apiVersion: dataflow.argoproj.io/v1alpha1
kind: Pipeline
metadata:
  annotations:
    dataflow.argoproj.io/description: |-
      This pipeline count the number of words in a document, not the number of count of each word as you might expect.

        It also shows an example of a pipelines terminates based on a single step's status.
    dataflow.argoproj.io/needs: word-count-input-configmap.yaml
    dataflow.argoproj.io/owner: argoproj-labs
    dataflow.argoproj.io/test: 'false'
  name: 201-word-count
spec:
  steps:
  - container:
      args:
      - sh
      - -c
      - cat /in/text | tee -a /var/run/argo-dataflow/out
      image: ubuntu:latest
      volumeMounts:
      - mountPath: /in
        name: in
    name: read-text
    sinks:
    - stan:
        subject: lines
    sources:
    - kafka:
        topic: input-topic
    volumes:
    - configMap:
        name: word-count-input
      name: in
  - container:
      args:
      - bash
      - -c
      - |-
        set -eux -o pipefail
        while read sentence; do
          for word in $sentence; do
              echo $word
          done
        done > /var/run/argo-dataflow/out < /var/run/argo-dataflow/in
      image: ubuntu:latest
      in:
        fifo: true
    name: split-lines
    scale:
      desiredReplicas: minmax(pending, 0, 1)
    sinks:
    - stan:
        subject: words
    sources:
    - stan:
        subject: lines
  - container:
      args:
      - bash
      - -c
      - |-
        set -eux -o pipefail
        i=0
        while read word && [ $word != EOF ]; do
          i=$((i+1))
        done < /var/run/argo-dataflow/in
        echo $i > /var/run/argo-dataflow/out
      image: ubuntu:latest
      in:
        fifo: true
    name: count-words
    sinks:
    - kafka:
        topic: output-topic
    sources:
    - stan:
        subject: words
    terminator: true