Skip to content

Latest commit

 

History

History
71 lines (49 loc) · 1.49 KB

README.md

File metadata and controls

71 lines (49 loc) · 1.49 KB
Raw

Tweet UI

Sentiment analysis UI to check whether tweets are ok or not.

Format

The UI will query the backend API using a POST to $BASE_URL/api/analysis containing the following json body:

{
    "content": "This is a bad sentence"
}

It expects a response in the format:

{
    "toxicity":false,
    "threat":false,
    "sexualExplicit":false,
    "severeToxicity":false,
    "obscene":false,
    "insult":false,
    "identityAttack":false
}

The $BASE_URL environment variable needs to be set at build time and must start with http:// or https://

Usage

You can run the service locally by running:

npm install
BASE_URL=http://localhost:8080 npm run webpack:server

Deployment

You can deploy the UI as a static website or as a service, it's up to you.

Static website

To bundle the UI for static website deployment, run

npm install
BASE_URL=http://localhost:8080 npm run webpack:bundle

This will bundle the HTML and js files into the dist/ folder which you can then deploy. To test that the site works, run

npm run serve

and open http://localhost:9000

Service

To deploy the UI as a service, you can use the provided Dockerfile which runs the UI in an integrated HTTP server, e.g. by running:

docker build --no-cache --build-arg base_url=http://localhost:8080 -t="hivemind/tweet-ui" -f docker/Dockerfile .
docker run -ti -p 9000:9000 hivemind/tweet-ui:latest

and open http://localhost:9000