tests/testthat/testFindSub.R
In hyginn/rete: Network Analysis for Cell- and Systems Biology
#findSub Tests
testthat::context("Test1 for Network Discovery")
testthat::test_that("2 Networks form of expected size, have correct heat
score assigned", {
#Note that Lines 11:47 are repeated in each context so that
#variables are reassigned, and do not carry over from previous contexts
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#Here lies the border of the common variable setup.
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 2 #setting minOrd to 2
#below, applying findSub function
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog=FALSE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#below 4 lines manually constructing expected networks
expectNet1Edges<-BigNetEdges[1:10,]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[1:4,]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
expectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID),]
expectNet2Verts<-BigNetVertices[5:7,]
#below 6 lines loading iGraph object contents from output into data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(length(outputGraphs),2)
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
})
testthat::context("findSub Test 2: testing effect of raising minOrd to 4")
testthat::test_that("1 subnetwork formed, with appropriate aggregate heat score, edges, and vertices", {
#Here we only expect 1 subnetwork to form, with ABC1, DEF2, GHI4, and JKL3
#again, the setup of variables
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end assignment of common variables
#begin: changed inputs for test 2
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 4
#We should only get 1 subnetwork this time
#run findSub
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog = FALSE,silent=FALSE)
#method: method of edge removal; EGG: iGraph object containing
#network with assigned 'influence' to each edge; noLog: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#Note:
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
#below 2 lines manually constructing expected networks
expectNet1Edges<-BigNetEdges[1:10,]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[1:4,]
#below 3 lines loading iGraph object contents from output into data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
#checking the following conditions:
#number of vertices in expected and produced (netwk1) graphs consistent
#every vertex in expected graph is present in produced graph
#every row and column of expected graph reproduced in produced graph
#only one subnetwork produced given input parameters and delta value
#checking: expected vertices and edges, expected number of networks formed
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(length(outputGraphs),1)
#checking: expected aggregate heat score
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
})
testthat::context("checking the effect of lowered delta and minOrd")
testthat::test_that("3 subnetworks formed with appropriate edges, vertices, heat scores", {
#Begin setup of variables common to all tests
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end setup of common variables
#begin setup of other inputs for test
#delta set to 3.1, minOrd to 2
igraph::graph_attr(EGG,"delta")<-3.1 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 2
#Here, we expect three subnetworks to form. UVW8 and XYZ9 have directed edges going each way of Influence 3.5
#Thus, they were eliminated in Test1 with a delta of 4
#run findSub
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog = TRUE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#Note
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
#below 6 lines: set up expected edges and vertices
expectNet1Edges<-BigNetEdges[1:10,]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[1:4,]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
expectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID),]
expectNet2Verts<-BigNetVertices[5:7,]
expectNet3Edges<-BigNetEdges[21:22,]
expectNet3Edges<-expectNet3Edges[order(expectNet3Edges$edgeID),]
expectNet3Verts<-BigNetVertices[8:9,]
#below 9 lines: set up edges and vertices from output as data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
netwk3<-outputGraphs[[3]]
netwk3Edges<-igraph::as_data_frame(netwk3,what="edges")
netwk3Verts<-igraph::as_data_frame(netwk3,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
edgeCompar3<-vector(length=ncol(expectNet3Edges))
if (ncol(expectNet3Edges)==ncol(netwk3Edges) && nrow(expectNet3Edges)==nrow(netwk3Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet3Edges[,i]==netwk3Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet3Edges)) {
edgeCompar3[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
vertCompar3<-vector(length=nrow(expectNet3Verts))
for (i in 1:nrow(expectNet3Verts)) {
ind3<-grepl(expectNet3Verts$HGNC_symbol[i],netwk3Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk3Verts$Gene_Score[i]==expectNet3Verts$Gene_Score[ind3]) {
vertCompar3[i]<-TRUE
}
}
}
#do we have 3 networks present with expected edges and vertices
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(nrow(expectNet3Verts),nrow(netwk3Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar3)),nrow(expectNet3Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar3)),ncol(expectNet3Edges))
testthat::expect_equal(length(outputGraphs),3)
#do we have the expected heat scores for each network
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
testthat::expect_equal(igraph::graph_attr(netwk3,"aggHeatScore"),14)
})
testthat::context("Test 4, part 1: Do we fail to get a network created if not strongly connected")
testthat::test_that("1 subnetwork forms with expected edges and vertices and heat score", {
#Begin construction of common variables
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end construction of common variables
#Here, we will be assigning an edge influence attribute of 5 to edge GHI4->XYZ9. Since this does not
#create a 'strongly connected' network that liks XYZ9 & UVW8 to ABC1:GHI4,
#ABC1 DEF2 JKL3 GHI4 should form a network of 4 nodes and LMN5, OPQ6, RST7 should form
#a subnetwork of 3 nodes.
