/contented

Delivery of results (notebooks, tables) for data science projects in a user-authorised way (Django based)

Primary LanguageJavaScriptMIT LicenseMIT

contented

This is a django project for delivery of data-analysis results (notebooks, tables etc) in a user-authorised way: Only those users with access to a given project should be able to view it's deliverables.

To use contented:

  • obtain a webserver, register a website domain and point the website URL to the webserver;
  • put your analysis deliverables in a directory on the webserver;
  • follow the provisioning notes in ./deploy_tools/provisioning_notes.md (eg, install python / nginx etc on the webserver)
  • run the fabric script in ./deploy_tools - this will clone the contented repository onto the webserver, install the python env / update the database etc
  • then configure the files that users can see by modifying the ./.env file, as described in the 'CONFIG' section below

ENV

Environment is managed by pipenv

django 3.1.0 is used, as in "Django for Beginners" book by WS Vincent.

black is used for styling: installed it into the dev environment using pipenv install black --dev --pre

pylint and pylint_django are used for linting

CONFIG

The file ./.env is used to configure each deployment of contented.

Each line of this file defines an environment variable used while contented runs.

A template .env file can be found in ./deploy_tools/template.env

The config variables are:

  • PROJECTS_DIR: A directory that contains all projects that should be made visible on the website. The default value points to a set of projects that are used during testing.

  • RESTRICTED_PROJECTS: Access to a subset of the projects in PROJECTS_DIR may be restricted (users must be logged in to view them) by adding their names to this comma-separated string. The default is for all projects to be publicly accessible (this occurs when RESTRICTED_PROJECTS is missing or the empty string).

Tests

contented is developed using TDD (based brazenly on the tests in TDD with Python by H Percival). Write functional tests first, then unit tests, then code.

selenium is used for functional testing

  • Some differences between current selenium (4.0.0a7) and the one used in the TDD book project (3.141.0)
  • Use find_element(By.ID, "ID") and find_elements(By.TAG_NAME, "TAG-NAME") instead of find_element_by_id("ID") and find_elements_by_tag_name("TAG-NAME") functions

Use Django test client to check that which templates are used in response to a http request, eg, self.client.get("/")

Note you can use @override_settings decorator, if you want to change the project data that is presented by contented during tests (plan is to use env variable to define the location of the data that is presented; this env var will be parsed by settings.py)

Notes

This is my first (non-tutorial) Django project. So there may be much better ways of achieving the project's goals.

  • Github complained about the first push of this project since django SECRET_KEY was present in config/settings.py. Therefore, copied code from superlists project into settings.py (so the SECRET_KEY is obtained from os.environ)