CG:SHOP - Official pyutils25 for the 2025 Challenge on Non-Obtuse Triangulation

Utilities for verifying solutions of the CG:SHOP 2025 Competition. Feel free to use the code, but it is optimized for exact verification not for sampling or other purposes.

Installation

You can install the utilities via pip:

pip install cgshop2025-pyutils

Alternatively, you can install this package via source:

pip install --verbose .

You can also install it without the need to clone the repository:

pip install --verbose git+https://github.com/CG-SHOP/pyutils25

During the installation, CGAL and other dependencies will be downloaded and compiled. This can take a while but should happen mostly automatic. You need to have a C++ compiler (and the Python development environment) installed. Most systems will come with all dependencies preinstalled. Otherwise, you can for example install them on Ubuntu with the following command

sudo apt install build-essential python3-dev

You can test the installation with

pytest -s tests

Please check for updates of the utils frequently as we are still working on them.

Usage

You can use the utilities to verify solutions of the CG:SHOP 2025 Competition.

1. Instance Database: Utilize the instance database to iterate over all instances in the benchmark effortlessly.
2. Naive Solver: Compute a solution for each instance using the naive solver, which performs a constrained Delaunay triangulation without using Steiner points.
3. ZipWriter: Use the ZipWriter to create a valid submission file easily.
4. SolutionChecker: Verify the validity of the solution using the SolutionChecker.

from pathlib import Path

from cgshop2025_pyutils import (
    DelaunayBasedSolver,
    InstanceDatabase,
    ZipSolutionIterator,
    ZipWriter,
    verify,
)

# Load the instances from the example_instances folder. Instead of referring to the folder,
# you can also give a path to a zip file.
idb = InstanceDatabase("example_instances/")

# If the solution zip file already exists, delete it
if Path("example_solutions.zip").exists():
    Path("example_solutions.zip").unlink()

# Compute solutions for all instances using the provided (naive) solver
solutions = []
for instance in idb:
    solver = DelaunayBasedSolver(instance)
    solution = solver.solve()
    solutions.append(solution)

# Write the solutions to a new zip file
with ZipWriter("example_solutions.zip") as zw:
    for solution in solutions:
        zw.add_solution(solution)

# Verify the solutions
for solution in ZipSolutionIterator("example_solutions.zip"):
    instance = idb[solution.instance_uid]
    result = verify(instance, solution)
    print(f"{solution.instance_uid}: {result}")
    assert not result.errors, "Expect no errors."

Additional Features

The utils also expose a number of additional features that may be useful for developing your own solvers, as you will need to use exact arithmetic as provided by CGAL. However, if you want to use Python, we recommend to actually just fork this repository and adapt/extend the provided code to your needs. Check out the test cases (in ./tests/) for some examples. We may provide some further utilities in the near future, but for now, you can use the following classes:

FieldNumber

Represents an exact number (rational or floating-point) for precise arithmetic operations. If you are not sure about exact arithmetic, this single class will save you a lot of time. Just use it for all your arithmetic operations and you will be fine. For extracting the exact representation of the number, use the exact() method, which returns a string that is an exact representation and accepted by our verifier.

• FieldNumber(value: int | float | str): Initialize with an integer, float, or string.
• exact() -> str: Returns the exact representation of the number as a string.
• Supports +, -, *, / arithmetic and comparison operators (<, >, ==, !=).

Point

Represents a 2D point with exact coordinates.

• Point(x: FieldNumber, y: FieldNumber): Initialize a point with two FieldNumber coordinates.
• x() -> FieldNumber: Returns the x-coordinate.
• y() -> FieldNumber: Returns the y-coordinate.
• scale(factor: FieldNumber) -> Point: Returns a new point scaled by the given factor.
• Supports +, - operators for point arithmetic.

Segment

Represents a 2D segment between two Point objects.

• Segment(source: Point, target: Point): Initialize a segment with source and target points.
• source() -> Point: Returns the source point of the segment.
• target() -> Point: Returns the target point of the segment.
• squared_length() -> FieldNumber: Returns the squared length of the segment.
• does_intersect(other: Segment) -> bool: Checks if the segment intersects with another segment.

Polygon

Represents a simple polygon.

• Polygon(points: List[Point]): Initialize a polygon with a list of points.
• is_simple() -> bool: Checks if the polygon is simple (no self-intersections).
• contains(point: Point) -> bool: Checks if the polygon contains the given point.
• on_boundary(point: Point) -> bool: Checks if the point is on the polygon boundary.
• area() -> FieldNumber: Returns the area of the polygon.

VerificationGeometryHelper

Verifies geometric structures, detecting non-triangular faces, bad edges, and isolated points.

• VerificationGeometryHelper(): Initialize the geometry helper.
• add_point(point: Point) -> int: Adds a point to the geometry and returns its index.
• add_segment(index1: int, index2: int): Adds a segment between two points by their indices.
• get_num_points() -> int: Returns the number of points in the geometry.
• search_for_non_triangular_faces() -> Optional[Point]: Searches for any non-triangular faces.
• search_for_bad_edges() -> Optional[Segment]: Searches for edges with the same face on both sides.
• search_for_isolated_points() -> List[Point]: Returns a list of isolated points.
• count_obtuse_triangles() -> int: Counts the number of obtuse triangles in the geometry.

ConstrainedTriangulation

Handles 2D constrained Delaunay triangulation.

• ConstrainedTriangulation(): Initialize the triangulation.
• add_point(point: Point) -> int: Adds a point and returns its index.
• add_boundary(indices: List[int]): Adds a polygon boundary defined by the point indices.
• add_segment(index1: int, index2: int): Adds a segment between two points by their indices.
• get_triangulation_edges() -> List[Tuple[int, int]]: Returns a list of edges as tuples of point indices.

compute_convex_hull

Computes the convex hull of a set of points.

• compute_convex_hull(points: List[Point]) -> List[int]: Returns the indices of points on the convex hull.

intersection_point

Finds the intersection point of two segments, if they intersect.

• intersection_point(seg1: Segment, seg2: Segment) -> Optional[Point]: Returns the intersection point or None.

License

Our code is licensed under the MIT License. However, the code depends on CGAL, which can have implications for non-academic users. Please check the CGAL license for more information.

Reporting Issues

If you encounter any issues, please report them in the issue tracker. We will try to fix them as soon as possible.

Changelog

• 0.0.8: Adding file information to exceptions of ZipSolutionIterator
• 0.0.7: Fix for only accepting .solution.json (there was a , missing)
• 0.0.6: Fix for CGAL 6
• 0.0.5: Improved error messages for isolated points. Only accepting .solution.json as solution file extension.
• 0.0.4: Fixed bug with negative numbers in FieldNumber with string input
• 0.0.3: Allows negative Steiner points and snsure coordinates are converted to FieldNumber before Point construction