Pipelines / DAG about qaboard

[arthur-flam]

Update: thanks to Itamar Persi and Ela Shahar, there is a pipeline implementation in "user-land":

my-pipeline:
  configs:
  - run: echo "Step 1"
  - batch: first-batch
  - batch:
    - second-batch
    - third-batch
    - label: batches running in parallel
  - run: some-postprocessing-script.py

Features include

• using PIPELINE_OUTPUT_DIR to save data across the batch
• providing to run steps info on the previous batch (what qa batch --list returns)

It's much simpler than a full DAG, and good enough in most cases.

Next steps

• We'll contribute it to the project as a default run() if the input type is pipeline
• Until then, we'll provide the code here on request as sample code (just comment here)...