paths of zork - Visual Diary

Code

library(tidyverse)

Playsession highlights: found out I can run zork from VSC’s terminal, data capture needs to be done in googlesheets for ease of categorical variable handling.

So now I need better data capture than editing the raw tsv, and a better way of organising my code. Currently I have data in zork/data.

Already I see I need to split the data into two datasets. One where each row is uniquely defined by node

Code

read_tsv(
  paste0(here::here(), '/raw_data/zork-nodes.tsv'),
  show_col_types = FALSE) %>%
  head() |> # why has my keybinding for pipe stopped working? argh
  gt::gt()

perspective_location	region	interactivity
west of house	forest	mailbox
north of house	forest	NA
forest path	forest	NA
clearing	forest	pile of leaves
clearing	forest	NA
behind house	house	window

and one where each row is uniquely defined by edge.

Code

read_tsv(
  paste0(here::here(), '/raw_data/zork-edges.tsv'),
  show_col_types = FALSE) %>%
  head() |>
  gt::gt()

perspective_location	perspective_to_target	target_location	playthrough
west of house	north	north of house	2024-10-22
north of house	north	forest path	2024-10-22
forest path	north	clearin	2024-10-22
clearing	north	forest	2024-10-22
clearing	east	forest	2024-10-22
clearing	west	forest	2024-10-22

No doubt there will be other things. (Do I need a dataset for region?)

Instantiate the data

I will create a new dataset for this, but use the original Playsession as the template.

Code

# Create sets required for graph object
source('init-graph.R')

# Created an empty dataframe for edges

zork_edges

# A tibble: 0 × 3
# ℹ 3 variables: perspective_location <chr>, perspective_to_target <chr>,
#   target_location <chr>

Code

# Created an empty dataframe for nodes

zork_nodes

# A tibble: 0 × 4
# ℹ 4 variables: perspective_location <chr>, region <chr>, interactivity <chr>,
#   loot <chr>

Update the data

I could do something fancy, but I think I’ll go with add_row to update the data. Let’s add one row for the start.

Set levels of columns

Code

# Set levels of columns

source('set-levels.R')

# Define a custom print function
print_vector <- function(name, vector) {
  cat(name, ":\n", paste(vector, collapse = ", "), "\n\n")
}

Locations

Code

# Define the dimensions of the matrix (e.g., 2 rows and 3 columns)
n_cols <- 4
n_rows <- ceiling(length(locations) / n_cols)

# Create an empty matrix with the specified dimensions and fill it row-wise
# Suppressed warning because ragged length of vector, i.e., no guarantee it will
# fill out matrix
locations_matrix <- matrix(locations, nrow = n_rows, ncol = n_cols, byrow = TRUE)

# Print the matrix
locations_matrix

     [,1]                   [,2]            [,3]             [,4]            
[1,] "unexplored"           "west of house" "north of house" "forest"        
[2,] "clearing"             "forest path"   "behind house"   "south of house"
[3,] "kitchen"              "living room"   "attic"          "canyon view"   
[4,] "rocky ledge"          "cellar"        "inaccessible"   "east of chasm" 
[5,] "gallery"              "studio"        "the troll room" "maze"          
[6,] "east west passageway" "unexplored"    "west of house"  "north of house"

Regions

Code

# Print regions
print_vector("Regions", regions)

Regions :
 unexplored, forest, house, cellar

Directions

Code

# Print directions
print_vector("Directions", directions)

Directions :
 north, south, east, west, up, down

Set nodes

Code

unknown = 'unknown'

source('set-nodes.R')

zork_nodes

# A tibble: 21 × 4
   perspective_location region interactivity            loot                
   <chr>                <chr>  <chr>                    <chr>               
 1 west of house        forest solved: mailbox          leaflet             
 2 north of house       forest none                     none                
 3 forest               forest song bird?               none                
 4 clearing             forest unsolved: pile of leaves none                
 5 forest path          forest none                     none                
 6 behind house         forest window to house          none                
 7 south of house       forest none                     none                
 8 kitchen              house  sack                     water, lunch, garlic
 9 living room          house  rug, locked door to west sword, lantern      
10 attic                house  none                     rope, knife         
# ℹ 11 more rows

Set edges

Code

source('set-edges.R')

zork_edges

# A tibble: 29 × 3
   perspective_location perspective_to_target target_location
   <chr>                <chr>                 <chr>          
 1 west of house        north                 north of house 
 2 west of house        west                  forest         
 3 forest               north                 clearing       
 4 clearing             west                  forest         
 5 clearing             south                 forest path    
 6 forest path          south                 north of house 
 7 north of house       east                  behind house   
 8 behind house         south                 south of house 
 9 south of house       south                 forest         
10 behind house         west                  kitchen        
# ℹ 19 more rows

Instantiate the plot

I wonder if I can join the nodes dataframe to the nodes of the ggraph object?