#Influence is reassigned to several edges such that after removing edges below the delta threshold,
#ABC1 through GHI4 and UVW8:XYZ9 are liked by 1 directional edge. This does not create a strongly
#connected graph that includes all of those vertices
Influence[(From=="UVW8")&(To=="XYZ9")]=5
Influence[(From=="XYZ9")&(To=="UVW8")]=5
Influence[(From=="GHI4")&(To=="XYZ9")]=5
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #remaking the data frame
inputGraph<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #and inputGraph
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 3 #setting minOrd to 3s
#below, applying findSub function
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog=FALSE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#below 6 lines: set up expected edges and vertices
expectNet1Edges<-BigNetEdges[c(1:10),]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[c(1:4),]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
exoectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID)]
expectNet2Verts<-BigNetVertices[5:7,]
#below 9 lines: set up edges and vertices from output as data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
#Have we got the expected subnetworks?
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(length(outputGraphs),2)
#and do they have the right aggregate heat scores?
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
})
testthat::context ("Test 4, Part 2: When we make the previously 'not strongly' connected network strongly connected,
does it get constructed?")
testthat::test_that("2 large networks found and are appropriately formed and scored", {
#once again setting up common variables
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end setup of common variables
#Like FINDSUB_Test4, but setting edge weights such that after edge removal, a path will exist between
#GHI4 and XYZ9 going in both directions, and consequently UVW8 and XYZ9 are included in a strongly connected
#subnetwork with ABC1:GHI4
#Here, we will be assigning an edge influence attribute of 5 to edge GHI4->XYZ9, and an
#edge influence attribute of 5 to edge XYZ0->GHI4
#ABC1 DEF2 JKL3 GHI4 and should form a network of 6 nodes and LMN5, OPQ6, RST7 should form
#a subnetwork of 3 nodes.
#Another of this test is to show that construction of subnetworks can involve grouping
#edges that are not grouped together i.e. consecutive in the list of edges.
Influence[(From=="UVW8")&(To=="XYZ9")]=5
Influence[(From=="XYZ9")&(To=="UVW8")]=5
Influence[(From=="GHI4")&(To=="XYZ9")]=5
Influence[(From=="XYZ9")&(To=="GHI4")]=5
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #remaking the data frame
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #and inputGraph
igraph::graph_attr(EGG,"EGGversion")<-"1.0"
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 3 #setting minOrd to 3s
#below, applying findSub function
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog=FALSE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#below 6 lines: set up expected edges and vertices
expectNet1Edges<-BigNetEdges[c(1:10,12,14,21:22),]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[c(1:4,8:9),]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
exoectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID)]
expectNet2Verts<-BigNetVertices[5:7,]
#below 9 lines: set up edges and vertices from output as data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
#Do we have our two expected networks, one of them of size 7, with appropriate
#edges and vertices?
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(length(outputGraphs),2)
#Do the networks constructed have appropriate aggreagate heat scores assigned?
