The project treon (https://github.com/ReviewNB/treon) was recently brought to my attention. This project would allow us to more easily run notebooks as "tests" and know how many succeeded and how many failed. If we create travis jobs that are getting the new versions of pytroll packages (satpy, pyspectral, etc) and run them on cron (every week?) this would really help show us where examples are failing and are out of date.
Would also be good for testing dependency changes that break satpy usage but may be out-of-scope for our packages to test specifically (ex. creating cartopy plots from satpy data).
The project
treon(https://github.com/ReviewNB/treon) was recently brought to my attention. This project would allow us to more easily run notebooks as "tests" and know how many succeeded and how many failed. If we create travis jobs that are getting the new versions of pytroll packages (satpy, pyspectral, etc) and run them on cron (every week?) this would really help show us where examples are failing and are out of date.Would also be good for testing dependency changes that break satpy usage but may be out-of-scope for our packages to test specifically (ex. creating cartopy plots from satpy data).