Reference/Debug use: Using the Django ORM to explore the Dataverse database
JavaScript Python HTML CSS
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.


This repository may be configured to directly read an existing Dataverse 4.4+ database. This read access includes the ability to pull metrics from a Dataverse installation.

The repository was initially created as a way to explore/debug/prototype.

Live version:

database setup]


  • In use:
  • Explore/Prototype
    • Using the Django ORM to explore the Dataverse database, create queries that are later translated to Java code
    • Store cached metadata in JSON documents.

Miniverse set-up


This describes a set-up with two databases:

  1. An existing Dataverse db in Postgres and credentials to at least read the db tables
    • This Database is unmanaged. e.g. Django does not update or attempt to update the tables.
    • With proper user credentials, you can edit records via the Django Admin--but DON'T do this against a production Dataverse
  2. A database for Django to create tables needed for the map and administration.
    - This can be any relational db. (Tested with Postgres and sqlite)

Step 1: Pip install, y'all

This assumes you have installed pip and virtualenvwrapper.

# Create a virtualenv
mkvirtualenv miniverse

# cd into the repo
cd miniverse

# install the basic requirements
pip install -r requirements/local.txt

Step 2: Make a settings file

  • Create a settings file for a two database set-up.
# Make your own settings file: copy "miniverse/settings/"
cp miniverse/settings/ miniverse/settings/
  • Within your new file, change the following values:
  1. SECRET_KEY - Set a new secret key value.
  • Dataverse URL related to the Dataverse db being read. No trailing slash.
  • Set the default credentials to a new database for holding the Django/Miniverse apps
  • Set the dataverse credentials to your existing Dataverse Postgres db
  • This defaults to If you are not using the Django dev server, set it to your host. e.g. ""
  • Related to the host aboe. Set this value to http or https. Default is http
  1. Testing database
  • If you are running the Django tests, go to (about) line 135 where it reads if 'test' in sys.argv or 'test_coverage' in sys.argv:
  • Make these changes:
    • Make sure the test db is a Postgres db
    • Make sure the test db user has access to create a database and tables

Step 3: Make your settings file load when you invoke your virtualenv

  • From the top miniverse project directory, edit this file. (Note: your virtualenv must be active.)
# open with vim
vim $VIRTUAL_ENV/bin/postactivate
# OR open with Atom (or open with anything else)
atom vim $VIRTUAL_ENV/bin/postactivate
  • Add this line to the bottom of the file:
    • export DJANGO_SETTINGS_MODULE=miniverse.settings.local_settings
  • The complete file, including your line, should be
# This hook is run after this virtualenv is activated.
export DJANGO_SETTINGS_MODULE=miniverse.settings.local_settings
  • Save and close the file

  • Type: source $VIRTUAL_ENV/bin/postactivate

  • Now, each time you use your virtualenv, Django will know which settings to use

Step 4: Test your settings and run the dev server

  • Type python check
  • If everything looks good, run these lines.
    • If any provide errors (not warnings), then re-examine.
    • You will see a warning related to a "OneToOneField", that's OK
python migrate    # creates needed tables
python createsuperuser    # create a superuser for yourself
python collectstatic    # collect your static files

See if things are working. Try this:

python runserver

Try these urls:

Step 5: Run tests

If you have your Testing Database setting in place from Step 2, you can run the local tests:

python test dv_apps.metrics

If they work, you'll see something similar to this:

Installing json fixture 'test_2016_0819' from '/Users/rmp553/Documents/iqss-git/miniverse/dv_apps/metrics/fixtures'.
Processed 10033 object(s).
Resetting sequences
Installed 10033 object(s) from 1 fixture(s)
Test date params
.01 - Count total dataverses: published, unpublished, all
.02 - Test published dataverse counts by month
.03 - Test unpublished dataverse counts by month
.04 - Test all dataverse counts by month
.05 - Count total datasets: published, unpublished, all
.06 - Test published dataset counts by month
.07 - Test unpublished dataset counts by month
.08 - Test all dataset counts by month
.09 - Count total files: published, unpublished, all
.10 - File downloads by month: published,
.11 - File downloads by month: unpublished,
.12 - File downloads by month: all
.13 - Content types of published files
.14 - Content types of published files
.15 - Content types of all files
.16 - Affiliations of published dataverses types
.17 - Affiliations of unpublished dataverses types
.18 - Affiliations of all dataverses types
.19 - File extensions within type
.flush fixtures

... (lots of other stuff) ...

