################################################################################
# 
#
#             ggplot2 and loon
#             
#             The loon.ggplot package
#
################################################################################
#  
#  
#  
#  ggplot2 is one of the most popular R packages
#  
#    - produces high-quality presentation graphics
#    - uses a pipeline style to build visualizations
#    - based on a variation on the original grammar of graphics (hence the gg)
#    - part of the tidyverse (with magrittr, dplyr, purrr, ...)
#
#
#  Packages used here
#  
library(loon)
library(loon.data)
library(ggplot2)
library(magrittr)
library(dplyr)
library(loon.ggplot)


########
#
#
#  ggplot2   
#  
#  
#  ggplots are built up from a series of components
#  that can sometimes be thought of as computational parts,
#  
pIris <- ggplot(data = iris, 
                mapping = aes(x = Sepal.Width, 
                              y = Sepal.Length, 
                              col = Species)) +
    geom_point(size = 5)
# 
#  which is not displayed until printed.
pIris
#  
#  You can add to it later
# 
pIris + ggtitle("Iris data")

#  It has some natural layers for statistics
# 
pIris + geom_density2d() 
#
#  Histograms
#  
hIris <- ggplot(data = iris, 
                mapping = aes(x = Sepal.Width, 
                              col = Species, 
                              fill = Species)) +
    geom_histogram(binwidth = 0.5)

hIris

#
#  One of its most useful features is facetting
pIris + geom_density2d() + facet_wrap(vars(Species))

#
# It can get complicated 
# 
#  SAheart from Hastie and Tibshirani "Elements of Statistical Learning"
#  Here from loon.data
#  
data("SAheart")

pSAheart <- ggplot(data = SAheart, 
                   mapping = aes(group = famhist, 
                                 col = famhist)) +
           geom_point(mapping = aes(x = age, y = chd)) +
           geom_smooth(mapping = aes(x = age, y = chd)) +
           xlab("Age in years") +
           ylab("Heart disease")
pSAheart

#
#  It is integral to the tidyverse and works
#  nicely with pipes
#  
#  Here the 1974 Motor Trends car data

mtcars %>%
    rename(transmission = am, weight = wt) %>%
    mutate(lp100km = (100 * 3.785411784) / (1.609344 * mpg)) %>%
    select(weight, lp100km) %>%
    ggplot(aes(x = weight, y = lp100km)) +
    geom_point(size = 5, col = "steelblue") +
    geom_smooth(method='lm', formula= y~x) +
    ylab("Litres per 100 kilometres") +
    ggtitle("Gas usage")  ->   # Note assignment occurs here
    pMTcars

pMTcars

########
#
# loon.ggplot
#  
#  
#  It would be nice if these were interactive plots though.
#  
#  That's what loon.ggplot() function is for.
#  
l_pIris <- loon.ggplot(pIris, linkingGroup = "iris")
l_hIris <- loon.ggplot(hIris, linkingGroup = "iris")
l_pSAheart <- loon.ggplot(pSAheart,
                          size = 10,
                          linkingGroup = "SAheart")
l_pMTcars <- loon.ggplot(pMTcars,
                         itemLabel = row.names(mtcars),
                         color = mtcars$am,
                         showItemLabels = TRUE,
                         linkingGroup = "mtcars")

l_pIrisDens <- loon.ggplot(pIris + geom_density2d(),
                           linkingGroup = "iris")
# And to grid
# 
plot(l_pIrisDens)
#
# facets
l_pIrisDens <- loon.ggplot(pIris + 
                            geom_density2d() +
                            facet_wrap(vars(Species)),
                            linkingGroup = "iris")


#
# loon.ggplot() is bi-directional
# 
p1 <- with(mtcars,
           l_plot(mpg, qsec, 
                  itemLabel = row.names(mtcars),
                  showItemLabels = TRUE,
                  linkingGroup = "mtcars")
           )

p1["selected"] <- mtcars$am == 1
loon.ggplot(p1)
loon.ggplot(p1) + geom_density2d()

loon.ggplot(loon.ggplot(p1) + geom_density2d(),
            glyph = "ctriangle")
# Which makes identifying which pointa are highlighted difficult!

loon.ggplot(l_facet(p1, by =  mtcars$am))


########
#
# gg_pipe()
# 
#  It would be nice to just add the loon conversion
#  to a pipeline.
#  
#  Unfortunately this fails.
#  loon.ggplot() needs a REALIZED ggplot to transform to loon
#  NOT the PROMISE of one.
#  ggplots are only a promise until they are printed.
#  
#  That's the job of gg_pipe.
#  
#  For example

mtcars %>%
    rename(transmission = am, weight = wt) %>%
    mutate(lp100km = (100 * 3.785411784) / (1.609344 * mpg)) %>%
    select(weight, lp100km) %>%
    # Have to wrap the ggplot in gg_pipe
    # which forces the ggplot to be realized.
    gg_pipe(
           {
            ggplot(aes(x = weight, y = lp100km)) +
                geom_point(size = 5, col = "steelblue") +
                geom_smooth(method='lm', formula= y~x) +
                ylab("Litres per 100 kilometres") +
                ggtitle("Gas usage")
            }
        )  %>%
    loon.ggplot(linkingGroup = "mtcars") ->  
    l_pMTcars2

#
#  Which admittedly seems less than ideal.
#  

########
#
# l_ggplot()
# 
#   Wouldn't be nicer if we could just use the 
#   grammar of ggplot2 without an actual ggplot?
#   
#   This seems like a fast way to get an interactive 
#   plot to play with.
#   
#   Simply use l_ggplot() in place of ggplot()
#   

mtcars %>%
    rename(transmission = am, weight = wt) %>%
    mutate(lp100km = (100 * 3.785411784) / (1.609344 * mpg)) %>%
    select(weight, lp100km) %>%
    # use l_ggplot() in place of ggplot()
    l_ggplot(aes(x = weight, y = lp100km)) +
    geom_point(size = 5, col = "steelblue") +
    geom_smooth(method='lm', formula= y~x) +
    ylab("Litres per 100 kilometres") +
    ggtitle("Gas usage") ->  
    l_pMTcars3

# which, just like ggplot, produces no plot until it is printed
l_pMTcars3
#
# To which we could add stuff
# 
l_pMTcars3 + geom_density2d()
#
#
########
#
#  NOTE
#  
#    loon.ggplot is still being polished and further fleshed out.
#  
#    There is lots more and in some cases some attention needs
#    to be paid to multiple data layers and linking.
#  
#    And more to come.
#    
#    And then there is loon.shiny ... another time.
#  
################################################################################