Python Networkx to R igraph conversion with errors

Question

I am attempting to translate the following Python code (for Networkx) into R (for igraph).

# Python code

import csv
import networkx as nx
import urllib

DG=nx.DiGraph()

dogcsv=csv.DictReader(open("dogs.csv","rU"))

dogid=0

nodeStr=['id', 'species',   'Location 1', 'Location 2']

nodeInt=['sumflu']

fluCols=['Influenza-1', 'Influenza-2', 'Influenza-3', 'Influenza-4',
'Influenza-5', 'Influenza-6', 'Influenza-7', 'Influenza-8', 'Influenza-9']

for flu in fluCols:
    DG.add_node(flu,typ="species")

for row in dogcsv:
    dogID='animal'+str(dogid)
    DG.add_node(dogID,typ='animal')
    for dog in nodeStr:
        DG.node[dogID][dog]=row[dog]
    for dog in nodeInt:
        DG.node[dogID][dog]=int(row[dog])
    for flu in fluCols:
        if str(row[flu])=='1':
            DG.add_edge(dogID,flu)
    dogid=dogid+1

The following is my attempt at a partial conversion of from Python to R.

# R code

library("data.table")
library("igraph")

dogs <- make_empty_graph()

dogcsv <- fread("dogs.csv")

dogid <- 0

nodeStr <- c("id", "species", "Location 1", "Location 2")

nodeInt <- "sumflu"

fluCols <- c("Influenza-1", "Influenza-2", "Influenza-3", "Influenza-4", 
"Influenza-5", "Influenza-6", "Influenza-7", "Influenza-8")

for(flu in fluCols) {
dogs[flu] <- dogs %>% add_vertices(8) %>% add_vertices(dogs[flu], typ = "host")
    }

This is the error code that I receive when I run the last portion of R code:

Error in intI(i, n = x@Dim[1], dn[[1]], give.dn = FALSE) : 
  no 'dimnames[[.]]': cannot use character indexing

In Python Networkx, the graph can be called out by graphname.function (ex. DG.add_node(dogID,typ='animal')).

Is there an equivalent in R's igraph?

Is there any way to call out the nodes in R's igraph without having to define the number of nodes?

Also, will you offer hints or suggestions on completing the translation to R?

Thank you.

The following contains the .csv data:

id,species,Location 1,Location 2,animal id,sumvirus,Influenza-1,Influenza-2,Influenza-3,Influenza-4,Influenza-5,Influenza-6,Influenza-7,Influenza-8,Influenza-9
YUI-4322,host1,Park 4,Park,MW 391,1,0,0,0,0,0,0,0,0,0
YUI-4321,host2,Park 4,Park,MW 390,0,0,0,0,0,0,0,0,0,0
YUI-4320,host2,Park 4,Park,MW 389,1,0,0,0,0,0,0,0,0,0
YUI-4319,host2,Park 4,Park,MW 388,0,0,0,0,0,0,0,0,0,0
YUI-7318,host2,Park 4,Park,MW 387,1,0,0,0,0,0,0,0,0,0
YUI-4317,host1,Park 4,Park,MW 386,0,0,0,0,0,0,0,0,0,0
YUI-4316,host2,Park 4,Park,MW 385,1,1,0,0,0,0,0,0,0,0
YUI-2315,host1,Shelter 2,Shelter,MV319,1,0,0,0,0,0,0,1,0,0
YUI-4314,host1,Shelter 2,Shelter,MV 332,1,0,0,0,0,0,0,1,0,0

You may be interested in [http://stackoverflow.com/questions/17433402/r-igraph-rename-vertices](this). There exists an igraph Python version; perhaps as an intermediate step you could translate into that. Though how could you leave networkx? ;-) — Arya McCarthy, May 16 '17 at 05:26

score 0 · Answer 1 · answered Jun 08 '17 at 22:37

I'm confused with what you are doing, but I can tell you that dogs is a graph object, and you are trying to do multiple things in a line that don't make sense together.

Calling dogs[flu] gives all vertices connected to flu.

add_vertices(8) is adding 8 new vertices to the graph.

add_vertices(dogs[flu], typ="host") is adding the same nodes into the graph that would already exist if dogs[flu] is callable.

Just add the nodes all before (add_vertices[fluCols], although it may need ints instead of strings) and then modify those nodes after.

Python Networkx to R igraph conversion with errors

1 Answers1