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),36)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
})
testthat::context("Test 5: Checking all the checks and that incorrect input results in stoppage of function")
#setup EGG and other variables for test data one more time
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end common variable setup
#begin Test 5 tests
testthat::test_that("Checking the checks", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=NULL)) #silent is null
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=2)) #silent is numeric
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=1,
noLog = TRUE,silent="TRUE")) #silent with character
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = NULL,silent=FALSE)) #noLog is null
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = "NULL",silent=FALSE)) #noLog is Character
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = 43,silent=FALSE)) #noLog is numeric
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=-3,
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Blargh",EGG,minOrd=3, #character method
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method=NULL,EGG,minOrd=3, #null method
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method=2,EGG,minOrd=3, #numeric method
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=NULL, #null minord
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=2.2,
noLog = TRUE,silent=TRUE)) #minOrd with non-integer
testthat::expect_error(outputGraphs<-findSub(method=2,EGG,minOrd="Pizza",
noLog = TRUE,silent=TRUE)) #MinOrd w character
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=1,
noLog = TRUE,silent=TRUE)) #minOrd<2
})
#An added test: checking the version number
testthat::context("Checking that you can check version number")
igraph::graph_attr(EGG,"EGGversion")<-"1.2"
testthat::test_that("Wrong version number -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"EGGversion")<-1
testthat::test_that("EGGversion of type double -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"EGGversion")<-as.integer(1)
testthat::test_that("EGGversion of type integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::remove.graph.attribute(EGG,"EGGversion")
testthat::test_that("EGGversion nonexistent -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
testthat::context("Have a valid delta value")
igraph::graph_attr(EGG,"delta")<-"puppies"
testthat::test_that("Delta a character -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"delta")<-as.integer(1)
testthat::test_that("Delta an integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"delta")<-as.integer(1)
testthat::test_that("Delta an integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"delta")<-as.integer(1)
testthat::test_that("Delta an integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
testthat::test_that("No Delta-> Error message", {
igraph::remove.graph.attribute(EGG,"delta")
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
#finally, no EGG! (Expect an error)
testthat::test_that("No EGG -> error", {
igraph::graph_attr(EGG,"delta")<-4
testthat::expect_error(outputGraphs<-findSub(method="Leis",NULL,minOrd=3,
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis","EGG",minOrd=3,
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis",0,minOrd=3,
noLog = TRUE,silent=TRUE))
rm(EGG)
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
hyginn/rete documentation built on May 17, 2019, 9:16 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#findSub Tests
testthat::context("Test1 for Network Discovery")
testthat::test_that("2 Networks form of expected size, have correct heat
score assigned", {
#Note that Lines 11:47 are repeated in each context so that
#variables are reassigned, and do not carry over from previous contexts
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#Here lies the border of the common variable setup.
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 2 #setting minOrd to 2
#below, applying findSub function
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog=FALSE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#below 4 lines manually constructing expected networks
expectNet1Edges<-BigNetEdges[1:10,]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[1:4,]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
expectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID),]
expectNet2Verts<-BigNetVertices[5:7,]
#below 6 lines loading iGraph object contents from output into data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(length(outputGraphs),2)
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
})
testthat::context("findSub Test 2: testing effect of raising minOrd to 4")
testthat::test_that("1 subnetwork formed, with appropriate aggregate heat score, edges, and vertices", {
#Here we only expect 1 subnetwork to form, with ABC1, DEF2, GHI4, and JKL3
#again, the setup of variables
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end assignment of common variables
#begin: changed inputs for test 2
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 4
#We should only get 1 subnetwork this time
#run findSub
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog = FALSE,silent=FALSE)
#method: method of edge removal; EGG: iGraph object containing
#network with assigned 'influence' to each edge; noLog: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#Note:
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
