R + Python: The best of both worlds
‘R’ you ready for some Python??
Over the last few weeks I’ve been immersed in an internship where I’ve learned a TON about single cell genomics, trajectory inference methods, and Python. I’ll leave the fun details for another post but here I wanted to share something I’ve recently learned that’s been quite the game changer…
The first programming language I learned that opened me up to the wide world of computational biology was Python. Over the last few years, however, I transitioned into an avid R user. I’ve always held the misconception that the two languages were entirely independent, not easily integrated into a single analysis notebook. But thanks to RStudio, this is not true!!
The reticulate R package provides a comprehensive set of tools for interoperability between R and Python.
library(reticulate)
Now, as usual, I can run some R code –
#```{r}
some_R <- 'Hello World!'
print(some_R); class(some_R)
## [1] "Hello World!"
## [1] "character"
But with reticulate, I can also run some Python code –
#```{python}
some_Python = 'Good Morning!'
print(some_Python); type(some_Python)
## Good Morning!
## <class 'str'>
Accessing objects between languages
Objects in either language can be accessed in the other language thanks to the reticulate package.
Use R to call a variable made in Python:
# ```{r}
py$some_Python
## [1] "Good Morning!"
Use Python to call a variable made in R:
# ```{python}
r.some_R
## 'Hello World!'
Converting objects between languages
NOTE: When calling variables between environments using ‘r.’ and ‘py$’, the reticulate package translates the data type from one environment to another data type of the other environment. For example, a named list in R will be translated into a Python dictionary.
Besides simply accessing objects, you can also convert an R object into a Python object while still running R with r_to_py.
# ```{r}
# creating an R vector
rvector <- c(1,2,3,4,5)
# converting vector to Python list
rvector_converted <- r_to_py(rvector)
class(rvector); class(rvector_converted)
## [1] "numeric"
## [1] "python.builtin.list" "python.builtin.object"
Running a Python code chunk in R
There may be times when you want to run a code chunk written in Python within R. In this case, the py_run_string function from reticulate comes in handy. This function runs the Python code and saves the results into the Python environment. These results can then be accessed using ‘py$’ as before.
#```{r}
py_run_string("import numpy as np")
py_run_string("pyrun = np.array(range(5))")
py$pyrun
## [1] 0 1 2 3 4
Likewise, py_run_file allows you to run entire Python scripts in R.
And that’s it! Keeping it short and sweet for this one. In a later post I’ll showcase a bit more of the integration of R and Python in the genomics analysis I’ve recently learned.
