Installing Geospatial libraries in Docker

Question

Django's official documentation lists 3 dependencies needed to start developing PostGIS application. They list a table depending on the database.

I use docker for my local development and I am confused about which of those packages should be installed in the Django container and which in the PostgreSQL container. I am guessing some of them should be on both.

I would appreciate your help with this.

Answer 1

You need to install the Geospatial libraries only in the Django container because they are used for interacting with a spatially enabled DB (such as PostgreSQL with PostGIS). You can deploy such a DB by using as a base a ready-made image for that purpose, such as kartoza/postgis.
Here is a nice example of a Dockerfile that uses python:3.6-slim as a base and builds the GDAL dependencies into the container. The part of the Dockerfile that you need for that is the following:

FROM python:3.6-slim

ENV PYTHONUNBUFFERED=1

# Add unstable repo to allow us to access latest GDAL builds
# Existing binutils causes a dependency conflict, correct version will be installed when GDAL gets intalled
RUN echo deb http://deb.debian.org/debian testing main contrib non-free >> /etc/apt/sources.list && \
    apt-get update && \
    apt-get remove -y binutils && \
    apt-get autoremove -y

# Install GDAL dependencies
RUN apt-get install -y libgdal-dev g++ --no-install-recommends && \
    pip install pipenv && \
    pip install whitenoise && \
    pip install gunicorn && \
    apt-get clean -y

# Update C env vars so compiler can find gdal
ENV CPLUS_INCLUDE_PATH=/usr/include/gdal
ENV C_INCLUDE_PATH=/usr/include/gdal

ENV LC_ALL="C.UTF-8"
ENV LC_CTYPE="C.UTF-8"

You can deploy both the Django app and the DB using docker-compose, using the following docker-compose.yaml (from the same repo as the Dockerfile):

# Sample compose file for a django app and postgis
version: '3'

services:

  postgis:
    image: kartoza/postgis:9.6-2.4
    volumes:
      - postgis_data:/var/lib/postgresql
    environment:
      ALLOW_IP_RANGE: 0.0.0.0/0
      POSTGRES_PASS: ${POSTGRES_PASSWORD}
      POSTGRES_USER: ${POSTGRES_USER}
      POSTGRES_DB: postgis

  web:
    image: intelligems/geodjango:latest
    command: python manage.py runserver 0.0.0.0:8000
    environment:
      DEBUG: "True"
      SECRET_KEY: ${SECRET_KEY}
      DATABASE_URL: postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@postgis:5432/postgis
      SENTRY_DSN: ${SENTRY_DSN}
    ports:
      - 8000:8000
    depends_on:
      - postgis

volumes:
  postgis_data: {}

In this repository, you can find more info and interesting bits of configuration for your issue: https://github.com/intelligems/docker-library/tree/master/geodjango (the Dockerfile snippet above is from that repo).

---

As a Note:
If you want to create a PostgreSQL with PostGIS enabled database as a "local DB" to interact with your local Django, you can deploy the previously mentioned kartoza/postgis:

1. Create a Volume:

   $ docker volume create postgresql_data
   

2. Deploy the container:

   $ docker run \
         --name=postgresql-with-postgis -d \
         -e POSTGRES_USER=user_name \
         -e POSTGRES_PASS=user_pass \
         -e ALLOW_IP_RANGE=0.0.0.0/0 -p 5433:5432 \
         -v postgresql_data:/var/lib/postgresql \
         --restart=always \
         kartoza/postgis:9.6-2.4
   

3. Connect to the default DB (postgres) of the container and create your DB:

   $ psql -h localhost -U user_name -d postgres
   $ CREATE DATABASE database_name;
   

4. Enable PostGIS extension to the database:

   $ \connect database_name
   $ CREATE EXTENSION postgis;
   

This will result in a DB with the name database_name listening to port 5432 of your localhost and you can connect to that from your local Django app.