Qt platform plugin issue Rstudio

Question

I am trying to do a seaborn heatmap plot via RStudio.

I usereticulate package in R.

Below is my code:

library(reticulate)
use_condaenv("python36", conda = "auto", required = FALSE)
os <- import("os")
os$listdir(".")
py_available()


sns <- import('seaborn')
plt <- import('matplotlib.pyplot')
pd <- import('pandas')


dat <- AirPassengers
# convert time series to data frame
dat <- data.frame(matrix(dat, ncol=frequency(dat), dimnames=dimnames(.preformat.ts(dat)) ))
dat
sns$heatmap(r_to_py(dat), fmt = "g", cmap = "viridis")
plt$show()

However, I receive the following error and my R session gets aborted when it reaches the seaborn heatmap line. What should I do to fix this error?

score 5 · Accepted Answer · answered Nov 12 '18 at 18:38

I had the same problem with RStudio daily build 1.2.114 and a Anaconda Python 3.7 environment where I had PyTorch and matplotlib installed.

I followed the instructions by @Sheperd, with the following changes, pointing to the environment where you have matplotlib installed; in my case pytorch37:

import matplotlib
import matplotlib.pyplot as plt
import numpy as np

import os
os.environ['QT_QPA_PLATFORM_PLUGIN_PATH'] = 'C:/Users/user_name/Anaconda3/envs/pytorch37/Library/plugins/platforms'

t = np.arange(0.0, 2.0, 0.01)
s = 1 + np.sin(2 * np.pi * t)

fig, ax = plt.subplots()
ax.plot(t, s)

ax.set(xlabel='time (s)', ylabel='voltage (mV)',
       title='About as simple as it gets, folks')
ax.grid()

plt.show()

Now, PyQt is found and RStudio doesn't crash anymore.

I'm also working on Windows and this solved the issue so commenting and upvoting as this was the best answer I found. — KevOMalley743, May 06 '21 at 20:41

score 4 · Answer 2 · answered Jun 13 '18 at 14:57

this seems like a duplicate question. I am using RStudio-1.2.679 with R-3.4.4 to write and edit Python code. I was having the exact same problem, I tried many solutions but nothing seemed to work. Finally I found the solution here -I take absolutely no credit for it. This is, at the top of your Python code (file with extension .py) where you import libraries include:

import os
os.environ['QT_QPA_PLATFORM_PLUGIN_PATH'] = 'C:/Users/myusername/AppData/Local/Continuum/Anaconda3/Library/plugins/platforms'

Notice that is how the path looks like in my PC, it may look different in yours.

Following this example:

import matplotlib
import matplotlib.pyplot as plt
import numpy as np
import os
os.environ['QT_QPA_PLATFORM_PLUGIN_PATH'] = 'C:/Users/myusername/AppData/Local/Continuum/Anaconda3/Library/plugins/platforms'

t = np.arange(0.0, 2.0, 0.01)
s = 1 + np.sin(2 * np.pi * t)

fig, ax = plt.subplots()
ax.plot(t, s)

ax.set(xlabel='time (s)', ylabel='voltage (mV)',
       title='About as simple as it gets, folks')
ax.grid()

plt.show()

This shows the plot in the "Plots" panel in RStudio. Best wishes!

score 1 · Answer 3 · edited Sep 11 '20 at 05:21

Here is a script to run directly in R, for Anaconda users:

library(reticulate)

use_condaenv(condaenv = "your conda env", required = T)

py_run_string('import os')
py_run_string("os.environ['QT_QPA_PLATFORM_PLUGIN_PATH'] = '/path/to/Anaconda3/Library/plugins/platforms'")

np = import("numpy",delay_load = T)
plt = import("matplotlib.pyplot",delay_load = T)

t = np$arange(0.01, 10.0, 0.01)

data1 = np$exp(t)
data2 = np$sin(2 * np$pi * t)

#These last lines must be run together:

fig = plt$figure(figsize=c(14,8))
plt$plot(1:10,1:10)
plt$show()

Enjoy!!!

score 0 · Answer 4 · answered May 11 '20 at 05:53

0

I'm not using Anaconda but using miniconda and reticulate for R-studio

So I used

import os

os.environ['QT_QPA_PLATFORM_PLUGIN_PATH'] = 'C:\Users\<user-name>\AppData\Local\r-miniconda\envs\r-reticulate\Library\plugins\platforms'`

It worked, Thank you @Shepherd and @F0nzie

answered May 11 '20 at 05:53

user10777048

46
2

Please do not repeat an already given answer, just upvote it (once you have enough reputation) – xiawi May 11 '20 at 06:55

Qt platform plugin issue Rstudio

4 Answers4

Linked