sentence-classification

Via universal sentence encoder and k-means

Running the Application Locally (Without Docker)

Before running the application, it's recommended to rebuild the tensorflow-node for your environment.

## make sure you have the necessary packages installed.
## If you're unsure about which packages are required, refer to the Dockerfile for guidance.  
npm run rebuild:tsjs

Once the rebuild is complete

npm install
npm run start

Building the Docker Image

To build the Docker image for the application, make sure to set the HTTP_PORT build argument to match the port in your .env file.

docker build --build-arg HTTP_PORT=4544 --tag hermeslin/sentence-classification:latest .

Running the Docker Container with the .env File

To run the Docker container with your .env file for environment variable configuration, execute the following:

docker run -p 4544:4544 --env-file .env hermeslin/sentence-classification:latest

Overriding the Entry Point (Optional)

If you need to access the container's shell directly (for debugging or other purposes), you can override the entry point and open a Bash shell with:

docker run -it --entrypoint /bin/bash -p 4544:4544 --env-file .env hermeslin/sentence-classification:latest

Building and Pushing the Docker Image to GCP via GitHub Actions

Google authentication