Skip to content

Latest commit

 

History

History

bcra-app

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
app
app
 
 
func/LocalFunctionsProject
func/LocalFunctionsProject
 
 
README.md
README.md
 
 
deploy.sh
deploy.sh
 
 

README.md

Python + R API Example (Azure Function)

Etienne P Jacquot - 05/11/2021

DEVELOPMENT WORK FOR DECIDE2SCREEN-DEV!

Please visit func for latest example

This Azure Function Python HTTP Trigger leverages the ryp2 module to pass a pandas dataframe (created by GET request) to an R module (specifically, BCRA, more info here)

Azure Function Custom Linux Container for Python

More info here: https://docs.microsoft.com/en-us/azure/azure-functions/functions-create-function-linux-custom-image

Getting Started w/ Azure Function

Create a new directory on your host. Microsoft recommends setting up a virtual environment for this directory so you can do your respective pip installs for local azure function testing...

  • Configure your venv
# setup a folder for your new app (name it whatever youw want!) 
mkdir bcra_func_testing
cd bcra_func_testing

# setup your python virtual env for this specific project...
python3 -m venv .venv
source .venv/bin/activate
  • Start a new Python Custom Linux project
func init LocalFunctionsProject --worker-runtime python --docker
cd LocalFunctionsProject
  • Create a new function based on the HTTP trigger template
func new --name HttpExample --template "HTTP trigger"
func start

Develop your local function

Make your changes to func/LocalFunctionsProject/-init-.py, this is the Python script that gets fired off when the function is reached via URL.

  • Run the following to test & confirm your local test & dev
func start
  • When your function is working locally, you are ready to build!

Consider our Dockerfile which of course we customized for this example:

FROM mcr.microsoft.com/azure-functions/python:3.0-python3.8

ENV AzureWebJobsScriptRoot=/home/site/wwwroot \
    AzureFunctionsJobHost__Logging__Console__IsEnabled=true

# SET FOR SILENT R INSTALL
+ ENV DEBIAN_FRONTEND=noninteractive

# R DEPENDENCIES
+ RUN apt update && \
+     apt install -y r-base && \
+    R -e "install.packages('BCRA', repos='http://cran.rstudio.com/')"

# PYTHON DEPENDENCIES
+ RUN pip3 install azure-functions pandas rpy2

COPY . /home/site/wwwroot

Proceed with building, running, and pushing the docker container to ensure all is working

  • remember to set your tag version number!
  • Note: my login is atnjqt, you need your own Docker login
docker build --tag atnjqt/azurefunctionspython:v1.1.0 . 

# OPTIONAL step to confirm http://localhost:8080/
# See the microsoft guide on AuthLevel Anonymous (dev) vs Function (prod)
docker run -p 8080:80 -it atnjqt/azurefunctionspython:v1.1.0 

docker push atnjqt/azurefunctionspython:v1.1.0

Deploy Azure Function

We use the azcli, this uses existing ASC Azure resource group & storage groups that I have provisioned.

  • Run the following:
az login

az functionapp plan create --resource-group asc-etienne-rg --name myPremiumPlan --location eastus2 --number-of-workers 1 --sku EP1 --is-linux

az functionapp create --name bcra-app --storage-account etiennejapp --resource-group asc-etienne-rg --plan myPremiumPlan --runtime custom --deployment-container-image-name atnjqt/azurefunctionspython:v1.1.0

storageConnectionString=$(az storage account show-connection-string --resource-group asc-etienne-rg --name etiennejapp --query connectionString --output tsv)

az functionapp config appsettings set --name bcra-app --resource-group asc-etienne-rg --settings AzureWebJobsStorage=$storageConnectionString

Great, you now created your Azure Function for bcra-app!

Sample URL (INCLUDE code FUNCTION SECRET) -->

Enable CICD for your Azure Function Docker image

  • Run the following for your azure function app
az functionapp deployment container config --enable-cd --query CI_CD_URL --output tsv --name bcra-app --resource-group asc-etienne-rg

  • Copy the web hook url (sample, https://$bcra-app:[email protected]/docker/hook) and load it as a repository webhook for the respective image!

  • On subsequent docker image pushes, your azure function should automatically pull the latest image! I noticed ~1 minute delay from push to when the public endpoint was continously deployed!

Finally, use the Qualtrics API to pass values to our public endpoint

I just setup a sample Qualtrics Survey Web Servce usecase and this works really nicely!!

TODO --> ...

Okay so the AZ Func is really a nice way to BYOA (Bring-Your-Own-API)... This represents great progress but still a lot to go!

  • To have multiple functions for various bcra operations, this is probably an additional directory like HttpExample ...

  • Review Qualtrics api use case to pass survey question values, piped values ... This is embedded data I believe.

  • set up branching for repo, or private bitbucket repository

  • enable CICD for this development project so qualtrics survey always gets latest api endpoint results

  • look at azure resource consumption for estimates over lifecycle of decide2screen project

  • demo presentation for this work w/ emphasis on serverless computing + qualtrics survey -- dev for best of both worlds