h1. Maze

Maze generator and A* path finder.

It's horrible, and needs refactoring, but it works.

h2. Server/client

There is a ruby/sinatra server, which does all of the hard work, and exposes an api.

There is a JS/p5.js front end to it, which only does UI.

h2. API

- GET /api/v0.1/maze := Get the MML definition of the maze
- POST /api/v0.1/maze := Generate a new maze (returns MML)
- GET /api/v0.1/path := Get the PML definition of the path, if any
- GET /api/v0.1/path/valid := Returns the work "true" or "false" if a path has been found
- POST /api/v0.1/path := Generate a new path (returns nothing)

h2. Maze Markup Language / Path Markup Language

h3. MML

MML describes the maze in the following format:

bc. 40 30
10 11
22 13
1 3 4 5
30 23 2 2
[...]

* Line 1 is the width and height of the maze
* Line 2 is the starting x and y coordinates
* Line 3 is the target x and y coordinates
* Line 4 and beyond are untraversable "blocks":
** The first two numbers are the anchor x and y coordinates of the block
** The second two numbers are the width and height of the block
* All other nodes are considered traversable
* There is no diagonal path between two untraversable nodes touching at a corner

h3. PML

PML describes the path in the following format:

bc. 11 12
12 13
13 13
14 13
15 13
[...]

* Each line contains the x and y coordinates of the next step in the path
* The start and target do not appear in the path