workflow-api

Local setup

1. Some kind of k8s cluster (e.g. kind), kubectl and argo is required.
2. Disable authentication and HTTPS for argo (it's not necessary but very helpful for local development):

   kubectl patch deployment \
     argo-server \
     --namespace argo \
     --type='json' \
     -p='[{"op": "replace", "path": "/spec/template/spec/containers/0/args", "value": [
     "server",
     "--auth-mode=server",
     "--secure=false",
     "--access-control-allow-origin=*",
   ]},
   {"op": "replace", "path": "/spec/template/spec/containers/0/readinessProbe/httpGet/scheme", "value": "HTTP"}
   ]'

3. You can port-forward argo UI:

   kubectl -n argo port-forward --address 0.0.0.0 svc/argo-server -n argo 2746:2746

   UI will be available at http://localhost:2746
4. Create cluster role, service account and cluster role binding for argo:

   kubectl apply -f deploy/02-workflow-api-svcaccount.yaml

5. Install project dependencies with poetry:

   poetry shell
   poetry install

6. Run the FastAPI server:

   uvicorn src.main:app --reload