############################
### Peak Search Function ###
############################

peaks <- function (series, span = 3, ties.method = "first")
{
  if ((span <- as.integer(span))%%2 != 1)
    stop("'span' must be odd")
  z <- embed(series, span)
  s <- span%/%2
  v <- max.col(z, ties.method = ties.method) == 1 + s
  pad <- rep(FALSE, s)
  result <- c(pad, v, pad)
  result
}

#### Peak Search Function Usage

# Create an extra column in the dataframe 'df', called 'peaks' containing
# either 'TRUE' or 'FALSE' depending on whether the value on any given row of 'temp'
# is a peak (based on a span of 3)
df$peaks <- peaks(df$temp, 3)

# Extract all rows where the value of 'temp' was a peak to a new dataframe 'dfpeaks'
dfpeaks <- subset(df, peaks=='TRUE') 


############################
### Linear Interpolation ###
############################

# This code assumes that 'df' is a dataframe containing 2 columns, 'time' & 'temp'.
# 'time' is assumed to be standardised to 'time since trap shut'.
# 'temp' is assumed to be standardised to temperature difference from baseline.

# Load Required Packages
require("zoo")
require(ggplot2)
require(Hmisc)

# Set Up Required Zoo Object to Data Frame Conversion Function
zoo.to.data.frame <- function(x, index.name="datetime") {
  stopifnot(is.zoo(x))
  xn <- if(is.null(dim(x))) deparse(substitute(x)) else colnames(x)
  setNames(data.frame(index(x), x, row.names=NULL), c(index.name,xn))
}

# Convert Dataframe to Zoo Object
dfzoo <- zoo(df)

# Make Relative Time Column Index for na.approx
index(dfzoo) <- dfzoo[,1]

# Interpolate Sequence of Temps for 0-120 sec, giving a single value for each second
dfzoo.interpolated <- na.approx(object=bleeding1.zoo.for.interp,
                                xout=seq(from=0, to=120, by=1))

# Convert Zoo Object back to Dataframe
dfzoo.interpolated.df <- zoo.to.data.frame(dfzoo.interpolated)


#################
### Mean Plot ###
#################

# This code assumes that all interpolated dataframes have been joined into a single
# dataframe 'df' containing 3 columns, 'time', 'temp', and 'test' (with 'test' indicating
# the identity of each individual temperature response curve)

ggplot(df,aes(x=time,y=temp)) +
  stat_summary(fun.data ="mean_cl_boot", mult=1, geom = "smooth", colour="red")