Search interface for www.nemi.gov, written in Django.

Place local settings in nemi/nemi_project/settings/local.py. Here's a basic example that connects to the dev database:

#pylint: disable=W0401,W0614
from .dev import *

# Should be of form: oracle://USER:PASSWORD@HOST:PORT/NAME
# To override in environment, set DATABASE_URL environment variable.
# Add the if statement so that tests can be run if JENKINS_URL is defined
if not os.getenv('JENKINS_URL', False):
    DATABASES['default'] = dj_database_url.config(default='oracle://USER:PASSWORD@HOST:PORT/NAME')

# Make this unique, and don't share it with anybody.
# Get the dev SECRET_KEY from the live environment.
SECRET_KEY = ''

# Note that using this seriously slows down page rendering and if set to True
# the Browse page will not work.
#show_debug_toolbar(globals())

Create a virtualenv and install the python dependencies:

% virtualenv --python=python3 env
% env/bin/pip install -r requirements.txt

The style sheets must be built as well. In order to build the style sheets from Sass. You must have the compass Gem installed as follows:

% gem install -i Gem compass -v 0.12.7

To build the style sheets execute:

% cd nemi/compass
% ./compass.sh compile

% cd nemi
% ./manage.py runserver

A local sqlite database is used to perform testing on the python code.

% cd nemi
% DBA_SQL_DJANGO_ENGINE=django.db.backends.sqlite3 JENKINS_URL=anything ../env/bin/python ./manage.py jenkins --enable-coverage

The NEMI database is Oracle-based, and not originally managed by Django migrations. As a result, the column types do not necessarily map to what the Django ORM's Oracle driver would choose. Therefore, it is useful to inspect the SQL that Django's makemigrations management command produces before executing it against a dev database.

For example, to auto-create a migration corresponding to model changes in an app, run:

python manage.py makemigrations <app-name>

Then, inspect the SQL statements that would be executed by that migration:

python manage.py sqlmigrate <app-name> <migration-name>

If the SQL looks correct and is consistent with the patterns used in the NEMI database, it may be executed against a dev database:

python manage.py migrate

If another representation of the model change is more appropriate than the auto-created SQL, you may edit the auto-created migration and use raw SQL. For example, this migration changes the column type of a field and notifies the migration system of the field change it corresponds to:

class Migration(migrations.Migration):
    dependencies = [
        ('common', '0002_auto_20180926_1147')
    ]
    operations = [
        migrations.RunSQL(
            'ALTER TABLE "METHOD" MODIFY "SOURCE_METHOD_IDENTIFIER" VARCHAR2(45);',
            state_operations=[
                migrations.AlterField(
                    model_name='method',
                    name='source_method_identifier',
                    field=models.CharField(max_length=45, unique=True, verbose_name='method number/identifier')
                )
            ]
        )
    ]