AI9414

ai9414 is a Python teaching toolkit for interactive artificial intelligence demos. It opens small browser-based visualisations for search, logic, uncertainty, constraint satisfaction, planning, and foundation models, with simple installed commands that are easy to use in class.

It currently includes seventeen ready-to-run demos:

• labyrinth DFS search
• office delivery DFS search
• spatial graph DFS search
• spatial graph BFS search
• spatial graph greedy best-first search
• spatial graph A* search
• spatial graph uniform-cost search
• spatial graph branch-and-bound search
• propositional logic DPLL
• propositional resolution/refutation
• reasoning with uncertainty belief-state explorer
• foundation models tokenisation explorer
• foundation models next-token decoding explorer
• CSP map colouring
• CSP delivery time-slot assignment
• STRIPS planning
• Blocks World planning

Available demos

• labyrinth DFS search Start with ai9414 demo labyrinth
• office delivery DFS search Start with ai9414 demo delivery. This demo has fixed office layouts and an action-order control for DFS.
• spatial graph DFS search Start with ai9414 demo graph-dfs
• spatial graph BFS search Start with ai9414 demo graph-bfs
• spatial graph greedy best-first search Start with ai9414 demo graph-gbfs
• spatial graph A* search Start with ai9414 demo graph-astar
• spatial graph uniform-cost search Start with ai9414 demo graph-ucs
• spatial graph branch-and-bound search Start with ai9414 demo graph-bnb
• propositional logic DPLL Start with ai9414 demo logic-dpll
• propositional resolution/refutation Start with ai9414 demo logic-resolution
• reasoning with uncertainty belief-state explorer Start with ai9414 demo uncertainty
• foundation models tokenisation explorer Start with ai9414 demo foundation-models-tokenisation
• foundation models next-token decoding explorer Start with ai9414 demo foundation-models-decoding
• CSP map colouring Start with ai9414 demo csp-map
• CSP delivery time-slot assignment Start with ai9414 demo csp-delivery
• STRIPS planning Start with ai9414 demo strips
• Blocks World planning Start with ai9414 demo blocksworld

Why use it

• one installed package with a small command-line interface: ai9414 ...
• browser-based visualisations with packaged frontend assets and no separate web build step
• curated examples for search, logic, uncertainty, CSP, planning, tokenisation, and decoding
• graph-search examples use conventional S and G labels for start and goal nodes
• Python APIs for loading custom problems and launching demos directly from code
• replayable traces for classroom walkthroughs and static solution export
• automated tests plus contributor documentation

Quick start

Install from PyPI:

uvx ai9414 demo graph-bnb

Install directly from this repository:

uv sync
uv run ai9414 list
uv run ai9414 demo graph-bnb

If your environment does not put console scripts on PATH, the module entry point works too:

uv run python -m ai9414 demo graph-bnb

To see the curated example names for a demo:

uv run ai9414 list --examples graph-dfs

For contributors

uv sync --extra dev
uv run pytest
uv build
uv run ai9414 demo graph-bnb

Direct example scripts

The scripts under examples/ are still useful when working directly in the repository, but they are no longer the main installed interface.

To start the labyrinth example directly from the repository:

uv run python examples/labyrinth_demo.py

To start the delivery DFS example directly from the repository:

uv run python examples/delivery_demo.py

To start the spatial graph DFS example directly from the repository:

uv run python examples/graph_dfs_demo.py

To start the spatial graph BFS example directly from the repository:

uv run python examples/graph_bfs_demo.py

To start the spatial graph greedy best-first example directly from the repository:

uv run python examples/graph_gbfs_demo.py

To start the spatial graph A* example directly from the repository:

uv run python examples/graph_astar_demo.py

To start the spatial graph uniform-cost example directly from the repository:

uv run python examples/graph_ucs_demo.py

To start the spatial graph branch-and-bound example directly from the repository:

uv run python examples/graph_branch_and_bound_demo.py

To start the DPLL logic example directly from the repository:

uv run python examples/logic_dpll_demo.py

To start the resolution/refutation example directly from the repository:

uv run python examples/resolution_demo.py

To start the reasoning-with-uncertainty example directly from the repository:

uv run python examples/uncertainty_demo.py

To start the foundation models tokenisation example directly from the repository:

uv run python examples/foundation_models_demo.py

To start the foundation models decoding example:

uv run ai9414 demo foundation-models-decoding

To start the CSP map-colouring example directly from the repository:

uv run python examples/csp_demo.py

To start the CSP delivery scheduling example directly from the repository:

uv run python examples/delivery_csp_demo.py

To start the STRIPS planning example directly from the repository:

uv run python examples/strips_demo.py

To start the Blocks World planning example directly from the repository:

uv run python examples/blocksworld_demo.py

The same install is enough for the labyrinth live-Python stub as well. The student download now uses the built-in ai9414.search.run_labyrinth_solver(...) wrapper, so no separate Flask setup is required.