Create graph object

Code

library(tidygraph)

Code

get_zork_graph <- function(zork_edges, zork_nodes, selected_region='all') {

  if (selected_region != 'all') {
    zork_nodes <- zork_nodes  |>  filter(region %in% c(selected_region, 'unexplored'))
  }

  zork_base_graph <- 
    zork_edges |> 
      mutate(
        from = perspective_location,
        to = target_location
      ) |>
      # does order matter?
      select(from, to, everything()) |>
      as_tbl_graph()
  # Join the nodes data frame to the nodes of the graph
  zork_graph <- zork_base_graph |>
    activate(nodes) |>
    # add attributes to nodes
    inner_join(zork_nodes, by = c("name" = "perspective_location"))

  return(zork_graph)

}

Code

zork_cellar <- get_zork_graph(zork_edges, zork_nodes |> filter(region %in% c('cellar', 'unexplored')))

zork_graph <- get_zork_graph(zork_edges, zork_nodes)

Plot the graph object

Code

library(ggraph)
install.packages("viridis")
library(viridis)
library(RColorBrewer)

n_regions <- length(regions)

# cols_f <- colorRampPalette(RColorBrewer::brewer.pal(8, 'Dark2'))
cols_f <- viridis::viridis_pal(option = "D")

zork_paths <- function(zork_graph) {
  zork_graph |>
    ggraph() +

    # configure edges
      geom_edge_fan(aes(label = perspective_to_target), 
        alpha = 0.2, 
        angle_calc = 'along',
        label_colour = 'grey',
        label_dodge = unit(2.5, 'mm'),
        colour = 'white',
        family = 'Courier',
        arrow = arrow(length = unit(2, 'mm')), 
        end_cap = circle(1, 'cm')) +

  # configure nodes

        geom_node_label(
          aes(label = name, fill=region),
          colour = 'white',
          family = 'Courier',
          alpha = 0.3
        ) + 

  # configure styling
    theme_minimal() +
    scale_edge_colour_manual(values = cols_f(n_regions)) + 
      scale_colour_manual(values = cols_f(n_regions)) +
      theme(
        plot.background = element_rect(fill = "#333333", colour = NA),
        legend.position = 'right',
        text = element_text(family = "Courier", size = 20),
        axis.ticks = element_blank(),
        axis.text = element_blank(),
        panel.grid = element_blank(),
        plot.title = element_text(colour = 'white'),
        plot.subtitle = element_text(colour = 'white'),
        plot.caption = element_text(colour = 'white'),
        legend.text = element_text(colour = 'white'),
        plot.margin = margin(10, 10, 10, 10)  # Adjust plot margins
     ) + 
    scale_x_continuous(expand = expansion(mult = 0.2)) +  # Add space around x-axis
    scale_y_continuous(expand = expansion(mult = 0.2)) +  # Add space around y-axis

  # configure text
      labs(
        title = "Paths of Zork",
        subtitle = "Playsession 3 (2024-10-26)",
        caption = "Angle of edge determined by ggraph, not direction",
        x = "",
        y = "",
        colour = 'white'
      ) 
}

Full map

Code

zork_paths(zork_graph)

Code

ggsave('2024-10-26--graph.png', dpi=1000)

The cellar

Code

zork_paths(get_zork_graph(zork_edges, zork_nodes, 'cellar'))

Code

ggsave('2024-10-26--graph-cellar.png', dpi=1000)

The forest

Code

zork_paths(get_zork_graph(zork_edges, zork_nodes, 'forest'))

Code

ggsave('2024-10-26--graph-forest.png', dpi=1000)

Display the loot

Code

library(gt)

zork_nodes |>
  gt()

perspective_location	region	interactivity	loot
west of house	forest	solved: mailbox	leaflet
north of house	forest	none	none
forest	forest	song bird?	none
clearing	forest	unsolved: pile of leaves	none
forest path	forest	none	none
behind house	forest	window to house	none
south of house	forest	none	none
kitchen	house	sack	water, lunch, garlic
living room	house	rug, locked door to west	sword, lantern
attic	house	none	rope, knife
unexplored	unexplored	unknown	unknown
canyon view	forest	none	none
rocky ledge	forest	none	none
cellar	cellar	none	none
inaccessible	unexplored	none	none
east of chasm	cellar	none	none
gallery	cellar	painting	none
studio	cellar	note, 69 colours	none
the troll room	cellar	troll	none
maze	cellar	unknown	unknown
east west passageway	cellar	unknown	unknown

Finally! I can play!

Now to add rows and watch the plot update.

At 13 locations discovered

Next time will try using a google sheet and scraping it. Is too cumbersome coding. But will have to get workflow for setting levels of categorical variables.

At 21 locations

Don’t want to proceed without fixing the data entry. Especially now that I’m switching to command line zork, so much better.

Code

sudo snap install zork

write_tsv(zork_edges, ‘2024-10-26–edges.tsv’) write_tsv( zork_nodes, ‘2024-10-26–nodes.tsv’)

```54p