#below 2 lines manually constructing expected networks
expectNet1Edges<-BigNetEdges[1:10,]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[1:4,]
#below 3 lines loading iGraph object contents from output into data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
#checking the following conditions:
#number of vertices in expected and produced (netwk1) graphs consistent
#every vertex in expected graph is present in produced graph
#every row and column of expected graph reproduced in produced graph
#only one subnetwork produced given input parameters and delta value
#checking: expected vertices and edges, expected number of networks formed
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(length(outputGraphs),1)
#checking: expected aggregate heat score
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
})
testthat::context("checking the effect of lowered delta and minOrd")
testthat::test_that("3 subnetworks formed with appropriate edges, vertices, heat scores", {
#Begin setup of variables common to all tests
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end setup of common variables
#begin setup of other inputs for test
#delta set to 3.1, minOrd to 2
igraph::graph_attr(EGG,"delta")<-3.1 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 2
#Here, we expect three subnetworks to form. UVW8 and XYZ9 have directed edges going each way of Influence 3.5
#Thus, they were eliminated in Test1 with a delta of 4
#run findSub
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog = TRUE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#Note
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
#below 6 lines: set up expected edges and vertices
expectNet1Edges<-BigNetEdges[1:10,]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[1:4,]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
expectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID),]
expectNet2Verts<-BigNetVertices[5:7,]
expectNet3Edges<-BigNetEdges[21:22,]
expectNet3Edges<-expectNet3Edges[order(expectNet3Edges$edgeID),]
expectNet3Verts<-BigNetVertices[8:9,]
#below 9 lines: set up edges and vertices from output as data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
netwk3<-outputGraphs[[3]]
netwk3Edges<-igraph::as_data_frame(netwk3,what="edges")
netwk3Verts<-igraph::as_data_frame(netwk3,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
edgeCompar3<-vector(length=ncol(expectNet3Edges))
if (ncol(expectNet3Edges)==ncol(netwk3Edges) && nrow(expectNet3Edges)==nrow(netwk3Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet3Edges[,i]==netwk3Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet3Edges)) {
edgeCompar3[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
vertCompar3<-vector(length=nrow(expectNet3Verts))
for (i in 1:nrow(expectNet3Verts)) {
ind3<-grepl(expectNet3Verts$HGNC_symbol[i],netwk3Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk3Verts$Gene_Score[i]==expectNet3Verts$Gene_Score[ind3]) {
vertCompar3[i]<-TRUE
}
}
}
#do we have 3 networks present with expected edges and vertices
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(nrow(expectNet3Verts),nrow(netwk3Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar3)),nrow(expectNet3Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar3)),ncol(expectNet3Edges))
testthat::expect_equal(length(outputGraphs),3)
#do we have the expected heat scores for each network
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
testthat::expect_equal(igraph::graph_attr(netwk3,"aggHeatScore"),14)
})
testthat::context("Test 4, part 1: Do we fail to get a network created if not strongly connected")
testthat::test_that("1 subnetwork forms with expected edges and vertices and heat score", {
#Begin construction of common variables
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end construction of common variables
#Here, we will be assigning an edge influence attribute of 5 to edge GHI4->XYZ9. Since this does not
#create a 'strongly connected' network that liks XYZ9 & UVW8 to ABC1:GHI4,
#ABC1 DEF2 JKL3 GHI4 should form a network of 4 nodes and LMN5, OPQ6, RST7 should form
#a subnetwork of 3 nodes.
#Influence is reassigned to several edges such that after removing edges below the delta threshold,
#ABC1 through GHI4 and UVW8:XYZ9 are liked by 1 directional edge. This does not create a strongly
#connected graph that includes all of those vertices
Influence[(From=="UVW8")&(To=="XYZ9")]=5
Influence[(From=="XYZ9")&(To=="UVW8")]=5
Influence[(From=="GHI4")&(To=="XYZ9")]=5
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #remaking the data frame
inputGraph<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #and inputGraph
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 3 #setting minOrd to 3s
#below, applying findSub function
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog=FALSE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#below 6 lines: set up expected edges and vertices
expectNet1Edges<-BigNetEdges[c(1:10),]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[c(1:4),]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
exoectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID)]
expectNet2Verts<-BigNetVertices[5:7,]
#below 9 lines: set up edges and vertices from output as data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
#Have we got the expected subnetworks?
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(length(outputGraphs),2)
#and do they have the right aggregate heat scores?
