Docker Compose Demo
This repository contains a
docker-compose.yml file starting the following containers:
db- a PostgreSQL database housing a toy dataset
api- a mock prediction API (built as a responder app). In a real world setting this API could, for instance, run an input data point through a previously trained machine learning model and return a prediction.
dash- a simple Dash dashboard that reads data from
dband runs them trough the prediction
jnb- a Jupyter lab environment to run Jupyter notebooks being able to read/write data from
db, run data through the
apiand so on.
To run Docker Compose:
docker-compose up --build
The Jupyter lab server is available at: http://127.0.0.1:10000/lab
The dashboard is then available at: http://127.0.0.1:8050/
Shutdown and remove containers:
- fancier dashboard?
- how do volumes relate to containers? what about persistance?
- add more container components?
- show how to use a simple, pretrained tf model from tf hub in API?