work_file.R

In kmob/s4_intro: worked examples from Genolini's "A (Not So) Short Introduction to S4" for R (2008)

source('classTrajectories.R')
source('classPartition.R')
source('classTrajectoriesBis.R')
source('classTrajPartitioned.R')
source('classPartitionFather.R')
source('classPartitionSimple.R')
source('classPartitionEval.R')



#### Section 4 - Declaring Classes ##########

new(Class = 'Trajectories', times=c(1,3,4))

new(Class = 'Trajectories', times=c(1,3),traj=matrix(1:4,ncol=2))

trajHospital_01 <- new(Class='Trajectories')

trajHospital_02 <- new(Class = 'Trajectories',
                  times =c(1,3,4,5),
                  traj=rbind (
                    c(15,15.1,15.2,15.2),
                    c(16,15.9,16,16.4),
                    c(15.2,NA,15.3,15.3),
                    c(15.7,15.6,15.8,16)
                  )
)

trajHospital_03 <- new(Class='Trajectories',
                  times=c(1:10,(6:16)*2),
                  traj=rbind(
                    matrix(seq(16,19,length=21),ncol=21,nrow=50,byrow=TRUE),
                    matrix(seq(15.8,18,length=21),ncol=21,nrow=30,byrow=TRUE)
                  )+rnorm(21*80,0,0.2)
)

trajHospital_02@times
trajHospital_02@times<-c(1,2,4,5) #using @ re-writes the values of the instance. DON'T use @!
trajHospital_02

test <- new('Trajectories')
test
length(test)

slotNames('Trajectories')
getSlots('Trajectories')
getClass('Trajectories')

#### Section 5 - Methods ##########
####### EXISTING METHODS ###########

# try plot method set for class Trajectories
par(mfrow = c(1,2))
plot(trajHospital_02)
plot(trajHospital_03)

# try print method set for class Trajectories
print(trajHospital_02)

# try show method set for class Trajectories
show(trajHospital_03)
# OR
trajHospital_03

# what happens with new instance with no arguments?
new('Trajectories')

####### NEW METHODS ###########

# try countMissing method for class Trajectories
countMissing(trajHospital_02)

# see the methods for a class
showMethods(class = 'Trajectories')

# see the details for a method
getMethod(f ='plot',
          signature = 'Trajectories')

# verify the method is defined for the class
existsMethod(f = 'plot', signature = 'Trajectories')

#### Section 6 - Constructors ##########

# see results of inspector check on instance validity
new(Class = 'Trajectories', times = 1:2, traj = matrix(1:2,ncol=2))
new(Class = 'Trajectories', times = 1:3, traj = matrix(1:2,ncol=2))

## reminder that accessing values by @ is a bad idea
trajHospital_04 <- new(Class = 'Trajectories', times = c(1), traj = matrix(1))
## using @ bypasses checking done on initialization, so the instance can now be broken
(trajHospital_04@times <- c(1,2,3))

# see the results of initialize on labeling and inspecting
new(Class = 'Trajectories', times = c(1,2,4,8), traj = matrix(1:8, nrow = 2))
new(Class = 'Trajectories', times = c(1,2,48), traj = matrix(1:8, nrow = 2))

# see the result of using initialize to define the args for setting a new instance
# args in 'new' don't have to match the slots in the class
new(Class = 'TrajectoriesBis', nbWeek = 4, BMIinit = c(16,17,15.6))

# see the results of the constructor funtion for general users that hides 'new'
trajectories(time = c(1,2,4), traj = matrix(1:6, ncol = 3))
# see the results of the constructor function that takes special arguments
# shows that a 'global' initializor can deal with multiple cases
regularTrajectories(nbWeek = 3, BMIinit = c(14,15,16))

###### Section 7 - Accessors ########
##### GET ACCESSOR ########
# times for a Trajectory
getTimes(trajHospital_02)
# traj for a Trajectory
getTraj(trajHospital_02)
# first column of traj for a Trajectory
getTrajInclusion(trajHospital_02)


