Calling custom Python function in Shiny webapp

info-rchitect

Hi,

I have a Shiny webapp that takes user inputs. I need to call a custom Python package function and pass the user inputs from the webapp and return the function value (in this case a SQL string) to the webapp. Then I need to run that SQL from the webapp and return the dataset for download to CSV.

thx

---

Operating system used: Windows 10

Alexandru

Hi @info-rchitect
,

Using python code in rshiny app is possible with packages like: https://rstudio.github.io/reticulate/

Here is a sample

library(reticulate)
use_python("/home/dataiku/dss_11/code-envs/python/py36test/bin/python")
py_run_string('

import dataiku
client = dataiku.api_client()
def update_global_var(variable_name, variable_value):

    global_variables = client.get_variables()
    print(global_variables)

    global_variables[variable_name] = variable_value
    client.set_variables(global_variables)
    print(global_variables)
')

main <- import_main()
main$update_global_var(variable_name='some_variable', variable_value='some_value')
py_capture_output(main$update_global_var(variable_name='some_variable', variable_value='some_value'))

info-rchitect

AlexT,

Have you seen any performance, scalability issues when exchanging information between R Shiny and Python? The reason I ask is because, even though I love Python, DASH in Python is much more cumbersome and less elegant than Shiny IMO. However, R sucks for the dirty work under the hood. Hence, I hope this partnership will work well, if performance is not degraded too much.

thx