Ggplot2 hell with rpy2-2.0.7 + python 2.6 + r 2.11 (windows 7)

Question

I use rpy2-2.0.7 (I need this to work with Windows 7, and compiling binaries for newer versions of rpy2 is a mess) to insert a two-frame dataframe into r, create several layers in ggplot2, and display the image in < .png>.

I wasted countless hours fidgeting around with syntax; I managed to output the files that I need at some point, but (stupidly) did not notice and continued fidgeting with my code ...

I will sincerely appreciate any help; below is a (trivial) example of a demonstration. Many thanks for your help!!! ~ Eric Butter

---

import rpy2.robjects as rob
from rpy2.robjects import r
import rpy2.rlike.container as rlc
from array import array

r.library("grDevices")    # import r graphics package with rpy2
r.library("lattice")
r.library("ggplot2")
r.library("reshape")

picpath = 'foo.png' 

d1 = ["cat","dog","mouse"]
d2 = array('f',[1.0,2.0,3.0])

nums = rob.RVector(d2)
name = rob.StrVector(d1)

tl = rlc.TaggedList([nums, name], tags = ('nums', 'name'))
dataf = rob.RDataFrame(tl)

## r['png'](file=picpath, width=300, height=300)
## r['ggplot'](data=dataf)+r['aes_string'](x='nums')+r['geom_bar'](fill='name')+r['stat_bin'](binwidth=0.1)
r['ggplot'](data=dataf)
r['aes_string'](x='nums')
r['geom_bar'](fill='name')
r['stat_bin'](binwidth=0.1)
r['ggsave']()
## r['dev.off']()

* The result is just an empty image (181 b).

---

Here are a couple of common mistakes that R throws itself when I play ggplot2:

r['png'](file=picpath, width=300, height=300)
r['ggplot']()
r['layer'](dataf, x=nums, fill=name, geom="bar")
r['geom_histogram']()
r['stat_bin'](binwidth=0.1)
r['ggsave'](file=picpath)
r['dev.off']()

* RRuntimeError: Error: no layers in plot

r['png'](file=picpath, width=300, height=300)
r['ggplot'](data=dataf)
r['aes'](geom="bar")
r['geom_bar'](x=nums, fill=name)
r['stat_bin'](binwidth=0.1)
r['ggsave'](file=picpath)
r['dev.off']()

* RRuntimeError: : . : fill, x

Answer 1

rpy2 Nathaniel Smith rnumpy (. "API" rnumpy). :

from rnumpy import *

r.library("ggplot2")

picpath = 'foo.png' 
name = ["cat","dog","mouse"]
nums = [1.0,2.0,3.0]

r["dataf"] = r.data_frame(name=name, nums=nums)
r("p <- ggplot(dataf, aes(name, nums, fill=name)) + geom_bar(stat='identity')")
r.ggsave(picpath)

( , , , .)

- "R-" Python ipy_rnumpy. (. " IPython" rnumpy).

, R , . rpy2 rnumpy .

, ( ) , . , . . , -, - . :

In [59] R> long <- 1:20
Out[59] R>
  [1]   1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18
 [19]  19  20

In [60] R> long <- 1:100; 0
Out[60] R> [1] 0

( rnumpy, rnumpy.py, "from warnies import warn" "print" process_revents: " " warn ( "error in process_revents: ignored" )). , .)

Answer 2

dev(), , , () (, JD ) , dev.off().

from rpy2 import robjects                          
r = robjects.r                                                                                    
r.library("ggplot2")
robjects.r('p = ggplot(diamonds, aes(clarity, fill=cut)) + geom_bar()') 
r.ggsave('/stackBar.jpeg') 
robjects.r('print(p)')
r['dev.off']()

Answer 3

To make this a little easier when you need to draw more complex plots:

from rpy2 import robjects
from rpy2.robjects.packages import importr
import rpy2.robjects.lib.ggplot2 as ggplot2
r = robjects.r
grdevices = importr('grDevices')
p = r('''
  library(ggplot2)

  p <- ggplot(diamonds, aes(clarity, fill=cut)) + geom_bar()
  p <- p + opts(title = "{0}")
    # add more R code if necessary e.g. p <- p + layer(..)
  p'''.format("stackbar")) 
  # you can use format to transfer variables into R
  # use var.r_repr() in case it involves a robject like a vector or data.frame
p.plot()
# grdevices.dev_off()