Azure auto running

Azure devops pipelines lets us run PUMA automatically!

Directories which are kept together when running PUMA locally are split out into different repositories in this implementation due to a variety of reasons:

• The actual PUMA code (https://github.com/OllyButters/puma)
• The cache is in a private repo (these are not allowed to be public due to licencing rules).
• The configs are in a private repo (these have API keys in them).
• The html output set up to be served as gh-pages (this is in a separate repo as it was making the source repo too large).

The configuration which runs the PUMA pipeline (spin up a VM, check out and manage all of the other repos, run the pipeline, commit files back) is in the azure-pipelines.yml file in the html repo (https://github.com/OllyButters/puma_web).

Every time the pipeline is run it (likely) generates new cache files and output html files. Overtime the history in the repos gets really large and it starts to take a long time to check them out. This is why they are separate repos and not in the main source repo any more. The puma_web repo serves the html from a branch (gh-pages), so every now and then you could delete the branch and make a new one with the same name. The next time the pipeline runs it will just add fresh copies of everything. Note that the azure-pipelines.yml file that is used by Azure Pipelines to run PUMA is in the main branch of the puma_web repo, the one in the branch with the html files is just a copy. The same principle for managing the repo history is true for the cache repo.