To export a backend-free replay bundle:

PYTHONPATH=src uv run python - <<'PY'
from pathlib import Path
from ai9414.search import SearchDemo

app = SearchDemo()
output_dir = app.export_solution_bundle(Path("build/search-solution"))
print(output_dir)
PY

Student-facing examples

Labyrinth example:

from ai9414.labyrinth import LabyrinthDemo

app = LabyrinthDemo()
app.load_example("small")
app.show()

Spatial graph DFS example:

from ai9414.graph_dfs import GraphDfsDemo

app = GraphDfsDemo()
app.load_example("small")
app.show()

Spatial graph BFS example:

from ai9414.graph_bfs import GraphBfsDemo

app = GraphBfsDemo()
app.load_example("small")
app.show()

Spatial graph greedy best-first example:

from ai9414.graph_gbfs import GraphGbfsDemo

app = GraphGbfsDemo()
app.load_example("small")
app.show()

Spatial graph A* example:

from ai9414.graph_astar import GraphAStarDemo

app = GraphAStarDemo()
app.load_example("small")
app.show()

Spatial graph branch-and-bound example:

from ai9414.search import SearchDemo

app = SearchDemo()
app.load_example("small")
app.set_options(playback_speed=1.0)
app.show()

Spatial graph uniform-cost example:

from ai9414.graph_ucs import GraphUcsDemo

app = GraphUcsDemo()
app.load_example("small")
app.show()

Visual DPLL example:

from ai9414.logic import DpllDemo

app = DpllDemo()
app.load_example("unit_chain")
app.show()

Reasoning with uncertainty example:

from ai9414.uncertainty import BayesFilterDemo

app = BayesFilterDemo.example("two_doors")
app.show()

In the browser, click Sense door, Move right, then Sense no door. The final belief has two equally likely locations, cells 3 and 7. Reset clears the interactive history and restores the initial belief and random stream.

Download Python stub supplies the three Bayes-filter functions to implement. Run the downloaded script to start the local solver, then click Solve with Python to replay its result. The same script supports --check and --demo for standalone work. See the uncertainty exercise guide for the complete workflow.

Set up a custom corridor:

from ai9414.uncertainty import BayesFilterDemo

app = BayesFilterDemo(
    cells=10,
    landmarks={"door": [2, 6]},
    initial_belief="uniform",
)
app.show()

Foundation models tokenisation example:

from ai9414.foundation_models import TokenisationExplorer

app = TokenisationExplorer()
app.load_example("simple_sentence")
app.show()

Custom tokenisation text:

from ai9414.foundation_models import TokenisationExplorer

app = TokenisationExplorer()
app.load_corpus("office_messages")
app.learn_merges(12)
app.set_text("Deliver parcel to Office A before 10:30.")
app.show()

Foundation models decoding example:

from ai9414.foundation_models import DecodingExplorer

app = DecodingExplorer()
app.load_example("lecture_activity")
app.show()

Visual CSP example:

from ai9414.csp import CSPDemo

app = CSPDemo(example="australia")
app.set_algorithm("backtracking_forward_checking")
app.show()

Custom CSP map:

from ai9414.csp import CSPDemo

app = CSPDemo()
app.load_map_problem(
    regions=["a", "b", "c", "d"],
    adjacency={
        "a": ["b", "c"],
        "b": ["a", "c", "d"],
        "c": ["a", "b", "d"],
        "d": ["b", "c"],
    },
    colours=["red", "green", "blue"],
)
app.set_variable_ordering("mrv")
app.show()

Delivery scheduling CSP example:

from ai9414.delivery_csp import DeliveryCSPDemo

app = DeliveryCSPDemo(example="weekday_schedule")
app.set_algorithm("backtracking_forward_checking")
app.show()

Custom delivery scheduling CSP:

from ai9414.delivery_csp import DeliveryCSPDemo

app = DeliveryCSPDemo()
app.load_delivery_problem(
    deliveries=[
        {"id": "a", "label": "Delivery A", "short_label": "A", "colour": "#c75b4a"},
        {"id": "b", "label": "Delivery B", "short_label": "B", "colour": "#4d79ab"},
    ],
    slots=[
        {"id": "slot_1", "label": "09:00", "order": 0},
        {"id": "slot_2", "label": "11:00", "order": 1},
    ],
    rooms=[
        {"id": "dock", "label": "Dock"},
    ],
    values=[
        {"id": "slot_1_dock", "slot": "slot_1", "room": "dock", "label": "09:00 @ Dock"},
        {"id": "slot_2_dock", "slot": "slot_2", "room": "dock", "label": "11:00 @ Dock"},
    ],
    domains={
        "a": ["slot_1_dock", "slot_2_dock"],
        "b": ["slot_2_dock"],
    },
    constraints=[
        {
            "kind": "precedence",
            "left": "a",
            "right": "b",
            "label": "A before B",
            "description": "Delivery A must happen before delivery B.",
        }
    ],
)
app.set_variable_ordering("mrv")
app.show()