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),22)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
})
testthat::context ("Test 4, Part 2: When we make the previously 'not strongly' connected network strongly connected,
does it get constructed?")
testthat::test_that("2 large networks found and are appropriately formed and scored", {
#once again setting up common variables
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end setup of common variables
#Like FINDSUB_Test4, but setting edge weights such that after edge removal, a path will exist between
#GHI4 and XYZ9 going in both directions, and consequently UVW8 and XYZ9 are included in a strongly connected
#subnetwork with ABC1:GHI4
#Here, we will be assigning an edge influence attribute of 5 to edge GHI4->XYZ9, and an
#edge influence attribute of 5 to edge XYZ0->GHI4
#ABC1 DEF2 JKL3 GHI4 and should form a network of 6 nodes and LMN5, OPQ6, RST7 should form
#a subnetwork of 3 nodes.
#Another of this test is to show that construction of subnetworks can involve grouping
#edges that are not grouped together i.e. consecutive in the list of edges.
Influence[(From=="UVW8")&(To=="XYZ9")]=5
Influence[(From=="XYZ9")&(To=="UVW8")]=5
Influence[(From=="GHI4")&(To=="XYZ9")]=5
Influence[(From=="XYZ9")&(To=="GHI4")]=5
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #remaking the data frame
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #and inputGraph
igraph::graph_attr(EGG,"EGGversion")<-"1.0"
igraph::graph_attr(EGG,"delta")<-4 #setting delta parameter (score threshold for edge inclusion)
minOrd <- 3 #setting minOrd to 3s
#below, applying findSub function
outputGraphs<-findSub(method="Leis",EGG,minOrd,noLog=FALSE,silent=FALSE)
#MethType: method of edge removal; Thresh: influence threshold; inputgraph: iGraph object containing
#network with assigned 'influence' to each edge; logResults: do we save a log of the process? ; silent
#if false, print what appears in the process log while the process is in action, as well as additional
#material (e.g. list edges removed, vertices removed)
#below 6 lines: set up expected edges and vertices
expectNet1Edges<-BigNetEdges[c(1:10,12,14,21:22),]
expectNet1Edges<-expectNet1Edges[order(expectNet1Edges$edgeID),]
expectNet1Verts<-BigNetVertices[c(1:4,8:9),]
expectNet2Edges<-BigNetEdges[c(15:16,19:20),]
exoectNet2Edges<-expectNet2Edges[order(expectNet2Edges$edgeID)]
expectNet2Verts<-BigNetVertices[5:7,]
#below 9 lines: set up edges and vertices from output as data frames
netwk1<-outputGraphs[[1]]
netwk1Edges<-igraph::as_data_frame(netwk1,what="edges")
netwk1Verts<-igraph::as_data_frame(netwk1,what="vertices")
netwk2<-outputGraphs[[2]]
netwk2Edges<-igraph::as_data_frame(netwk2,what="edges")
netwk2Verts<-igraph::as_data_frame(netwk2,what="vertices")
#edgeCompar: vector of falses with length(ncol(expectNetXEdges)). If all column in the expected
#dataset are identical to that of the test dataset (netwkXEdges), edgeCompar becomes a vector of
#TRUES with the same length. If this condition is fulfilled, then we know the output for
#the given network is identical to the expected edges.
