C. Lente

Chess in R with {chess}

{chess} is an opinionated wrapper for R around python-chess, an amazing library created by Niklas Fiekas. It allows users to read and write PGN files as well as create and explore game trees such as the ones seen in chess books.

The package is still maturing! So I very much encourage you to send your suggestions and bugs as issues in the package repository.

Installation

You can install the most recent version of {chess} from CRAN with:

install.packages("chess")

This should automatically install python-chess to your {reticulate} environment, but you can also explicitly do it with a convenient function:

chess::install_chess()

Example

To read an existing game, simply use read_game(). To explore it you can use forward()/back(), as well as variations()/variation() to see all variations listed for the next move and choose one of them.

library(chess)

# Read first game from My 60 Memorable Games
file <- system.file("m60mg.pgn", package = "chess")
fischer_sherwin <- read_game(file, n_max = 1)

# Starting position
fischer_sherwin
#>         <Start>
#> r n b q k b n r
#> p p p p p p p p
#> . . . . . . . .
#> . . . . . . . .
#> . . . . . . . .
#> . . . . . . . .
#> P P P P P P P P
#> R N B Q K B N R

# Navigate to 4. g3
fischer_sherwin %>%
  forward(7)
#>         <4. g3>
#> r . b q k b n r
#> p p . p . p p p
#> . . n . p . . .
#> . . p . . . . .
#> . . . . P . . .
#> . . . P . N P .
#> P P P . . P . P
#> R N B Q K B . R

# See all variations for 4...
fischer_sherwin %>%
  forward(7) %>%
  variations()
#>      <4... Nf6>          <4... d5>
#> r . b q k b . r    r . b q k b n r
#> p p . p . p p p    p p . . . p p p
#> . . n . p n . .    . . n . p . . .
#> . . p . . . . .    . . p p . . . .
#> . . . . P . . .    . . . . P . . .
#> . . . P . N P .    . . . P . N P .
#> P P P . . P . P    P P P . . P . P
#> R N B Q K B . R    R N B Q K B . R

# Follow the sideline
fischer_sherwin %>%
  forward(7) %>%
  variation(2)
#>       <4... d5>
#> r . b q k b n r
#> p p . . . p p p
#> . . n . p . . .
#> . . p p . . . .
#> . . . . P . . .
#> . . . P . N P .
#> P P P . . P . P
#> R N B Q K B . R

You can also create your own game with game() and add variations to it: the move() function adds moves as well as branches the tree of the game. Strings are converted to simple moves, while list()s behave exactly as parenthesis in PGN, creating a variation of the last move. Here you can see how to recreate a Scholar’s mate and some ways to avoid it:

# Scholar's mate and some defenses
scholars_mate <- game() %>%
  move("e4") %>%
  move("e5", list("e6"), list("d5")) %>%
  move("Bc4") %>%
  move("Nc6", list("Nf6")) %>%
  move("Qh5") %>%
  move("Nf6", list("g6", "Qf3", "Nf6")) %>%
  move("Qxf7")

# Last mainline move
scholars_mate
#>      <4. Qxf7#>
#> r . b q k b . r
#> p p p p . Q p p
#> . . n . . n . .
#> . . . . p . . .
#> . . B . P . . .
#> . . . . . . . .
#> P P P P . P P P
#> R N B . K . N R

Note that there are many ways to structure the input to move(). See vignette("chess") for more information.

{chess} also features many ways of seeing both the game as a whole and the board at a specific point in time.

# Print with unicode (doesn't look good on the web)
print(scholars_mate, unicode = TRUE)
#>      <4. Qxf7#>
#> ♜ . ♝ ♛ ♚ ♝ . ♜
#> ♟ ♟ ♟ ♟ . ♕ ♟ ♟
#> . . ♞ . . ♞ . .
#> . . . . ♟ . . .
#> . . ♗ . ♙ . . .
#> . . . . . . . .
#> ♙ ♙ ♙ ♙ . ♙ ♙ ♙
#> ♖ ♘ ♗ . ♔ . ♘ ♖

# Export the FEN of the board
fen(scholars_mate)
#> [1] "r1bqkb1r/pppp1Qpp/2n2n2/4p3/2B1P3/8/PPPP1PPP/RNB1K1NR b KQkq - 0 4"

# See the PGN after some move
str(back(scholars_mate, 3))
#> 2... Nc6 3. Qh5 Nf6 ( 3... g6 4. Qf3 Nf6 ) 4. Qxf7#

# Export the PGN after some move
pgn(back(scholars_mate, 3))
#> [1] "2... Nc6 3. Qh5 Nf6 ( 3... g6 4. Qf3 Nf6 ) 4. Qxf7#"

# Plot current board
plot(scholars_mate)

Motivation

python-chess served as the inspiration (and backbone) for {chess}. While the original version (and {rchess} for that matter) broadly handles “move generation, move validation” (with powerful classes and object-oriented syntax), {chess} focuses on making it easy to create and explore PGNs as trees.

By narrowing down the scope of the API, I believe the package becomes more intuitive to people who just want to quickly create shareable game analyses or easily explore other people’s games without having to resort to point and click software.

{chess}’s first use was helping me study Bobby Fischer’s My 60 Memorable Games. After some very difficult parsing, I was able to convert the whole book to PGN and upload it to lichess, but I still felt like the interface was too clumsy…