Terralego Backend for OPP module

Requirements

• To handle pictures in templates, please install weasyprint requirement librairies https://weasyprint.readthedocs.io/en/stable/install.html#linux

First, create a data layer for observatory

./manage.py create_observatory_layer -n observatory

Then get the given primary key, for example 10.

Settings needed to be set

TROPP_OBSERVATORY_LAYER_PK=10  # replace by primary key given by command

TROPP_OBSERVATORY_ID = 20 # set the nationnal id of your observatory

VERSATILEIMAGEFIELD_RENDITION_KEY_SETS = {
  'terra_opp': [
    ('original', 'url'),
    ('full', 'thumbnail__1500x1125'),
    ('list', 'thumbnail__300x225'),
    ('thumbnail', 'thumbnail__180x120'),
  ]
}

AUTH_USER_MODEL = 'terra_accounts.TerraUser'

Media files

If your project is not using the default django storage, then you must define and set a url fetcher in order to tell weasyprint where to find your media files.

An example of url fetcher using media files from S3 storage :

from django.conf import settings
from terra_opp.renderers import django_url_fetcher


def custom_url_fetcher(url, *args, **kwargs):
   scheme = 'https' if settings.AWS_S3_SECURE_URLS else 'http'
   url_prefix = f"{scheme}://{settings.AWS_S3_CUSTOM_DOMAIN}"

   if url.startswith(url_prefix):
       url = url.replace(
           url_prefix,
           settings.AWS_S3_ENDPOINT_URL + settings.AWS_STORAGE_BUCKET_NAME
       )

   return django_url_fetcher(url, *args, **kwargs)

And then you must refer to this custom url fetcher in your settings. Example if your fetcher is define in custom/fetcher.py:

TROPP_URL_FETCHER = 'custom.fetcher.custom_url_fetcher'

To start a dev instance

Define settings you wants in test_opp django project.

docker-compose up

Then initialize the database:

docker-compose exec web /code/venv/bin/python3 /code/src/manage.py migrate

and create the base layer:

docker-compose exec web /code/venv/bin/python3 /code/src/manage.py create_observatory_layer -n observatory

You can now edit your code. A django runserver is launched internally so the this is an autoreload server.

You can access to the api on http://localhost:8000/api/

Test

To run test suite, just launch:

docker-compose exec web /code/venv/bin/python3 /code/src/manage.py test

Releasing a new version

# on the master branch
# update the file CHANGES.md with your latest changes
git changelog

# Next update the version in the file terra-opp/VERSION.md
echo X.X.X > terra-opp/VERSION.md

# Next use git release to create the tag et push your branch & tag to origin
git releasse X.X.X

After this few steps, you can now go on the github repo to create a new release with the corresponding tag.

0.5.6 / 2021-03-30

• Update readme with release procedure (#55)
• Add more concistency settings naming and documentation
• Add checks for TROPP_OBSERVATORY_ID

0.5.5 / 2021-03-30

• Broken release

0.5.4 / 2021-03-29

• Add active status and endpoint for active viewpoint only

0.5.3 / 2021-03-17

• Add owner_id to picture to be to modify it
• Add business identifier for picture

0.5.2 / 2021-03-11

• Add active status on viewpoint

0.5.1 / 2021-02-17

• Add permissions generation on migrate

0.5.0 / 2021-02-17

• Add translations
• Add permissions and check
• Update installation process
• Enhance command

!! WARNING : BREAKING CHANGES !!

• You need to adapt your project settings and requirements
• Use new django-terra-settings instead of django-terra-utils
• Some terra-utils functions are directly integrated

0.4.2 / 2020-09-09

• Taking out city and themes from JSON properties, making it related objects on each viewpoint instance
• Improving related document serializer, it does not return the file as base64 anymore but only the access url

0.4.1 / 2020-08-26

• Removing signals and doing all the logic in the related viewsets
• Add possibility to make your own url fetcher :
  • Add TROPP_URL_FETCHER settings and instructions
• Fix factory boy version
• using django.db.models.JSONField instead of django.contrib.postgres.fields

0.4.0 / 2020-07-30

• Warning ! Breaking changes

  • Due to uniformity of storage usages, you need to use file:// schema in your template to access pictures
  • Deprecate defining observatory layer by name.
  • Now you need to define layer pk instead -> TROPP_OBSERVATORY_LAYER_PK.
  • Existing frontend applications should be fixed by TROPP_OBSERVATORY_LAYER_PK=1

• fix picture property

• fix prefetched data as list and not queryset

• fix restframework not listed in setup requirements

• Fix storage usage and make it working with weasyprint with default and custom storages

• add correct dependencies to handle JPG

• allow argument set layer name at creation

• dont block if observatory layer pk not defined, to let ability to create it

• block start without defined correct layer settings

• add and improve configuration checking

• add command to create point layer to used as observatory

• set defined observatory layer by pk and provide info to frontend to avoid expecting hardcoded pk=1 layer for tiles

• PK and corresponding endpoints are auto added to /api/settings to send dynamic configuration to frontend

• Storage bucket is not required anymore.

0.3.8 / 2020-06-23

• Fix default settings

0.3.7 / 2020-06-19

• Support django 3.0
• add "as_versatile" filter for template usage

0.3.6 / 2019-12-19

• Compatibility with python 3.8, django 3.0 and DRF 3.11
• Add flake8 for linting

0.3.5 / 2019-11-04

• Fix MEDIA_URL may be empty, breaking url fetcher

0.3.4/ 2019-10-10

• Remove remarks field on Picture

0.3.3 / 2019-10-09

• Fix Manifest to include md files

0.3.0 / XXXX-XX-XX

First standalone release

• Extract from terra-common package to make it a standalone package