# Bookstore — Part 12 ch.07 "ML pipelines and workflows": Argo Events
# `EventSource` — a webhook listener that turns "a new training dataset is
# available" (an HTTP POST from the dataset-publishing pipeline OR a
# Git provider's webhook) into an event the `recommender-sensor.yaml`
# `Sensor` translates into a Workflow run.
#
# !!! CRD-INTRINSIC DRY-RUN (identical precedent to recommender-workflow.yaml
#     in this dir, raw-manifests/51-/70-/83-, argocd/, operators/, chaos/,
#     ml/batch/, ml/serve/recommender-inferenceservice.yaml) !!!
#   `EventSource` is an Argo Events CRD (argoproj.io/v1alpha1). WITHOUT Argo
#   Events installed a client dry-run prints:
#     no matches for kind "EventSource" in version "argoproj.io/v1alpha1"
#   EXPECTED and SCHEMA-CORRECT — install Argo Events first (Part 12 ch.07
#   Hands-on: pinned Helm `argo/argo-events --version 2.4.7` -> ns `argo-events`,
#   plus the default `EventBus` from the same chart). Schema verified against
#   argoproj.io/v1alpha1 (service + webhook event source endpoints).
#
# WHAT IT DOES
#   Listens on `:12000` (the Argo Events default for webhook event sources)
#   at the path `/recommender-dataset-ready`. An HTTP POST with a JSON body
#   (e.g. `{"dataset_uri": "s3://bookstore/datasets/2026-05-19/"}`) becomes
#   an event picked up by the Sensor in the next file.
#
# WHY illustrative (kept small)
#   In prod the EventSource is a `git`/`gitlab`/`github`/`sqs`/`pubsub`
#   event source; the webhook one is the smallest, most-portable, kind-
#   testable shape (`curl http://recommender-dataset.../recommender-dataset-ready`
#   from inside the cluster). Part 12 ch.07 calls out the prod replacements.
#
# PRODUCTION NEEDS AUTH — this webhook EventSource has none. In prod, put it
# behind an Ingress with auth (OAuth2 Proxy, mTLS, token header check) OR use
# a git/s3/pubsub EventSource where authentication is part of the source itself.
apiVersion: argoproj.io/v1alpha1
kind: EventSource
metadata:
  name: recommender-dataset
  namespace: argo-events
  labels:
    app.kubernetes.io/part-of: bookstore-ml
    app.kubernetes.io/component: ml-pipeline-events
spec:
  # Argo Events scales the EventSource as its own Pod(s) (`service.ports`
  # creates a Service in front). One small Pod is plenty for nightly /
  # on-commit events.
  service:
    ports:
      - port: 12000
        targetPort: 12000
  webhook:
    # The route name `dataset-ready` becomes the path `/recommender-dataset-ready`
    # at the Pod's :12000.
    dataset-ready:
      port: "12000"
      endpoint: /recommender-dataset-ready
      method: POST