PipeMania Solver

Python project that solves an adaptation of the PipeMania puzzle using state-space search and problem-solving techniques from Artificial Intelligence.

---

Project Description

This project was developed for the Artificial Intelligence course. The objective is to read a PipeMania board from standard input, solve the puzzle, and print the solved board to standard output.

In this version of PipeMania, each board position contains a pipe piece, and the goal is to rotate the pieces so that the entire grid becomes a valid leak-free pipe network. According to the project statement, each instance can be assumed to have a unique solution.

The implementation is centered around:

• a Board representation for the puzzle grid
• a PipeManiaState class to represent search states
• a PipeMania problem class compatible with the provided search framework
• a pipe.py entry point that parses input, solves the puzzle, and prints the solution

---

Piece Types

The board includes four types of pieces:

• F — closing pieces
• B — branching pieces
• V — corner pieces
• L — straight pieces

Each piece also has an orientation, such as:

• C, B, E, D for directional pieces
• H, V for straight horizontal/vertical pieces

Examples:

• FB, VC, VD
• BC, BB, LV

---

Input Format

The program reads a square PipeMania instance from standard input.

Example:

FB VC VD
BC BB LV
FB FB FE

Each token identifies:

- the piece type
- its current orientation

## Output Format

The program prints the solved board in the same grid format.

### Example

```text
FB VB VE
BD BE LV
FC FC FC

---

Project Structure

.
├── README.md
├── pipe.py
├── search.py
├── utils.py
├── Visualizador/
│   ├── visualizer.py
│   ├── images/
│   └── test-*.txt / test-*.out
└── __pycache__/

---

Main Files

• pipe.py — main implementation of the PipeMania solver
• search.py — provided search algorithms framework
• utils.py — support utilities
• Visualizador/ — visualizer, test boards, expected outputs, and piece images

---

Approach

The project models PipeMania as a search problem.

A state represents a board configuration, and actions correspond to changing the orientation of pieces. The solver then applies a search technique to reach a valid goal state where:

• all pipes are correctly connected
• there are no leaks
• the grid satisfies the puzzle constraints

The official statement explicitly frames the project as a problem-solving/search task and provides the Problem abstraction and search algorithms as the intended base infrastructure.

---

Running the Project

Run the solver by redirecting an input board into pipe.py:

python3 pipe.py < input.txt

Example

python3 pipe.py < Visualizador/test-01.txt

---

Example Usage

Input

FB VC VD
BC BB LV
FB FB FE

Output

FB VB VE
BD BE LV
FC FC FC

---

Technical Highlights

• Developed in Python
• Models the puzzle as an AI search problem
• Uses state representation and action generation
• Reads from stdin and writes to stdout
• Compatible with the provided search framework
• Includes test boards and a visualizer for experimentation

---

Notes

• The repository includes a Visualizador/ folder with example boards, expected outputs, piece images, and a visualizer script.
• search.py and utils.py were part of the provided support code used in the course.
• The exact search strategy and heuristics may vary depending on the implementation in pipe.py.
• The focus of the project is on correctness and search-based problem solving rather than UI design.

---

Credits

This project was developed by:

• Martim Afonso
• Rita Melo