eha-nutsurv-django

NutSurv - Data collection and quality assurance tools for Nutrition Surveys on mobile devices

Installation

The following has been tested on Ubuntu 14.04. Other versions of Linux/Ubuntu or Mac OS X will have to make minor adjustments.

First install all needed dependencies:

sudo apt-get install git postgresql-9.4 postgresql-9.4-postgis-2.1 python-virtualenv python-dev
libpq-dev postgresql-server-dev-all gfortran libopenblas-dev liblapack-dev postgresql-server-dev-all
g++ gfortran nodejs npm nodejs-legacy python-numpy python-scipy

Install Bower for front-end packages

sudo npm install bower -g

Then get the sources (enter github username/password):

git clone https://github.com/eHealthAfrica/eha-nutsurv-django.git nutsurv

Now create a python virtual environment:

virtualenv --no-site-packages nutsurv-venv

Activate the virtual env (has to be done every time you want to run the application)

source nutsurv-venv/bin/activate

Enter the source folder:

cd nutsurv

Either install python requirements for a production environment:

pip install -r requirements/production.txt

or for development:

pip install -r requirements/development.txt

Create the database nutsurv_dev (ignore the errors) and enable Postgis:

sudo -u postgres psql -f make_nutsurv_dev.sql
sudo -u postgres psql -d nutsurv_dev -f enable_postgis.sql

Install JavaScript (bower) dependencies:

bower install

Enter the application:

cd nutsurv

Add structure to the database:

./manage.py migrate

Create cache tables:

./manage.py createcachetable

Collect static files:

./manage.py collectstatic

Load some structural data:

./manage.py loaddata SecondLevelAdmin

Create a superuser for Django access:

./manage.py createsuperuser

Run the server either locally:

./manage.py runserver

Or run the server with global access (for Vagrant access):

./manage.py runserver 0.0.0.0:8000

Access it in your web browser at http://localhost:8000/

Contributing

Please run flake8 on your code before making a PR

flake8 --exclude migrations --ignore E501,C901 .

Good job!

The website should be available at http://HOST:8001

Preload GIS data

After the system has been installed and configured you might want to load some geographic data set with LGAs instead of adding them via the Django admin site. Please continue reading this section.

Example GIS data

In case you do not have your own shapefile, the shapefile used for testing has been downloaded from http://www.diva-gis.org/gdata by selecting 'Nigeria' from the drop-down list of countries and 'Administrative areas' from the drop-down list of subjects. The thus acquired NGA_adm.zip file was then unzipped and the Nigerian LGA level data was imported (see below for instructions) from the NGA_adm2.shp file (which was one of the files contained in the downloaded archive).

Simple scenario

In order to preload the database with some LGA data, copy the shapefile (e.g. somefile.shp + somefile.shx + somefile.dbf) you want to use to the directory of your choosing, invoke the Django shell:

$ path/to/the/manage.py shell

WARNING: This scenario assumes that the shapefile you prepared is compatible with the default mapping dictionary defined in the load module. If it is not, check the optional steps described later in the 'Custom mapping dictionary' section. As of this writing, the shapefile mentioned in section 'Example GIS data' above is compatible and importing it does not require any additional steps.

Next, import the load module and call the import_data_from_file function passing the path to your .shp file as an argument:

>>> import dashboard.load
>>> dashboard.load.import_data_from_file('/path/to/your/shapefile.shp')

After the function finishes, exit the shell and enjoy all those newly added LGAs.

Custom mapping dictionary

The default mapping dictionary defined in the load module assumes that the shapefile has the following attribute fields:

• NAME_0 - which stores the name of the top-level area (e.g. a country)
• NAME_1 - which stores the name of the mid-level area (e.g. a state)
• NAME_2 - which stores the name of the area of interest (e.g. a county) - i.e. the area contained by the mid-level area (which in turns is contained by the top-level area)
• VARNAME_2 - an alternative name of the area of interest (optional, can be NULL)

Since the LGA model defines the following fields: name_0, name_1, name_2, varname_2, mpoly, the attribute fields of the shapefile have to be mapped to the fields of the LGA model by defining the following dictionary (do not worry about the mpoly mapping, just remember to always have it in your mapping dictionary):

$ path/to/the/manage.py shell

>>> some_mapping_dictionary = {
>>>     'name_0': 'NAME_0',
>>>     'name_1': 'NAME_1',
>>>     'name_2': 'NAME_2',
>>>     'varname_2': 'VARNAME_2',
>>>     'mpoly': 'MULTIPOLYGON',
>>> }

Now, when you have defined your new dictionary, you can use it and load your GIS data:

>>> import dashboard.load
>>> dashboard.load.import_data_from_file(filepath='/path/to/your/shapefile.shp',
                                         mapping=some_mapping_dictionary)

===================

Override settings

You can override settings by copying the "configuration.py-default" file to "configuration.py" and customize it. The file is in the nutsurv directory where also manage.py can be found.