Dockerfile

# docker image
FROM python:3.11-slim

ENV PYTHONDONTWRITEBYTECODE=1
ENV PYTHONUNBUFFERED=1

RUN echo "Starting Docker build..."
RUN echo "Build time:: $(date)"

# set the working directory (app folder here)
# container has it's own filesystem within that we put all the files that need for a particular api like here app is the working directory.
WORKDIR /app

# install dependencies
COPY ./requirements.txt /app
# --no-cache-dir --upgrade-> prevents pip from using its local cache \
#  and forces pip to download the package fresh from PyPI.
RUN pip install --no-cache-dir --upgrade -r requirements.txt

# copy the scripts to the app folder
# everything on the root folder specify by `.` (dot), copy to app folder of docker container
COPY . /app

# . (destination) → current WORKDIR and WORKDIR =/app
# so, `COPY . .` also same.

# start the server (we internally run on port 80 (default port) on docker container)
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "80"]


# How Docker builds images:
    # Docker builds images layer by layer.
    # Each instruction in the Dockerfile creates one immutable layer.
    # Docker reuses layers if nothing has changed in the layer.

# Why requirements.txt is copied separately ?

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

# Dependency installation runs only when requirements.txt changes
# Application code changes don't trigger reinstalling dependencies
# Builds become much faster