# To test if we can load a package, use the require() function.
# if require() returns False, we didn't load the package,
# and then we know we need to install it.
if(!require("pacman"))
{
    install.packages("pacman")
}

# To access another function in another .R files, use the source() function.
# For example, if you had a script saved [in the same folder] as other_script.R, 
# which contained a function() other_function(a,b) which returns a + (b/2), 
# you could call it like this:
source("other_script.r")
res <- other_function(1,2)
res

# Note you only need to call 'source()' once, then you can use the function multiple times.
res2 <- other_function(40,50)
res2

# first we load pacman package manager
if(!require("pacman"))
{
    install.packages("pacman")
}

# First load the datasets package,
p_load(datasets)

# then we can use p_data() function to get details about builtin packages 
p_data(datasets)

# We can now load some useful datasets that contain medically relevant datasets
p_load(carData)
p_load(caret)

print("carData, Companion to Applied Regression")
p_data(carData)
print("caret, Classification and Regression Training")
p_data(caret)

# first, load the R Input/Output package
if(!require("rio"))
{
    install.packages("rio")
}


# read in an xlsx Excel file
df1 <- import("data/ExampleDataClean.xlsx") %>% as_tibble()
print(df1)

# read an xlsx Excel file and get specific fields
df2 <- import("data/ExampleDataClean.xlsx") %>%
    as_tibble() %>%
    select(Field1, 
        Field3:Field6) %>%
print(df2)

# first, load the R Input/Output package
if(!require("rio"))
{
    install.packages("rio")
}

# now we read in the csv data
df1 <- import("data/ExampleDataClean2.csv")
print(colnames(df1))

# we can use slicing notation as [rows, columns] of a matrix-like object to 
# select certain data from the whole.
df1[1:5,1:4]

# if we don't want the slice to fail if we request more rows or columns than we have, we can use 
# the head() or tail() function to get the first couple or last couple rows.  This can be useful 
# when previewing data to see if it loaded correctly.

# first five rows, all columns
head(df1, 5)
# last 3 rows, all columns
tail(df1, 3)

# we often need to check number of rows or columns to slice or subselect data.
# do this with nrow() or ncol().  Can also check with dim() to return both
rs <- nrow(df1)
cs <- ncol(df1)

# To make unpacking multiple return statements easier, use the zeallot package.
if(!require("zeallot"))
{
    install.packages("zeallot")
}
# we can now use the magic pipe, %<-% to unpack values to a collection, c()
c(rows, cols) %<-% dim(df1)

sprintf(fmt = "rows = %d, cols = %d, dim = %dx%d", rs, cs, rows, cols)

# the apply function is useful to use a function on many elements of a matrix or dataframe
# without looping over each row/column.

# first, load the R Input/Output package
if(!require("rio"))
{
    install.packages("rio")
}

# To make unpacking multiple return statements easier, use the zeallot package.
if(!require("zeallot"))
{
    install.packages("zeallot")
}

# now we read in the csv data to dataframe
df1 <- import("data/ExampleDataClean2.csv")
df1

# Now if we wanted the max of each column, we can use apply():
maxes <- apply(df1, 2, max)
print('max of each column:')
maxes
# for columns, change the second parameter from '2' to '1'.  For all elements, c(1,2)


# if you want a list or vector result, use lapply() or sapply() respectively on just a part of the dataframe/matrix.
# example:
max_row_4 <- lapply(df1[4,], max)
print('max of row 4:')
max_row_4

min_column_3 <- sapply(df1[3:4,3], min)
print('min of rows 3 and 4, column 3:')
min_column_3


# You can also define your own function to apply.  Just define it, then reference it in apply.
inv_sq <- function(x){   
    # takes the square of the inverse of non-zero elements.
    ix <- x[!is.infinite(1.0/x)]
    return(1/(ix^(2)))
}
# apply to the last ten rows of column 4 of df1:
Val3_invsq <- sapply(tail(df1[,4],10), inv_sq)
Val3_invsq

# first, load the R Input/Output package
if(!require("rio"))
{
    install.packages("rio")
}

# Use the ggplot2 library.  This is also included in tidyverse.
require("ggplot2")
require("reshape2")

# now we read in the csv data to dataframe
df1 <- import("data/ExampleDataClean2.csv")
df1

# first, add an index column by row
df1$id <- 1:nrow(df1)

# extract subset of data columns to plot, including the index
df = data.frame("id" = df1$id,
                "Val2" = df1$Val2,
                "Val3" = df1$Val3)
# melt the data so each data point has a label and value tied to the index
d <- melt(df, id="id")

# name the melted data to make plotting easier
names(d) <- c('id', 'func', 'val')

# plot the data as lines all at once:
# ggplot() + 
#     geom_line(data=d, aes(x=id, y=val, color=func, group=func))




# alternatively, you could plot like this as well with fewer steps, 
# but much more manually needing to specify each line specifically:
ggplot(data = df1, aes(x=id)) +
    geom_line(aes(y = Val2, color = "Val2")) + 
    geom_line(aes(y = Val3, color = "Val3")) + 
    labs(x = "datapoints", y = "Values")