This system provides a multi-threaded, constraint-aware task execution and scheduling engine built with SWI-Prolog, extended by Python tools and MeTTa integration for flexible pipeline-based AI planning.
- Activate the development environment:
./from_venv.shOr, if you want to keep your shell:
source ./from_venv.sh-
Run the full pipeline (example):
./run_pipeline.sh
-
Load core modules in Prolog:
?- [prolog/reorder_executor, prolog/reorder_task_static, prolog/threaded_attvar].
-
Run tests:
?- [prolog/run_tests], run_tests.
prolog/reorder_task_static.plβ Task reorderingprolog/reorder_executor.plβ Dynamic and parallel execution engineprolog/threaded_attvar.plβ Attributed variables for cross-thread syncprolog/goal_dag.plβ DAG-based task modelingprolog/run_pipeline.plβ Top-level composition runner
test/β Full test suite for executors, annotations, DAGs, and sync logic
python/β Python modules for metta β prolog conversion and automationpython/metta_to_prolog.pyβ DSL converterpython/requirements.txtβ Python dependencies
example/greedy_example/β End-to-end GreedyChess planning demo
from_venv.shβ Smart virtualenv setup (detects shell, installs deps, symlinks pack)Makefileβ Common targets for building, running, and testingrun_pipeline.shβ Launch pipeline logic end-to-end
WALKTHROUGH.mdβ Full usage walk-through and test planREADME-ANNOTATOR.mdβ How annotators and goal analysis workTODO.mdβ Development roadmap and backlog
| Mode | Threads | Dynamic | Shared Vars | Best Use Case |
|---|---|---|---|---|
run_static_sequential/3 |
β | β | β | Deterministic, step-by-step |
run_static_parallel/3 |
β | β | β | Independent tasks, no var sharing |
run_dynamic_ready_loop/3 |
β | β | β | Tasks become ready at runtime |
run_dynamic_parallel_ready/3 |
β | β | β | Parallel with full readiness + sync |
- Requires SWI-Prolog β₯ 8.2
- Automatically symlinks the local folder as a SWI-Prolog pack
- Compatible with
.envfor configuration .venv/is created automatically byfrom_venv.shif missing
For detailed developer usage, see WALKTHROUGH.md. See AGENTS.md for how agents (human or AI) coordinate work and document reasoning.