This is the feature-limited port of the original Tweet Mashup from F# to R. It uses Shiny and rtweet to recreate the original interface. The Shiny application is hosted in a Shiny Server using a Docker image from a modified T-Mobile Tensorflow R container
Before you can use this app you first need to setup a Twitter application for it:
1. Register the application - Go to https://developer.twitter.com/en/apps and create an application. This make take some time to get approval
2. Set the callback url - Finally, in the Twitter developer site you'll need to add the appropriate callback urls. This is required and needs to be anywhere you will host the application. This means you'll want to include:
http://127.0.0.1- for testing the application locally[url you host it on]- for when it's running livehttp://127.0.0.1:1410[optional] - if you want to do general experimentation with rtweet this will allow you to easily authenticate
3. Get the keys - If you click the details of the app you created you'll see four keys: consumer key, consume secret key, access token, and access token secret. Keep these stored somewhere
4. Add google analytics (optional) - If you want to add Google analytics, create a GA account and get the ID for your page.
5. Set the environmental variables - now with your keys from (3) and (4), set the environmental variables below. You'll need to do this for each new R session that runs the app.
Sys.setenv("TWITTER_CONSUMER_KEY"="[CONSUMER KEY]")
Sys.setenv("TWITTER_CONSUMER_SECRET"="[CONSUMER SECRET]")
Sys.setenv("TWITTER_ACCESS_TOKEN"="[ACCESS TOKEN]")
Sys.setenv("TWITTER_ACCESS_TOKEN_SECRET"="[ACCESS TOKEN SECRET]")
Sys.setenv("GOOGLE_ANALYTICS_ID"="[Google Analytics ID]") #optional, disregard if you don't have a google analytics IDAfter following the setup instructions, you can go into the src folder and run the following r command:
shiny::runApp("app",port=80L, launch.browser = FALSE)After running this command you can navigate to http://127.0.0.1 in your broswer and test it out.
Instead of directly running the R code, you can create a Docker image with the R code inside it.
Before starting ensure you have installed Docker To build the container with docker installed use the following command from the commany line in the base folder of the repository:
docker build -t tweet-mashup-r .
To run it locally, do the following:
docker run -it --rm -p 80:80 tweet-mashup-r
Then navigate to http://127.0.0.1 to try it out!
One of the easiest ways to publically host this app is with Google Cloud Compute Engine Virtual Machines. I have been able to host it for less than $10 a month, but it's likely possible to do it fore more cheaply. To do so, first create a Google Cloud project. Then run the following commands:
First, tag the container with how you want to add it to Google Cloud Registry (this will be how it's available to the virtual machine):
docker tag tweet-mashup-r gcr.io/[PROJECT NAME]/tweet-mashup-r:latest
Then, push it to Google Cloud Registry. This actually makes it to GCP:
docker push gcr.io/[PROJECT NAME]/tweet-mashup-r:latest
Finally, create an instance that runs it. We'll use an instance that size "e2-micro" which is the smallest container that is sufficiently powerful for the task. We also will specify that the server should be open to http traffic. The zone defaults to with the united states (us-central1-a), but this could be adjusted to your region.
gcloud compute instances create-with-container tweet-mashup-r-vm --zone us-central1-a --machine-type e2-micro --tags http-server --container-image=gcr.io/[PROJECT NAME]/tweet-mashup-r:latest
After the container is deployed, if you make changes to the container you can redeploy them by building, tagging, and pushing the new container and restarted the VM. You can do this by running the following:
docker build -t tweet-mashup-r .
docker tag tweet-mashup-r gcr.io/[PROJECT NAME]/tweet-mashup-r:latest
docker push gcr.io/[PROJECT NAME]/tweet-mashup-r:latest
gcloud compute instances reset tweet-mashup-r-vm