout and a list of conflicts. See Parsing tables.
Callbacks on every reduction. Each rule's callback gets one token per right-hand side symbol, and its return value becomes the value of the left-hand side. See Parsing.
Primitives in the grammar. Tag a rule with a primitive, like add(1,3) % expr -> expr PLUS expr or int % term -> NUMBER, and the parser computes its value without a callback. See Primitives.
No dependencies. It uses only the standard library.

Here's a calculator in a few lines:

from syntactes import Grammar, Token
from syntactes.parser import LR1Parser

grammar = Grammar.from_text("""
expr -> expr PLUS term
expr -> term
term -> term TIMES NUMBER
term -> NUMBER
""")
start, add, pass_term, multiply, number = grammar.rules

parser = LR1Parser.from_grammar(grammar)


@parser.execute_on(start)
def result(expr):
    return expr.value


@parser.execute_on(add)
def on_add(expr, plus, term):
    return expr.value + term.value


@parser.execute_on(pass_term)
def on_term(term):
    return term.value


@parser.execute_on(multiply)
def on_multiply(term, times, number):
    return term.value * number.value


@parser.execute_on(number)
def on_number(number):
    return number.value


def lex(text):
    for word in text.split():
        if word == "+":
            yield Token("PLUS", True)
        elif word == "*":
            yield Token("TIMES", True)
        else:
            yield Token("NUMBER", True, int(word))
    yield Token.eof()


print(parser.parse(lex("2 * 3 + 4")))  # 10

Need a scanner to produce the tokens? The sibling project lectes scans text with named regexes, and its tokens map onto syntactes Tokens.

Write the grammar.Get the parser.

Write the grammar.
Get the parser.