RStudio Connect & Python

RStudio Connect is a publishing platform for the work your team creates in R and Python. This repository contains examples of Python content you can deploy to Connect, including:

Interactive apps

• Streamlit
• Dash
• Flask

Web APIs

• Flask
• FastAPI

Documents

• Jupyter Notebooks

Reticulate

Reticulate allows you to call Python from within an R session. This enables you to use models built in Python to power Shiny apps, visualize pandas dataframes with ggplot2, and much more.

Interactive apps

• Serving Sentiment Analysis with Plumber and spaCy
• Image Classification with PyTorch and Shiny

Documents

• Visualizing pandas dataframes with ggplot2

Getting Started

You can deploy examples from this repo to your Connect server via git-backed deployment, or clone the repository and deploy examples from their manifests with the rsconnect CLI.

If you want to explore an example more closely before deploying it:

• Clone this repository
• create a virtual environment in the folder you want to work in
• restore the needed packages into the virtual environment

$ cd flask-sentiment-analysis-api
$ python -m venv .venv
$ source .venv/bin/activate
$ python -m pip install -U pip setuptools wheel
$ python -m pip install -r requirements.txt

For reticulated content, set the RETICULATE_PYTHON environment variable to point to your virtual environment, by placing an .Renviron file in the folder containing the following:

RETICULATE_PYTHON=.venv/bin/python

• Run the examples locally
• Publish the examples with the rsconnect cli

rsconnect deploy api . -n <SERVER-NICKNAME>