logic-asts 1.5.4

Grammars and parsers for various logic

Logic ASTs: Abstract Syntax Trees for Logical Specifications

A collection of grammars, parsers, and abstract syntax trees (ASTs) for various logical formalisms.

The goal is to serve as a reusable foundation for academics and developers building tools that require logical expression parsing and manipulation, eliminating the need to create new parsers for each application.

Supported Logics

The library implements complete support for four logical systems:

1. Propositional Logic (base): Classical Boolean logic with conjunction, disjunction, negation, implication, equivalence, and exclusive-or operators.

2. Linear Temporal Logic (ltl): Temporal extension adding operators for reasoning about sequences of states over time. Includes Next (X), Eventually (F), Always (G), and Until (U) operators with optional time constraints.

3. Spatio-Temporal Reach-Escape Logic (strel): Combines temporal and spatial reasoning for multi-agent and distributed systems. Adds spatial operators (Everywhere, Somewhere, Reach, Escape) with distance constraints.

4. Signal Temporal Logic with Graph Operators (stl_go): Extends temporal logic with graph-based operators for specifying properties over multi-agent communication networks. Includes incoming and outgoing edge quantifiers with weight and count constraints.

Installation

Install from PyPI:

pip install logic-asts

Or if you'd like the latest main branch:

pip install git+https://github.com/anand-bala/logic-asts.git

Quick Start

Parse logical expressions:

import logic_asts

# Propositional logic
prop = logic_asts.parse_expr("(p & q) | ~r", syntax="base")

# Linear temporal logic
ltl = logic_asts.parse_expr("G(request -> F response)", syntax="ltl")

# Spatio-temporal logic
strel = logic_asts.parse_expr("G everywhere[0,5] ~obstacle", syntax="strel")

# Graph-based temporal logic
stl_go = logic_asts.parse_expr("in^[0,1]{E}_{c}[1,n] consensus", syntax="stl_go")

Create expressions programmatically:

from logic_asts.base import Variable, And, Or, Not
from logic_asts.ltl import Eventually, TimeInterval

p = Variable("p")
q = Variable("q")

# (p & q) | ~p
formula = (p & q) | ~p

# F[0,10] (p & q)
temporal = Eventually(p & q, TimeInterval(0, 10))

Evaluate propositional formulas:

from logic_asts.base import simple_eval

p = Variable("p")
q = Variable("q")
formula = p & q

# Evaluate: p=true, q=true -> Result: true
result = simple_eval(formula, {"p", "q"})

# Evaluate: p=true, q=false -> Result: false
result = simple_eval(formula, {"p"})

Type-safe tree traversal

The most convenient way to walk an expression tree is expr.iter_subtree(), but its return type is Iterator[Expr]. If you need mypy (or pyright) to know the precise element type, reach for one of the patterns below.

Pattern 1 -- typed iterator (preferred)

When you already hold a typed expression, call the matching iterator directly:

from logic_asts import ltl_expr_iter, parse_expr

expr = parse_expr("G (p U q)", syntax="ltl")  # LTLExpr[str]
for node in ltl_expr_iter(expr):              # Iterator[LTLExpr[str]]
    ...

Your type	Iterator to use
BoolExpr[AP] / BaseExpr[AP]	bool_expr_iter(expr)
LTLExpr[AP]	ltl_expr_iter(expr)
STRELExpr[AP]	strel_expr_iter(expr)
STLGOExpr[AP]	stlgo_expr_iter(expr)

All four functions also validate that the subtree contains no out-of-dialect nodes and raise TypeError at runtime if it does.

Pattern 2 -- type-guard then typed iterator

When the static type is just Expr (e.g. coming from an untyped API), narrow it first:

from logic_asts import Expr, is_ltl_expr, ltl_expr_iter

def process(expr: Expr) -> None:
    if is_ltl_expr(expr, str):            # narrows to LTLExpr[str]
        for node in ltl_expr_iter(expr):  # Iterator[LTLExpr[str]]
            ...

Pattern 3 -- filter to a single node class

Use the kind= argument on iter_subtree to visit only one concrete class. The full tree is still traversed, but only matching nodes are yielded:

from logic_asts import Variable, parse_expr

expr = parse_expr("G (p U q)", syntax="ltl")
for v in expr.iter_subtree(kind=Variable):  # Iterator[Variable[Any]]
    print(v.name)

Pattern 4 -- custom type subset with ExprVisitor

For arbitrary subsets of node types, construct an ExprVisitor directly. It both validates the tree and yields a typed iterator:

from logic_asts import ExprVisitor, And, Or, Not, Variable

for node in ExprVisitor((And, Or, Not, Variable), expr):
    # node: And | Or | Not | Variable[Any]
    ...

Contributing

Contributions are welcome. Please ensure all tests pass and documentation is updated for new features.

License

This project is licensed under the BSD 2-clause license.