packageset_update_process.md

Steps required to deploy a new dependency package

Langpacks

1. Create new langpacks branch, from latest master
   • eg: git checkout -b ALGO-xxx-package-name
2. create new directory in libraries
   • eg: package-name-version
3. create new Dockerfile in new package directory
4. copy existing dependency template to new directory
   • eg: cp -LrR templates/package-name-version-1 templates/package-name-version
   • Update as necessary
5. run packageset_validator in tools to test it works
   • eg: ./tools/packageset_validator.py -g python3 -s python37 -d tensorflow-gpu-2.1 -t dependency -n tensorflow-gpu-2.1
   • If it doesn't work, lets figure out why not at this stage before moving forward
6. Push to langpacks repo, open a PR
   • get algo team & kenny to review, provide packageset_validator commands for reviewers to test locally
7. Merge PR into master

Legit

8. Create a new branch in legit, using git flow
   • git flow feature create ALGO-xxx-package-name
9. copy latest langpacks master hash (including the changes you made) to the LANGPACKS_VERSION variable in docker/algorithmia/legit/Dockerfile in a new branch
10. Open a PR in legit with your langpack version change
    • Get infrastructure team approval, ensure CI tools path; merge into develop

Stagetools

12. Create a new branch in stagetools, using gitflow
    • git flow feature create ALGO-xxx-package-name
13. Create stagetools packageset registration file in deployment/scripts/packages/beta
    • You can copy the a previous version's package set file, and tweak the variables to use the latest package that you created in langpacks.
14. Create a new PR
    • Get infrastructure team approval, ensure CI tools pass; merge into develop

• Wait for next test deploy or if urgent, request a supplimentary test deploy from the secondary oncall