Use multi-stage build for Docker

Dockerfile

@@ -1,24 +1,46 @@
-FROM python:3.10-slim
+#
+# Backend build
+#
+FROM python:3.11-slim-bookworm AS backend
+
+# build deps
+RUN apt-get update && apt-get upgrade && apt-get install -y --no-install-recommends \
+        build-essential \
+        libpq-dev
+
+# get latest version of pip
+RUN pip install pip -U
+
+# install requirements
+COPY /requirements/* /app/requirements/
+RUN pip install -r /app/requirements/base.txt
+
+# pyppeteer deps (https://stackoverflow.com/a/71935536)
+RUN xargs apt-get install -y --no-install-recommends < /app/requirements/pyppeteer_deps.txt
+
+
+#
+# Final build
+#
+FROM python:3.11-slim-bookworm AS final
 
-ENV PIP_DISABLE_PIP_VERSION_CHECK 1
 ENV PYTHONDONTWRITEBYTECODE 1
 ENV PYTHONUNBUFFERED 1
+ENV DJANGO_ENV "BASE"
 
-COPY requirements/* requirements/
+RUN apt-get update && apt-get upgrade -y && apt-get install -y --no-install-recommends \
+        postgresql-client
 
-RUN apt-get update \
-  # psycopg2 + deps
-  && apt-get install -y --no-install-recommends build-essential \
-  && apt-get install -y --no-install-recommends libpq-dev \
-  && pip install psycopg2 \
-  # pyppeteer deps (cf. https://stackoverflow.com/a/71935536)
-  && xargs apt-get install -y --no-install-recommends < requirements/pyppeteer_deps.txt \
-  && pip install -r requirements/production.txt
+# copy backend deps
+COPY --from=backend /usr/local/lib/python3.11 /usr/local/lib/python3.11
+COPY --from=backend /usr/local/bin/ /usr/local/bin/
 
-COPY . /usr/src/app
-WORKDIR /usr/src/app
+COPY . /app
+WORKDIR /app
 
-RUN python manage.py collectstatic --link --no-input
+# create user and drop privileges
+RUN useradd -m pi-sigma
+RUN chown -R pi-sigma /app
+USER pi-sigma
 
-RUN useradd -m myuser
-USER myuser
+RUN python manage.py collectstatic --link --no-input

README.md

@@ -68,8 +68,8 @@ http://localhost:8000
 ```
 
 If all went well, you should see the homepage of the app with a list of news sources arranged in a grid.
-The grids are empty to begin with, but the celery workers will start right away and you should see the
-first articles displayed shortly thereafter.
+The grids are empty to begin with and fill up when the celery workers start
+(depends on the schedule in `scraper.tasks`).
 
 In order to extract data about the sources from the database, use the following command while the web container is running (the commands for the other tables are analogous):
 ```sh

docker-compose.yml

@@ -14,7 +14,7 @@ services:
     image: nous_aggregator
     build: &web_build
       context: .
-      dockerfile: Dockerfile.dev
+      dockerfile: Dockerfile
     environment: &web_env
       - DJANGO_SETTINGS_MODULE=${DJANGO_SETTINGS_MODULE}
       - SECRET_KEY=${SECRET_KEY}

scraper/tasks.py

@@ -1,6 +1,6 @@
 magazines = {
     "en": {
-        "schedule": 3600 * 4,  # 4h
+        "schedule": 3600,
         "titles": [
             "Al Jazeera",
             "Associated Press",