Getting Started

Install (development)

git clone https://github.com/pisanuw/ltms
cd ltms
python -m venv .venv && . .venv/bin/activate
pip install -e ".[dev]"      # add ".[dev,sat]" for the PySAT differential tests
pytest                        # everything green

The package targets Python 3.10+ and ships type hints (py.typed). The continuous integration build runs ruff, mypy --strict, and pytest on Python 3.10–3.13.

The reasoning engine (LTRE)

LTRE is the high-level entry point: a forward-chaining, pattern-directed reasoning engine sitting on top of the LTMS. You assert facts and implications, make retractable assumptions, and ask what is currently believed.

from ltms import LTRE

e = LTRE()
e.assert_(("or", ("p",), ("q",)))   # p v q
e.assert_(("not", ("p",)))          # ~p
e.is_true(("q",))                   # True  — forced by unit propagation

Assumptions are belief you can take back. Retracting one automatically un-derives everything that depended on it:

e.assume(("rain",), "guess")        # "guess" is the informant / reason
e.assert_(("implies", ("rain",), ("wet",)))
e.is_true(("wet",))                 # True
e.retract(("rain",), "guess")
e.is_unknown(("wet",))              # True  — belief revised, no manual cleanup

Queries are three-valued: a proposition is TRUE, FALSE, or UNKNOWN. The helpers is_true, is_false, is_known, and is_unknown read the label, inverting the sign for negated queries.

World models in .kb files

A world model can live in a .kb data file, separate from your Python. The DSL is a small Lisp-like language with directives for assertions, assumptions, implication rules, taxonomies, contradictions, queries, and self-checking expect lines:

# examples/kb/belief_revision.kb
rain         -> wet ground
sprinkler on -> wet ground
assume rain
expect wet ground true

from ltms.dsl import load_kb_file

load_kb_file("examples/kb/belief_revision.kb")   # runs it; `expect` lines self-check

See the Examples page for the full set of runnable demos and .kb knowledge bases, and the ltms.dsl source for the directive reference.

Where to go next

• Understand the design: the Architecture page maps every module and explains the core algorithms.
• Learn the theory: the Study Companion walks through the book chapter by chapter, with runnable commands against this code.
• See it work: the Examples and Exercises pages collect the runnable material.