Visual STRIPS planning example:

from ai9414.strips import StripsDemo

app = StripsDemo()
app.load_example("canonical_delivery")
app.show()

Visual Blocks World planning example:

from ai9414.strips import BlocksworldDemo

app = BlocksworldDemo()
app.load_example("sussman_anomaly")
app.show()

Custom STRIPS problem:

from ai9414.strips import StripsDemo, StripsProblem

problem = StripsProblem(
    rooms=["corridor", "mail_room", "office_a", "office_b", "lab"],
    robot_start="corridor",
    parcel_start="mail_room",
    keycard_start="office_a",
    locked_edge=("corridor", "lab"),
    door_locked=True,
    goal=[("at", "parcel", "lab")],
)

app = StripsDemo(problem=problem)
app.solve()
app.show()

Suggested STRIPS exercises:

• Move the keycard from office A to office B and predict how the plan changes.
• Start the robot in the mail room and explain why collecting the parcel too early is a bad idea.
• Unlock the lab door at the start and identify which action disappears from the plan.
• Move the locked door to the office B connection and compare the new plan structure.

Suggested Blocks World exercises:

• Compare the 3-, 4-, and 5-block tower examples and count how many temporary moves are needed.
• Use the Sussman anomaly example to explain why solving one goal first can force the planner to undo work.
• Compare the 5- and 6-disc Hanoi examples and relate the STRIPS actions to disc moves.

Custom CNF with DPLL:

from ai9414.logic import DpllDemo

app = DpllDemo()
app.load_cnf([
    ["A", "B"],
    ["~A", "C"],
    ["~B", "C"],
])
app.show()

Entailment with DPLL:

from ai9414.logic import DpllDemo

app = DpllDemo(mode="entailment")
app.load_kb(
    formulas=["A -> B", "B -> C", "A"],
    query="C",
)
app.show()

Curated Wumpus entailment examples:

from ai9414.logic import DpllDemo

app = DpllDemo(mode="entailment", example="wumpus_no_breeze")
app.show()

app = DpllDemo(mode="entailment", example="wumpus_forced_pit")
app.show()

Resolution/refutation proof attempts:

from ai9414.resolution import ResolutionDemo

app = ResolutionDemo(example="chain_refutation")
app.load_proof_attempt(
    clauses=[["~A", "B"], ["A"], ["~B"]],
    steps=[
        {"left": 1, "right": 2, "pivot": "A"},
        {"left": 4, "right": 3, "pivot": "B"},
    ],
)
app.show()

Live labyrinth solver wrapper:

from typing import Any
from ai9414.search import run_labyrinth_solver


def solve_dfs(labyrinth: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_labyrinth_solver(solve_dfs)

Live graph solver wrapper:

from typing import Any
from ai9414.search import run_graph_solver


def solve_dfs(graph: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_graph_solver(solve_dfs)

Live graph BFS solver wrapper:

from typing import Any
from ai9414.search import run_graph_bfs_solver


def solve_bfs(graph: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_graph_bfs_solver(solve_bfs)

Live graph A* solver wrapper:

from typing import Any
from ai9414.search import run_graph_astar_solver


def solve_astar(graph: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_graph_astar_solver(solve_astar)

Live graph UCS solver wrapper:

from typing import Any
from ai9414.search import run_graph_ucs_solver


def solve_ucs(graph: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_graph_ucs_solver(solve_ucs)

Live graph greedy best-first solver wrapper:

from typing import Any
from ai9414.search import run_graph_gbfs_solver


def solve_gbfs(graph: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_graph_gbfs_solver(solve_gbfs)

Live spatial graph branch-and-bound solver wrapper:

from typing import Any
from ai9414.search import run_weighted_graph_solver


def solve_dfbb(graph: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_weighted_graph_solver(solve_dfbb)

Live DPLL solver wrapper:

from typing import Any
from ai9414.logic import run_dpll_solver


def solve_dpll(problem: dict[str, Any], options: dict[str, Any]) -> dict[str, Any]:
    ...


if __name__ == "__main__":
    run_dpll_solver(solve_dpll)

Repository structure

src/ai9414/
  core/
  demo/
  graph_bfs/
  graph_dfs/
  graph_astar/
  graph_gbfs/
  graph_ucs/
  labyrinth/
  logic/
  search/
  frontend/
examples/
tests/
docs/

Acknowledgements

Thanks to Yi Luo yi.luo7@student.unsw.edu.au for reporting the uncertainty demo reset-state bug fixed in version 0.2.1.