Smart Forests Atlas

A tool to capture and narrate smart forests data, including fieldwork, interviews, social network analysis, maps and environmental data from instrumented forests.

See the live site at https://atlas.smartforests.net

Development

Requirements

• Docker Desktop
• VSCode
• VSCode Remote Development extension

Process

• Clone the repository from https://github.com/planetarypraxis/smartforests
• Open the repository folder in VSCode
• You should get a notification asking if you want to 'reopen in container'. Say yes.
  • If you don't get one, you should be able to 'reopen in container' via the command palette
  • If you can't see an option to do that, make sure you have the Remote Development extension installed
• Wait for the dev container to build.
• Check your VSCode terminal and respond to any setup prompts it asks for
• Copy the settings from Bitwarden into smartforests/settings/local.py
• Use the VSCode run and debug view (should appear if you press F5 on your keyboard) to run the app.
  • Make sure you use the 'App' configuration, which will start both the django app and the frontend build pipeline.
  • The setup_pages command in manage.py will set up the basic page structure in the development site.
• To import the production database
• Go to localhost:8000/admin for the CMS
• Go localhost:8000

Issues

• (Optional: if necessary) Make sure VSCode has selected the correct python interpreter ('command palette > Python: Select Interpreter'). It should be set to the Python in the virtualenv (venv).

Other Technical documentation

Build / deployment process

This repository has a development dockerfile (.devcontainer/Dockerfile) and a production one (./Dockerfile).

They both run .bin in ./.bin to configure their environments and install dependencies:

• Base container configuration, which is run infrequently (installing apt packages, etc) should be configured in prepare.sh.
• Frequently-run .bin (installing pip packages, etc) should go in install.sh. The difference between these is that changing prepare.sh triggers a full rebuild of the development container, whereas install.sh is only run after the container is built.
• build.sh is the last thing run on deploy to production

 Regenerating tag clouds and re-indexing logbooks

Producing tag clouds is a relatively computationally intensive operation, so we do it only occasionally with a Django management command.

python manage.py reindex_logbooks

This command:

• Saves all pages in order to update their location.
• Regenerates thumbnails for logbooks.
• Generates tag clouds.
• Creates a contributor page for each user.

You should run this command if you modify the save behaviour of models, or change the tag cloud generation algorithm.

 Background tasks to regenerate tag clouds

Tag clouds are also generated with a background task using django-background-tasks. This task is run 15 seconds after a tag is saved.

Reindexing all tags takes around 30 minutes.

Translations

1. To (re)generate the message files (.po), run yarn makemessages or python manage.py makemessages --locale=pt --locale=es --locale=fr --extension=html --ignore=env/**/*
2. We use django-deep-translate to automatically translate the message files. Ensure DEEPL_TRANSLATE_KEY is set in smartforests/settings/local.py then run python manage.py translate_messages
3. After translation, you must compile the message files using django-admin compilemessages

Stack

• Django (fullstack django app, no js frontend)
• Bootstrap (css an' stuff)
• Wagtail (CMS, administration, editor workflows and storage)
• Webpack (asset pipeline)
• PostgreSQL (Database & search index)

Hosting

• Digital Ocean (production)
• Fly.io (staging)

Third party services

• Digital Ocean App Platform (Compute, database hosting, object storage & CDN)
• Mapbox