edgeCompar1<-vector(length=ncol(expectNet1Edges))
if (ncol(expectNet1Edges)==ncol(netwk1Edges) && nrow(expectNet1Edges)==nrow(netwk1Edges)) {
for (i in 1:ncol(expectNet1Edges)) {
truRows<-expectNet1Edges[,i]==netwk1Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet1Edges)) {
edgeCompar1[i]<-TRUE
}
}
}
edgeCompar2<-vector(length=ncol(expectNet2Edges))
if (ncol(expectNet2Edges)==ncol(netwk2Edges) && nrow(expectNet2Edges)==nrow(netwk2Edges)) {
for (i in 1:ncol(expectNet2Edges)) {
truRows<-expectNet2Edges[,i]==netwk2Edges[,i]
sumTruRows<-sum(as.numeric(truRows))
if (sumTruRows==nrow(expectNet2Edges)) {
edgeCompar2[i]<-TRUE
}
}
}
#vertCompar: because of how unique() sorts things, you can get your vertices out of original order
#Therefore, the aim here is to see that for every vertex/gene score pair in the expected vertex dataframe
#is present in the output vertex data frame (by vertex dataframe I mean dataframe for vertices of
#a given network, expected or produced by the FINDSUB function)
#If that condition is fulfilled, AND the nrow of the expected vertex set is the same as that of the
#output vertex set, then you know that the two data frames of vertices have identical contents, and
# by extension the network's iGraph object's vertices/attributes were added correctly
vertCompar1<-vector(length=nrow(expectNet1Verts))
for (i in 1:nrow(expectNet1Verts)) {
ind1<-grepl(expectNet1Verts$HGNC_symbol[i],netwk1Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk1Verts$Gene_Score[i]==expectNet1Verts$Gene_Score[ind1]) {
vertCompar1[i]<-TRUE
}
}
}
vertCompar2<-vector(length=nrow(expectNet2Verts))
for (i in 1:nrow(expectNet2Verts)) {
ind2<-grepl(expectNet2Verts$HGNC_symbol[i],netwk2Verts$name)
if (sum(as.numeric(ind1))>0) {
if (netwk2Verts$Gene_Score[i]==expectNet2Verts$Gene_Score[ind2]) {
vertCompar2[i]<-TRUE
}
}
}
#Do we have our two expected networks, one of them of size 7, with appropriate
#edges and vertices?
testthat::expect_equal(nrow(expectNet1Verts),nrow(netwk1Verts))
testthat::expect_equal(nrow(expectNet2Verts),nrow(netwk2Verts))
testthat::expect_equal(sum(as.numeric(vertCompar1)),nrow(expectNet1Verts))
testthat::expect_equal(sum(as.numeric(vertCompar2)),nrow(expectNet2Verts))
testthat::expect_equal(sum(as.numeric(edgeCompar1)),ncol(expectNet1Edges))
testthat::expect_equal(sum(as.numeric(edgeCompar2)),ncol(expectNet2Edges))
testthat::expect_equal(length(outputGraphs),2)
#Do the networks constructed have appropriate aggreagate heat scores assigned?
testthat::expect_equal(igraph::graph_attr(netwk1,"aggHeatScore"),36)
testthat::expect_equal(igraph::graph_attr(netwk2,"aggHeatScore"),17)
})
testthat::context("Test 5: Checking all the checks and that incorrect input results in stoppage of function")
#setup EGG and other variables for test data one more time
HGNCsymb<-c("ABC1","DEF2","GHI4","JKL3","LMN5","OPQ6","RST7","UVW8","XYZ9")
VertScores<-c(10,2,4,6,7,8,2,5,9)
BigNetVertices<-data.frame(HGNC_symbol=HGNCsymb,Gene_Score=VertScores)
From<-c("ABC1","ABC1","DEF2","JKL3","DEF2","DEF2","JKL3","GHI4","JKL3","GHI4","GHI4","GHI4","LMN5",
"XYZ9","LMN5","OPQ6","OPQ6","UVW8","OPQ6","RST7","UVW8","XYZ9")
To<-c("DEF2","JKL3","ABC1","ABC1","JKL3","GHI4","DEF2","DEF2","GHI4","JKL3","LMN5","XYZ9","GHI4",
"GHI4","OPQ6","LMN5","UVW8","OPQ6","RST7","OPQ6","XYZ9","UVW8")
edgeID<-character(length=length(From))
for (i in 1:length(edgeID)) {
edgeID[i]<-paste("BNE_",as.character(i),sep="")
}
#Influence vector created to assign influence levels to directed edges
Influence<-numeric(length(From))
Influence[1:length(Influence)]=5
Influence[(From=="GHI4")&(To=="LMN5")]=1
Influence[(From=="LMN5")&(To=="GHI4")]=2
Influence[(From=="GHI4")&(To=="XYZ9")]=1.5
Influence[(From=="XYZ9")&(To=="GHI4")]=2.5
Influence[(From=="OPQ6")&(To=="UVW8")]=2
Influence[(From=="UVW8")&(To=="OPQ6")]=3
Influence[(From=="UVW8")&(To=="XYZ9")]=3.5
Influence[(From=="XYZ9")&(To=="UVW8")]=3.5
