Using DVC in dockerized container in a production environment

[utshb]

Let's say I have a simple Python app (as a github repo) which loads some data and processes it. I have the data stored in an S3 bucket which I pull through DVC (dvc pull). Everything works fine in a local environment on machine.

Now, I would like to use DVC in production for a dockerized application. How can I make use of DVC within the Dockerfile? I couldn't find any examples in docs so it would be great if someone can point me the right direction.

Here is my dockerfile.

FROM python:3.8-slim 

WORKDIR /code
COPY ./requirements.txt /code/requirements.txt 
RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
COPY ./app /code/app

CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "80"]

Here is the repo structure:

some_github_repo: 
  -----.dvc 
  -----app 
  -------------- some_python_script.py
  -------------- data/somedata.dvc
  -----.dvcignore 
  -----Dockerfile
  -----requirements.txt

Locally this setup works fine. I have added the data to an S3 bucket through dvc push. DVC pull works fine too. Now how can I use dvc pull within a dockerfile such that it pulls this data in dockerised container?

I can imagine adding the following to my Dockerfile.

1) Make the AWS security key/secret available as ENV variables  
2)pip install dvc[s3]
3) dvc pull 

However, dvc needs to have access to its configuration files in a container to connect to the right s3 bucket and get the relevant data. How can I achieve this within a dockerfile? The dockerized application will run in Fargate cluster in AWS.