# The original data is from:  https://www.kaggle.com/c/dogs-vs-cats/data

# Download the data file from kaggle.  
# You need to log into kaggle to be able to download the file.
# If you do not have a kaggle account I have uploaded it to the class Blackboard site under Course Materials.

# The name of the data file is: dogs-vs-cats.zip
# Unzip the file dogs-vs-cats.zip.
# Inside the directory dogs-vs-cats, unzip train.zip.  
# There should be pictures of cats and dogs in the train subdirectory.
# Be sure to check that the pictures are in the subdirectories before proceeding.

# Set the path to where the original data from kaggle was downloaded and unzipped.

original_dataset_dir <- "~/Downloads/kaggle_original_data/train"  # Update this line to where you have downloaded your data.
                                                                  # In Windows E:/kaggle_original_data/train
                                                                  # Note: We will only use pictures from the train directory

# Set the paths to where the  train, validation, and test  data will be copied to.  The pictures will be from the original data directory train.

base_dir <- "~/Downloads/cats_and_dogs_small_TEST"         # name of new directory where the pictures will be copied to
train_dir <- file.path(base_dir, "train")                  # subdirectory for the tranining data
validation_dir <- file.path(base_dir, "validation")        # subdirectory for the validation data
test_dir <- file.path(base_dir, "test")                    # subdirectory for the test data
train_cats_dir <- file.path(train_dir, "cats")             # cats subdirectory in the training subdirectory
train_dogs_dir <- file.path(train_dir, "dogs")             # dogs subdirectory in the training subdirectory
validation_cats_dir <- file.path(validation_dir, "cats")   # cats subdirectory in the validation subdirectory
validation_dogs_dir <- file.path(validation_dir, "dogs")   # dogs subdirectory in the validation subdirectory
test_cats_dir <- file.path(test_dir, "cats")               # cats subdirectory in the testing subdirectory
test_dogs_dir <- file.path(test_dir, "dogs")               # dogs subdirectory in the testing subdirectory

# Create all of the new directories.

dir.create(base_dir)
dir.create(train_dir)
dir.create(validation_dir)
dir.create(test_dir)
dir.create(train_cats_dir)
dir.create(train_dogs_dir)
dir.create(validation_cats_dir)
dir.create(validation_dogs_dir)
dir.create(test_cats_dir)
dir.create(test_dogs_dir)

# Copy all of the pictures to the new directories.
# Note I have added the overwrite = TRUE option so you can do the copy over and over if you want.

fnames <- paste0("cat.", 1:1000, ".jpg")
file.copy(file.path(original_dataset_dir, fnames), 
          file.path(train_cats_dir), overwrite = TRUE) 

fnames <- paste0("cat.", 1001:1500, ".jpg")
file.copy(file.path(original_dataset_dir, fnames), 
          file.path(validation_cats_dir), overwrite = TRUE)

fnames <- paste0("cat.", 1501:2000, ".jpg")
file.copy(file.path(original_dataset_dir, fnames),
          file.path(test_cats_dir), overwrite = TRUE)

fnames <- paste0("dog.", 1:1000, ".jpg")
file.copy(file.path(original_dataset_dir, fnames),
          file.path(train_dogs_dir), overwrite = TRUE)

fnames <- paste0("dog.", 1001:1500, ".jpg")
file.copy(file.path(original_dataset_dir, fnames),
          file.path(validation_dogs_dir), overwrite = TRUE)

fnames <- paste0("dog.", 1501:2000, ".jpg")
file.copy(file.path(original_dataset_dir, fnames),
          file.path(test_dogs_dir), overwrite = TRUE)

# Check

cat("total training cat images:", length(list.files(train_cats_dir)), "\n")
cat("total training dog images:", length(list.files(train_dogs_dir)), "\n")
cat("total validation cat images:", length(list.files(validation_cats_dir)), "\n")
cat("total validation dog images:", length(list.files(validation_dogs_dir)), "\n")
cat("total test cat images:", length(list.files(test_cats_dir)), "\n")
cat("total test dog images:", length(list.files(test_dogs_dir)), "\n")