# install.packages("TDAmapper")
library("TDAmapper")

# The fastcluster package is not necessary.  By loading the
# fastcluster package, the fastcluster::hclust() function 
# automatically replaces the slower stats::hclust() function
# whenever hclust() is called.
install.packages("fastcluster") 
require(fastcluster) 

# The igraph package is necessary to view simplicial complexes
# (undirected graph) resulting from mapper1D().

library(igraph)

## examples modified from https://cran.r-project.org/web/packages/TDAmapper/README.html

######################################################
## Example 1: Using mapper1D to identify a figure 8 ##
######################################################

#plot 100 points from a figure 8
plot(x=2*cos(0.5*(1:100)), y=sin(1:100))

# Apply mapper to these 100 points
m1 <- mapper1D(
  distance_matrix = dist(data.frame( x=2*cos(0.5*(1:100)), y=sin(1:100) )),
  filter_values = 2*cos(0.5*(1:100)),
  num_intervals = 10,
  percent_overlap = 50,
  num_bins_when_clustering = 10)

# create and plot mapper graph
g1 <- graph.adjacency(m1$adjacency, mode="undirected")
plot(g1, layout = layout.auto(g1) )

##########################################################
## Example 2: Using mapper2D to identify an oval as S^1 ##
##########################################################

m2 <- mapper2D(
  distance_matrix = dist(data.frame( x=2*cos(0.5*(1:100)), y=sin(1:100) )),
  filter_values = list( 2*cos(1:100), sin(1:100) ),
  num_intervals = c(5,5),
  percent_overlap = 50,
  num_bins_when_clustering = 10)

g2 <- graph.adjacency(m2$adjacency, mode="undirected")
plot(g2, layout = layout.auto(g2) )

########################################################################
## Example 3: using mapper1D to identify two independent spirals as two line segments
###############################################################

# sample 200 points from two intertwined spirals
t <- runif(100, min=1, max=6.3) # theta
X <- data.frame( x = c( t*cos(t), -t*cos(t) ), y = c( t*sin(t), -t*sin(t) ) )
d <- dist(X)
plot(X[,1], X[,2])

# create variables used in mapper
filter <- X[,2] # height projection
num_intervals <- 10
percent_overlap <- 50
num_bins_when_clustering <- 10

# Apply mapper 
m3 <- mapper1D(
  distance_matrix = d, 
  filter_values = filter, 
  # num_intervals = 10, # use default
  # percent_overlap = 50, # use default
  # num_bins_when_clustering = 10 # use default
)

# create and plot mapper graph
g3 <- graph.adjacency(m3$adjacency, mode="undirected")
plot(g3, layout = layout.auto(g3) )


###########################
##  Create more datasets ##
###########################

# note the following is incorrect
knotdata <- read.csv("TDA/MAPPER/trefoilknot.txt", header = FALSE)

# note the following is the correct way to read in knot data
knotdata <- read.csv("TDA/MAPPER/trefoilknot.txt", sep = " ", header = FALSE)

# Grab first 2 columns for fun
2dim_data <- cbind(knotdata[1], knotdata[2])

# Create distance matrix
d <- dist(knotdata)
# Create filter
filter <- knotdata[1]
# run mapper
m3 <- mapper1D(
  distance_matrix = d, 
  filter_values = filter)

# create and plot mapper graph
g3 <- graph.adjacency(m3$adjacency, mode="undirected")
plot(g3, layout = layout.auto(g3) )

#### more data ####

footballdata <- read.csv("http://www.repole.com/sun4cast/stats/cfb20140906.csv")

data <- cbind(footballdata$ScoreOff, footballdata$RushAttOff)
plot(data)

######### WARNING #########
## you should clean your data before analyzing it