Ran 20 tests in 12.722s

Destroying test database for alias 'default'...

Step 6: Load map test database

  • Load the latest fixture files. The name will be in this format installations_YYYY_MMDD.json where YYYY_MMDD will be the date--use the latest.
  • From your virtual env run:
python loaddata dv_apps/installations/fixtures/installations_2016_0825.json
  • If the command above worked, try this url:

  • On Heroku, the command would be: heroku run python loaddata dv_apps/installations/fixtures/installations_2016_0825.json

Logo update notes

For logos (related to dv_apps.installations), we are developing locally and then moving the logos and model references to dev/prod via github. What does that mean?

Scenario 1: Logo file changed directly (on local environment)

If the logo is changed directly--not uploaded through the admin--then:

  1. Check it into github
  2. Push version to Heroku: git push heroku master
  3. Copy the logos to S3 (which Heroku uses for media files): - heroku run python scripts/ - The script above reads the AWS creds from the Heroku environment - If needed script may be run locally by putting the credentials in the local environ--or directly in the file (but don't check it in)

That should do it.

Scenario 2: Logo uploaded through local admin

In this scenario, the logo has changed, but also the references to it. The steps are similar, but note: 1 and 5:

  1. Create new installation fixtures, naming the output file with the current date: python dumpdata installations --indent=4 > dv_apps/installations/fixtures/installations_YYYY_MM.json
  2. Check the fixture and new image(s) into github. (The file should have been uploaded to media/logos)
  3. Push version to Heroku: git push heroku master
  4. Copy the logos to S3 (which Heroku uses for media files): - heroku run python scripts/ 5 Load the new fixtures on Heroku--using the name of the JSON file created in step 1: - ``````heroku run python loaddata dv_apps/installations/fixtures/installations_YYYY_MM.json```

(end of documentation, so far)

Maps Documentation

Swagger API docs

The swagger UI is available through the following url:

Restricted Django Admin - RestrictAdminMiddleware

Using middleware, the Django Admin access may be restricted to addresses listed in settings.INTERNAL_IPS.

When this middeware is active, users going to Django admin urls who are not coming from an IP listed in settings.INTERNAL_IPS will receive a 404 error.

  • To enable this restriction:
    • In settings.MIDDLEWARE_CLASSES, add: dv_apps.admin_restrict.middleware.RestrictAdminMiddleware
    • Example for settings.local_settings which is importing from settings.base:
    # Restrict by IP address
  • To add acceptable addresses to settings.INTERNAL_IPS
    • You can add full addresses. e.g.

      • e.g., INTERNAL_IPS = ('',)
    • You can add the 1st two segments of the address:

      • e.g., INTERNAL_IPS = ('140.247', '',)
    • You can add the 1st three segments of the address:

      • e.g., INTERNAL_IPS = ('140.247.10', '',)
    • For running django's test server, the default is:

      • e.g., INTERNAL_IPS = ('',)
    • Dev note: The code for RestrictAdminMiddleware is in dv_apps/

To fix: Site matching query does not exist.

This happens when the db doesn't have an entry for a Site object with an id of 1.

  1. Open the shell: python shell
  • On Heroku: heroku run python shell
  1. Add a Site object:
# import the Site object
from django.contrib.sites.models import Site
from django.conf import settings

# make sure there isn't a Site--should give an error ending with:
# "DoesNotExist: Site matching query does not exist."
my_site = Site.objects.get(pk=settings.SITE_ID)

# Clear any existing sites.  Case: Site exists but with wrong pk

# Now really make a site
site = Site() = settings.SITE_ID
site.domain = '' = ''

# ensure it exists - shouldn't be an error
my_site = Site.objects.get(pk=settings.SITE_ID)

Load Heroku markers

Preload to S3.

  • Can use boto script in:
heroku run python scripts/
  • Add creds to script and run

Load fixtures

Load the latest fixture files. The name will be in this format installations_YYYY_MMDD.json where YYYY_MMDD will be the date--use the latest.

  • heroku run python loaddata dv_apps/installations/fixtures/installations_2016_0826.json