The purpsose of the sentiment-analysis-mean-app is to provide a simple sentiment analysis mean app. The application accepts text in the form of a 'Post' and forwards that text to a python tensorflow application. The application is currently spawned as a child process of the node server when a new post comes in. HOwever, in the future it will become its own micro-service that gets the text through pub/sub between the itself and the server. The predicted sentiment is then returned and displayed to the web app.
Install docker sentiment-analysis-python-app
Run docker network create -d bridge sentiment-analysis-bridge
From this directory run: docker build -t sentiment-analysis:dev -f docker/Dockerfile.dev sentiment-analysis-mean-app
From this directory run: docker build --no-cache -t sentiment-analysis:prod -f docker/Dockerfile.prod sentiment-analysis-mean-app
See README in sentiment-analysis-mean-app
This app depends on a MongoDB server running on the same bridge network with the resovable hostname sentiment-analysis-db. The easiest way to get a mongo database up and runnign is with their Docker image.
docker run --network=sentiment-analysis-bridge --name=sentiment-analysis-db --rm mongo:bionic
This app depends on a redis server running on the same bridge network with the resovable hostname sentiment-analysis-broker. The easiest way to get redis up and runnign is with their Docker image.
docker run --network=sentiment-analysis-bridge --name=sentiment-analysis-broker --rm redis:alpine
Run docker run --network=sentiment-analysis-bridge -p 3000:3000 --rm sentiment-analysis:dev
Run docker run --network=sentiment-analysis-bridge -p 3000:3000 --rm sentiment-analysis:prod
- Make mongoDB hostname a cmd arg
- Split python app into seperate service
- Convert Prod Dockerfiles to Alpine after Python separation
- Make redis hostname a cmd arg
- Use web sockets for sentiment analysis