Compute comparative statistics for a maze
maze_stats.RdSummarizes the structural character of a maze: how many dead ends it has,
how long its corridors are, how branchy it is, how "wide" it is from
corner to corner, how much of the grid its shortest solution covers, and
how straight its corridors tend to run. Useful for comparing mazes
produced by different generate_maze() algorithms or
open_walls settings.
Usage
maze_stats(maze)
# S3 method for class 'maze_stats'
print(x, ...)
# S3 method for class 'maze_stats'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)Arguments
- maze
A
mazeobject returned bygenerate_maze().- x
A
maze_statsobject, as returned bymaze_stats().- ...
Additional arguments, currently unused.
- row.names, optional
Passed on for compatibility with
as.data.frame(); not otherwise used.
Value
An object of class maze_stats (a list, printed with its own
print method, and convertible to a one-row data frame with
as.data.frame()) containing:
- rows, cols
The maze's dimensions.
- algorithm
The generation algorithm recorded on
maze.- open_walls
The
open_wallsprobability recorded onmaze.- dead_ends, dead_end_pct
The number of degree-1 cells, and that count as a fraction of all cells.
- corridor_mean, corridor_max
The mean and maximum corridor length (in edges), across every corridor as defined above.
NAfor a 1x1 maze, which has no corridors.- degree1, degree2, degree3, degree4
The number of cells with exactly 1, 2, 3, or 4 open passages. These four counts sum to
rows * cols(every generated maze keeps every cell connected, so degree 0 does not occur except in a degenerate 1x1 maze).- diameter
The length, in steps, of the longest shortest path between any two cells in the maze – its graph diameter. Computed by a double-sweep BFS: farthest cell from an arbitrary corner, then farthest cell from that cell, then the shortest-path length between those two. This is exact for a perfect maze (a tree) but only an approximation, typically a slight underestimate, for a braided maze with loops.
- solution_length, solution_pct
The number of cells on the shortest path from
maze$starttomaze$end(viasolve_maze(maze, algorithm = "bfs")), and that count as a fraction of all cells.- river
A straightness measure in
[0, 1],NAif the maze has no corridor with at least two edges. Within a corridor, each step after the first either continues in the same direction as the step before it or turns;riveris the fraction of those steps that continue straight, averaged over every corridor with at least two edges (each corridor contributes one fraction, regardless of its length). Arivernear 1 means the maze is dominated by long, straight corridors; arivernear 0 means corridors turn at nearly every cell.
The print method returns x invisibly; as.data.frame()
returns a one-row data.frame with one column per element above.
Details
A corridor here is a maximal run of cells between two "special"
cells (a dead end, a branch, or an intersection – any cell whose number
of open passages, its degree, is not exactly 2), following the unique
route through the intervening degree-2 cells. In a perfect maze every
corridor has exactly two such endpoints. A braided maze (see
open_walls in generate_maze()) can additionally contain a
loop of degree-2 cells with no special cell on it at all; each such loop
is counted as one more corridor whose length is its own edge count.
Examples
maze <- generate_maze(15, 15, seed = 42)
maze_stats(maze)
#> <maze_stats> 15 x 15 (dfs)
#> dead ends: 25 (11.1%)
#> degree 1/2/3/4: 25 / 177 / 23 / 0
#> corridors: mean 4.77, max 28
#> diameter: 167
#> solution: 30 cells (13.3% of maze)
#> river: 0.358
# Braiding a maze (adding loops) tends to shorten its corridors and
# increase its branching.
braided <- generate_maze(15, 15, seed = 42, open_walls = 0.3)
as.data.frame(maze_stats(braided))
#> rows cols algorithm open_walls dead_ends dead_end_pct corridor_mean
#> 1 15 15 dfs 0.3 11 0.04888889 1.557377
#> corridor_max degree1 degree2 degree3 degree4 diameter solution_length
#> 1 7 11 102 93 19 29 12
#> solution_pct river
#> 1 0.05333333 0.4180791