build: introduce the worker as a separate node

This makes the Docker environment feature-complete since no worker and
therefore no tasks were ran at all before this.

It also paves the way for that same separation to happen in production
because currently both the worker and the web server happen to run on
the same machine, which isn't the best idea for scalabilty or
resilience.

Introduce:
- a Redis node;
- a new "worker" service that runs off the tagged server
image ("ma-cantine-server");
- a Makefile to help administrate the Docker setup.

.env.docker

@@ -27,7 +27,7 @@ DATA_WARE_HOUSE_PORT=
 DATA_WARE_HOUSE_DB=
 DEBUG_PERFORMANCE=True
 ENVIRONMENT=dev
-REDIS_URL=redis://localhost:6379/0
+REDIS_URL=redis://redis
 REDIS_PREPEND_KEY=LOCAL
 OVERRIDE_TEST_SEED=
 SIRET_API_KEY=

Makefile

@@ -0,0 +1,13 @@
+DOCKER-RUN = docker compose run -e TERM --rm --entrypoint=""
+
+build:
+	docker compose build
+
+up:
+	docker compose up
+
+down:
+	docker compose down
+
+sh:
+	$(DOCKER-RUN) server bash

compose.yaml

@@ -9,6 +9,7 @@
 # https://github.com/docker/awesome-compose
 services:
   server:
+    image: ma-cantine-server
     build: .
     volumes:
       - "./:/app"
@@ -65,6 +66,16 @@ services:
       interval: 10s
       timeout: 5s
       retries: 5
+  worker:
+    image: ma-cantine-server
+    env_file:
+      - ".env.docker"
+    command: ["celery", "-A", "macantine", "worker"]
+    depends_on:
+      - db
+      - redis
+  redis:
+    image: redis:latest
 volumes:
   # it's normal that this is empty, it simply defines the existance of persistent storage
   db-data: