I’m new to the OHDSI community and recently I’ve been working through the setup documentation for Atlas and been recording my steps in a series of ansible playbooks. Ansible playbooks are just YAML files, so they act almost like runnable documentation.
The playbooks in the repo atlas_dev_setup can be used with Vagrant to build a local instance of Atlas for development / experimentation, but I’ve also used those playbooks to deploy Atlas to a test server within our private network. I thought something like this might eventually be useful to others, especially those interested in getting a development environment setup, since Vagrant syncs the source code with your local machine, so you can have the server running in the VM, while using your preferred editor locally for development.
This is still very much a work in progress, so let me know if you find it useful or see anything that needs updating. Right now, the synpuf1k dataset is loaded, but if you look at the data playbook you can see how you could alter for other datasets (synpuf5pct is commented out, but if you uncomment and increase the VM size in the Vagrant file it should work).
There are still several things that I wasn’t sure about, so listing those below if anyone has any guidance regarding these issues:
- I was using postgres and the most recent version of the CDM has some DDL code for the results table, but I wasn’t sure what was supposed to be done pre-version 6.0, so I just reused the CDM DDL code for the results table. However, this requires that I run some additional SQL from https://github.com/OHDSI/Atlas/issues/1055 after everything has finished in order for the cohort generation tab to work.
- When I load the
synpuf5pctdataset the row counts are not showing up when inspecting concept sets. This just recently started happening. I saw this issue https://github.com/OHDSI/Atlas/issues/1190, so wasn’t sure if something changed, or if it was just setup issue.