Django model generator for chop-dbhi/data-models-service style JSON metadata.

If you're not using a Python virtual environment, please do. I prefer pyenv with pyenv-virtualenv.

Get the most recent stable version:

pip install dmdj

Or build and install the (unstable) HEAD version:

git clone https://github.com/chop-dbhi/data-models-django.git
cd data-models-django
make install

Place the following in your app's models.py file:

import requests
from django.db import models
from dmdj.settings import get_url
from dmdj.makers import make_model

model_url = get_url('pednet', '2.1.0')
model_json = requests.get(model_url).json()

for model in make_model(model_json, (models.Model,), module='yourapp.models',
                        app_label='yourapp')
    globals()[model.__name__] = model

This code causes a single web request at runtime, which may slow down your app's startup. Also, the models are dynamically generated and so may change over time, although efforts to improve the semantic versioning and stability practices in the data-models repo are under way.

Install the development requirements and the package in "editable" mode.

git clone https://github.com/chop-dbhi/data-models-django.git
cd data-models-django
make devinstall

Tests are run on the source code using tox to replicate them across the range of compatible Django and Python versions, for the Python versions you have available in your environment. I like to create a dmdj virtual environment from 3.5.x and then use pyenv local dmdj 3.4.x 2.7.x

make test

Generate test coverage information that prints in the terminal and creates HTML and XML format files.

make coverage

These tasks are routinely handled by the CI/CD workflow, but I'll document them here anyway.

Publish test coverage information to coveralls (the project must be registered and the token must be in the COVERALLS_REPO_TOKEN env var).

make coveralls

Release a new final version by pushing a new tag to GitHub and uploading the dist files to PyPi (the project must be registered at PyPi and you must have PyPi credentials available or in the PYPI_USER and PYPI_PASS env vars).

make release