GTest for running c++ unit tests for each compile groups, sequential and MPI parallel
RunAllNotebooks.sh for automatically running all python notebook examples

See bitbucket-pipelines.yml in MUQ2 repo
Coverage
- Basic c++ unit tests and python notebooks test on push
- Weekly extended runs mainly for different OS configs
- Weekly installation instructions script run to ensure they it still leads to a running MUQ setup

Hosted on bitbucket pages in the mituq.bitbucket.io repository.
Accessible at https://mituq.bitbucket.io/
When something is committed to this repository, bitbucket pipelines runs Jekyll and generate the html.
(in progress) HTML example pages are generated from the documentation/scripts/ProcessExamples.py in the MUQ2 repository. This script is run from the pipelines script in the MUQ2 repository when a new git tag is created.
The bitbucket.io site is static, but the slack invitation button requires extra processing that cannot occur on a static site. To overcome this, a "cloud function" has been set up on the Google cloud platform to handle POST requests when the invitation modal and interact with the slack API.

Built via CI when release tag is defined on master repo, pushed as commit to website repo

The docker images are built on dockerhub using their automated builds functionality.
Dockerfiles for the images can be found in muq/SupportScripts/docker
Built images can be found in https://hub.docker.com/r/mparno .

The MUQ2 conda image lives on conda-forge.
The recipe for the MUQ2 conda image lives in the muq-feedstock repository on github.
Directions for updating the recipe manually can be found here.
Linus and Matt are currently recipe maintainers.
In our bitbucket-pipelines.yml file, a fork of conda-forge's muq-feedstock repository is updated every time a new tag is created on the MUQ2 repository. To update the image on conda-forge, we then need to go into github and create a pull request to merge our fork into conda-forge's repo.