Repository files navigation

django-dcat is a Django app that provides a model layer forDCAT 3.0metadata and some command line tools to import data to it, to create vocabularies and more.

If you wanna see an example of what can be done checkoutCatalogo Social, adata catalog implemented with django-dcat.

django-dcatwas presented at the CSV Conf 2024 in talk abouthow to preserve and backup open data portals.

1. Install the package using pip:

   pip install django-dcat

2. Add "dcat" to your INSTALLED_APPS setting like this:

   INSTALLED_APPS = [    ...,    "dcat",]

3. Runpython manage.py migrate to create the DCAT models in your database.

4. Start the development server and visithttp://127.0.0.1:8000/admin/to start adding data (you'll need the Admin app enabled).

5. Runpython manage.py --help to see the available commands.

There are two commands that allows you to:

1. Make a dump of data from a CKAN data portal (that hasckanext-datajson installed)
2. Create a Catalog with its related entities (datasets, distributions, etc)

# Download the data.json file from the portal$ wget -O data.json https://www.ckan-example.org/data.json# Download all files from the portal (by default in a data/ folder)$ python manage.py dump_from_datajson# Import all data and files into a local catalog$ python manage.py import_from_datajson

There are two commands to import controlled vocabularies used in the EU Open Data Portal in order to standardise the metadata.

# Populate records for the MediaType model (CSV, PDF, etc)$ python manage.py import_filetypes# Populate records for Licences (MIT, Apache, etc)$ python manage.py import_licences

The goal of these commands is to provide data publishers with pre-filled options for the metadata fields. This will improvedata quality and avoid common problems like duplicated metadata values for typos or inconsistent data entry (like distributions with.csv, .CSV, CSV, etc)

Every object contains methods to serialize it to different formats. This way it is easy to implement a feed URL for your catalog.

• catalog.to_jsonld(): (WIP) exports the catalog to a JSONLD format. The implementation has been inspired by theFAO - Data in Emergencies feed.
• catalog.to_turtle(): TODO
• catalog.to_rdf(): TODO

django-dcat focuses on providing a model layer for DCAT metadata. However, if you require custom fields in your application,you can extend any model using a OneToOneField pattern (similar to what you use to extend Django's User model).

fromdjango.dbimportmodelsfromdcat.modelsimportDistributionclassDistributionExtras(Distribution):distribution=models.OneToOneField(Distribution,on_delete=models.CASCADE,related_name='extras')my_extra_field=models.CharField(max_length=255,blank=True,null=True)

And then you can call this fields from your code using the related name attribute:

fromdcat.modelsimportDistributiondistribution=Distribution.objects.get(pk=1)print(distribution.extras.my_extra_field)

Note: Instead of calling itextras You can play with more semantic names for the related_nameattribute like the name of your app.

The Django models in this package are inspired by the diagram presented in DCAT profiles.

See the diagram:UML diagram illustrating the DCAT-AP specification

For more information on DCAT:

• Data Catalog Vocabulary (DCAT) - Version 3.0
• DCAT-AP profile

The project uses twine to publish to PyPi:

• Update the version insetup.cfg
• Create a new release in Github.
• Build the package:python -m build
• Upload to PyPitwine upload dist/*