Create ml10m.R

data-raw/ml10m.R

@@ -0,0 +1,81 @@
+### Code to prepare `ml10m` dataset goes here
+
+library(Matrix)
+library(stringr)
+
+raw_data <- tempfile()
+download.file(
+ "https://files.grouplens.org/datasets/movielens/ml-10m.zip", raw_data
+)
+
+## Movie rating data
+
+# UserID::MovieID::Rating::Timestamp
+ratings <- read.table(
+ text = readLines(unzip(raw_data, "ml-10M100K/ratings.dat")),
+ sep = ":", header = FALSE,
+ colClasses = c(
+ "integer", "NULL", "integer", "NULL",
+ "numeric", "NULL", "integer"
+ )
+)
+# head(ratings)
+# dim(ratings)
+all_UserID <- paste0("user_", sort(unique(ratings[, 1])))
+all_MovieID <- paste0("movie_", sort(unique(ratings[, 2])))
+ratings[, 1] <- paste0("user_", ratings[, 1])
+ratings[, 2] <- paste0("movie_", ratings[, 2])
+
+N <- length(all_UserID)
+M <- length(all_MovieID)
+
+rating_matrix <- Matrix::sparseMatrix(
+ i = match(ratings[, 1], all_UserID), # Row indices
+ j = match(ratings[, 2], all_MovieID), # Column indices
+ x = ratings[, 3],
+ dims = c(N, M),
+ symmetric = FALSE, triangular = FALSE,
+ index1 = TRUE
+)
+# head(rating_matrix)
+
+## Movies file description
+
+# MovieID::Title::Genres
+movies <- stringr::str_split_fixed(
+ readLines(unzip(raw_data, "ml-10M100K/movies.dat")),
+ pattern = "::", n = 3
+)
+# head(movies)
+movies[, 1] <- paste0("movie_", movies[, 1])
+movies <- movies[movies[, 1] %in% all_MovieID, ]
+
+all_genres <- c(
+ "Action", "Adventure", "Animation", "Children's",
+ "Comedy", "Crime", "Documentary", "Drama",
+ "Fantasy", "Film-Noir", "Horror", "Musical",
+ "Mystery", "Romance", "Sci-Fi", "Thriller",
+ "War", "Western"
+)
+
+# Create the movie-genre matrix for rows (# of movies) and columns (# of genres)
+movie_genre_matrix <- t(sapply(
+ movies[, 3],
+ FUN = function(x) {
+ as.integer(all_genres %in%
+ unlist(strsplit(x, split = "|", fixed = TRUE))
+ )
+ },
+ USE.NAMES = FALSE
+))
+colnames(movie_genre_matrix) <- all_genres
+
+ml10m <- list(
+ rating = rating_matrix,
+ genre = movie_genre_matrix
+)
+
+# Save the compressed data
+usethis::use_data(ml10m, overwrite = TRUE, compress = "xz")
+# # Compress the data
+# tools::resaveRdaFiles(paths = "data/ml10m.rda")