#With a threshold influence of 3.5, LMN5 OPQ6 RST7 should form a network,
#ABC1,DEF2,GHI4,JKL3 should form a network
#When this is lowered to 3, LMN5,OPQ6,RST7,UVW8,and XYZ9 should form a network
#ABC1 DEF2 JKL3 GHI4 from same subnetwork
#Lower further to 2, and the 'discovered subnetwork' will include all vertices, but not all edges
#Lower influence threshold (delta) to 1, and every edge included
BigNetEdges<-data.frame(from=From,to=To,Influence=Influence,edgeID=edgeID) #edge data frame created for EGG
EGG<-igraph::graph_from_data_frame(BigNetEdges,directed=TRUE,vertices=BigNetVertices) #setup EGG graph
igraph::graph_attr(EGG, "EGGversion") <- "1.0" #set EGG version attribute
#end common variable setup
#begin Test 5 tests
testthat::test_that("Checking the checks", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=NULL)) #silent is null
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=2)) #silent is numeric
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=1,
noLog = TRUE,silent="TRUE")) #silent with character
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = NULL,silent=FALSE)) #noLog is null
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = "NULL",silent=FALSE)) #noLog is Character
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = 43,silent=FALSE)) #noLog is numeric
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=-3,
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Blargh",EGG,minOrd=3, #character method
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method=NULL,EGG,minOrd=3, #null method
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method=2,EGG,minOrd=3, #numeric method
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=NULL, #null minord
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=2.2,
noLog = TRUE,silent=TRUE)) #minOrd with non-integer
testthat::expect_error(outputGraphs<-findSub(method=2,EGG,minOrd="Pizza",
noLog = TRUE,silent=TRUE)) #MinOrd w character
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=1,
noLog = TRUE,silent=TRUE)) #minOrd<2
})
#An added test: checking the version number
testthat::context("Checking that you can check version number")
igraph::graph_attr(EGG,"EGGversion")<-"1.2"
testthat::test_that("Wrong version number -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"EGGversion")<-1
testthat::test_that("EGGversion of type double -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"EGGversion")<-as.integer(1)
testthat::test_that("EGGversion of type integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::remove.graph.attribute(EGG,"EGGversion")
testthat::test_that("EGGversion nonexistent -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
testthat::context("Have a valid delta value")
igraph::graph_attr(EGG,"delta")<-"puppies"
testthat::test_that("Delta a character -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"delta")<-as.integer(1)
testthat::test_that("Delta an integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"delta")<-as.integer(1)
testthat::test_that("Delta an integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
igraph::graph_attr(EGG,"delta")<-as.integer(1)
testthat::test_that("Delta an integer -> Error message", {
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
testthat::test_that("No Delta-> Error message", {
igraph::remove.graph.attribute(EGG,"delta")
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
#finally, no EGG! (Expect an error)
testthat::test_that("No EGG -> error", {
igraph::graph_attr(EGG,"delta")<-4
testthat::expect_error(outputGraphs<-findSub(method="Leis",NULL,minOrd=3,
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis","EGG",minOrd=3,
noLog = TRUE,silent=TRUE))
testthat::expect_error(outputGraphs<-findSub(method="Leis",0,minOrd=3,
noLog = TRUE,silent=TRUE))
rm(EGG)
testthat::expect_error(outputGraphs<-findSub(method="Leis",EGG,minOrd=3,
noLog = TRUE,silent=TRUE))
})
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.