Skip to main content

Service for triggering edX course imports from Github webhooks

Project description

gitreload

Build Status Coverage Status PyPi Downloads PyPi Version License AGPLv3

gitreload is a Flask/WSGI application for responding to github triggers asynchronously. Out of the box it is primarily intended for use with the edx-platform, but could be used for generally updating local git repositories based on a trigger call from github using its /update url.

The general workflow is that a github trigger is received (push), gitreload checks if the respository and branch are already checked out in the configured location, and then either imports that repository into the edx-platform via a ... manage.py lms --settings=... git_add_course <repo_dir> <repo_name> command, or if the trigger is set to go to /update instead of / or /gitreload, it will simply fetch the newset version of the currently checked out branch from the origin remote. Authorization is generally expected to be provided by the Web server in front of it (using basic authentication), as it currently doesn't support the use of github secrets. An additional layer of security is provided by the fact that a repository must be cloned on gitreload's host before it will accept payloads from github for said repository.

Installation

pip install gitreload

or to use the latest github version it would be:

pip install -e git+https://github.com/mitodl/gitreload

Usage

gitreload is a flask application, and thus can be run either in debug mode by directly using the gitreload command, or by using a wsgi application server. For more information on running a flask app in a production mode, see the excellent flask documentation. We generally have run it using gunicorn and supervisord in a similar manner that edx/configuration roles follow, and eventually we plan on submitting a role to install this via their ansible plays.

Upon running gitreload via command line, you should see that it starts up listening on port 5000. You can verify that it is working by going to the queue status page at http://localhost:5000/queue. If all is well, you should be greeted with some lovely json that looks like:

{"queue_length": 0, "queue": []}

Configuration

Configuration is done via a json file stored in order of precedence:

  • Path in environment variable: GITRELOAD_CONFIG
  • $(pwd)/gr.env.json
  • ~/gr.env.json
  • /etc/gr.env.json

It isn't strictly required, and the defaults are:

{
    "DJANGO_SETTINGS": "aws",
    "EDX_PLATFORM": "/edx/app/edxapp/edx-platform",
    "LOG_LEVEL": null,
    "NUM_THREADS": 1,
    "REPODIR": "/mnt/data/repos",
    "VIRTUAL_ENV": "/edx/app/edxapp/venvs/edxapp"
}

This setup means that it looks for the git repositories to be cloned in /mnt/data/repos, and expects the edx-platform settings to be the current edx/configuration defaults. It also leaves the LOG_LEVEL set to the default which is WARNING, and provides only one worker thread to process the queue of received triggers from github.

Use Cases

This is currently in use at MITx primarily for the following reasons.

Rapid Centralized Course Development

One of our primary uses of this tool is to enable rapid shared XML based edx-platform course development. It is basically the continuous integration piece for our courseware, such that when a commit gets pushed to a github repo on a specific branch (say devel), the changes are quickly and automatically loaded up with the use of this hook consumer.

Course Deployment Management

Along the lines of the rapid course development, we also use this method for controlling which courses get published on our production student facing LMS. For raw github XML developers, this means that we hook up our student facing LMS to a specific branch intended for production (say master or release). We use this to monitor that branch for changes they have vetted in their development branch and are ready to deploy to students.

We don't limit our usage of gitreload to XML development though, as we also gate our Studio course teams with this same method. There is a feature in the platform that allows course teams to export their course to Git. We use this function to control student access, allowing our Studio course authors to push at will to production once the trigger and repositories are setup for their course.

Update of external course graders

We use the regular /update feature to automatically update external code graders that are served via xqueue-watcher or xserver. We use this in a similar vain as the previous two cases and manages a development and production branch for the repository that contains the graders.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

gitreload-0.2.5.tar.gz (28.1 kB view details)

Uploaded Source

Built Distribution

gitreload-0.2.5-py3-none-any.whl (29.5 kB view details)

Uploaded Python 3

File details

Details for the file gitreload-0.2.5.tar.gz.

File metadata

  • Download URL: gitreload-0.2.5.tar.gz
  • Upload date:
  • Size: 28.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.0.5 CPython/3.8.3 Linux/5.7.2-1-MANJARO

File hashes

Hashes for gitreload-0.2.5.tar.gz
Algorithm Hash digest
SHA256 72bc1ea7efc4450465d86f29ecf15044906e8fd7264f30b7af5b34b9d902766d
MD5 197021d6da026c112cae0d898c0dddde
BLAKE2b-256 2e2d8ff29ba3f2197cc0932ff9a6b99b4ff7c2742e00c34eee491828bf21b206

See more details on using hashes here.

File details

Details for the file gitreload-0.2.5-py3-none-any.whl.

File metadata

  • Download URL: gitreload-0.2.5-py3-none-any.whl
  • Upload date:
  • Size: 29.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.0.5 CPython/3.8.3 Linux/5.7.2-1-MANJARO

File hashes

Hashes for gitreload-0.2.5-py3-none-any.whl
Algorithm Hash digest
SHA256 c4c8ea309f0509e794aa62ce7285646053af19157f339ae2cc9717a78f576223
MD5 90f2da77bf10387bf972e182c30d5c8c
BLAKE2b-256 8e21ec8bedf40f1744ac54d77834d4378dc5cf57e7e8f5a02dc1323f39f73d78

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page