##### SET ACCESSOR ########
setTimes(trajHospital_02) <- c(1,2,4,6)
setTimes(trajHospital_02) <- 1:4

###### Section 8 - Methods using several arguments #######
### new Partition instances ####
partHospital_02 <- new(Class="Partition",nbGroups=2,part=factor(c("A","B","A","B")))
partHospital_03 <- new(Class="Partition",nbGroups=2,part=factor(rep(c("A","B"),c(50,30))))

partTest <- new(Class="Partition",nbGroups=27,part=factor(rep(c("A","B"),c(50,30))))

partHospital_02
partHospital_03

par(mfrow=c(2,2))
### Plot for "Trajectory"
plot(trajHospital_02)
plot(trajHospital_03)
### Plot for "Trajectory" plus "Partition" 
plot(trajHospital_02,partHospital_02)
plot(trajHospital_03,partHospital_03)

###### Section 9 - Inheritance ########
### new TrajPartitioned instances ###
tdHospital_01 <- new('TrajPartitioned')

unclass(tdHospital_01)

partHospital_022 <- new(Class = "Partition",
                   nbGroups = 3,
                   part = factor(c("A","C","C","B")) 
                   )


getMethod("initialize","TrajPartitioned") #looks only at TrajPartitioned
existsMethod("initialize","TrajPartitioned") #looks only at TrajPartitioned 
hasMethod("initialize","TrajPartitioned") #looks up the inheritance for TrajPartitioned
selectMethod ("initialize", "TrajPartitioned") #tells which method up the inheritance chain is executing


tdHospital_02 <- new(Class = "TrajPartitioned",
                traj = trajHospital_02@traj,
                times = c(1,3,4,5),
                listPartitions = list(partHospital_02,partHospital_022) 
)

tdHospital_02
print(tdHospital_02)

### section 9.6 ###
# avoid callNextMethod issues with 'as'
print(as(tdHospital_01,"Trajectories"))

is(trajHospital_02,"TrajPartitioned")
is(tdHospital_02,"TrajPartitioned")
is(tdHospital_02,"Trajectories")

### Creation of empty TrajPartitioned
tdHospital_03 <- new("TrajPartitioned")

### Assignment of Trajectories to the attributes of TrajPartitioned 
##### as(objectSon,"ClassFather") <- objectFather ####
as(tdHospital_03,"Trajectories") <- trajHospital_03
tdHospital_03

### see object of class TrajPartitioned as a Partition ###
## specific instance of a class is called so doesn't belong in the class file
## can add a fifth argument, the function test: 
## it subordinates the transformation of class1 into class2 following a condition
## this setIs replaces the partition having the smallest number of groups by a new one
setIs(
  class1="TrajPartitioned",
  class2="Partition",
  coerce=function(from,to){
    numberGroups <- sapply(tdHospital_02@listPartitions,getNbGroups)
    Smallest <- which.min(-numberGroups)
    to<-new("Partition")
    to@nbGroups <- getNbGroups(from@listPartitions[[Smallest]])
    to@part <- getPart(from@listPartitions[[Smallest]])
    return(to) 
  },
  replace=function(from,value){
    numberGroups <- sapply(tdHospital_02@listPartitions,getNbGroups)
    smallest <- which.min(numberGroups)
    from@listPartitions[[smallest]] <- value
    return(from) 
  }
)

as(tdHospital_02,"Partition")
as(tdHospital_02,"Partition") <- partHospital_022

### section 9.8
### Virtual classes allow one method written for two classes
### since the method applies to the parent,
### both children run the same method (method looks up the inheritence chain)
a <- new("PartitionSimple",nbGroups=3,part=factor(LETTERS[c(1,2,3,2,2,1)]))
nbMultTwo(a)

b <- new("PartitionEval",nbGroups=5,part=ordered(LETTERS[c(1,5,3,4,2,4)])) 
nbMultTwo(b)

###### Section 10 - Internal modification of an object ########
### section 10.2 
### impute method creates a local object .Object, 
### it modifies the local trajectories (impute by average) and returns an object. 
### Impute(trajHospital_02) did not modify trajHospital_02.
impute(trajHospital_02) # local replacement of NA with the average
trajHospital_02 # globally, the data in trajHospital_02 is unchanged