version: "1" # version of the workflow pipeline; as of 2023 May, there's only version 1
name: "nmulyk-baseband-processing" # name of your pipeline
event_number: &event-number 254820502 # for easy replacement later in the yaml file
event_str: &event-str "254820502"
defaults: # every object in the pipeline will have these parameters
  user: "nmulyk" # to keep track of who's submitting what job on the workflow website
  site: "canfar" # to run this at canfar
  event: [*event-str]
pipeline:
  stage_data:
    stage: 1
    work:
      pipeline: "baseband-nmulyk-datatrail" # name of the bucket this object will be deposited into and run from
      command:
        - "datatrail"
        - "pull"
        - "chime.event.baseband.raw"
        - *event-str
        - "-vvv" # verbose
        - "-f" # force
        - "-c" # cores
        - "4"
      tags:
        - "lowDM"
        - "stage-data"
      timeout: 10000
      priority: 3 # this is the default. 1 is lowest priority, 5 is highest priority.
  initialize:
    stage: 2
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.initialize
      parameters:
        event_number: *event-number
        site: canfar
      tags:
        - "lowDM"
        - "initialize"
      timeout: 600
  refinement_tiedbeam:
    stage: 3
    matrix:
      job_id:
        range: [1,64]
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.tiedbeam
      parameters:
        event_number: *event-number
        site: canfar
        job_id: ${{ matrix.job_id }}
        parameters: ["baseband_pipeline", "refinement", "job_size", "64", "files_per_group", "2"]
      tags:
        - "lowDM"
        - "refinement"
        - "tiedbeam"
      retries: 3
      timeout: 900
  refinement_merge:
    stage: 4
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.merge
      parameters:
        event_number: *event-number
        site: canfar
        parameters: ["baseband_pipeline", "refinement"]
      tags:
        - "lowDM"
        - "refinement"
        - "merge"
      timeout: 600
  refinement_analysis:
    stage: 5
    work:
      pipeline: "baseband-nmulyk-brightest"
      function: baseband_analysis.pipelines.cluster.cluster_cli.brightest
      parameters:
        event_number: *event-number
        site: canfar
        parameters: ["distributor_name", "baseband-refinement", "baseband_pipeline", "refinement", "num_parallel_processes", "8"] # num_parallel_processes speeds up the MCMC localization component of this stage
      tags:
        - "lowDM"
        - "refinement"
        - "analysis"
      timeout: 7200
  localization_tiedbeam:
    stage: 6
    matrix:
      job_id:
        range: [1,64]
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.tiedbeam
      parameters:
        event_number: *event-number
        site: canfar
        job_id: ${{ matrix.job_id }}
        parameters: ["baseband_pipeline", "localization", "job_size", "64", "files_per_group", "2", "get_parameters_from_h5", "true"]
      tags:
        - "lowDM"
        - "localization"
        - "tiedbeam"
      retries: 3
      timeout: 900
  localization_merge:
    stage: 7
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.merge
      parameters:
        event_number: *event-number
        site: canfar
        parameters: ["baseband_pipeline", "localization"]
      tags:
        - "lowDM"
        - "localization"
        - "merge"
      timeout: 600
  localization_analysis:
    stage: 8
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.localization
      parameters:
        event_number: *event-number
        site: canfar
        parameters: ["distributor_name", "baseband-localization"]
      tags:
        - "lowDM"
        - "localization"
        - "analysis"
      timeout: 600
  singlebeam_tiedbeam:
    stage: 9
    matrix:
      job_id:
        range: [1,64]
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.tiedbeam
      parameters:
        event_number: *event-number
        site: canfar
        job_id: ${{ matrix.job_id }}
        parameters: ["baseband_pipeline", "singlebeam", "job_size", "64", "files_per_group", "2", "get_parameters_from_h5", "true"]
      tags:
        - "lowDM"
        - "singlebeam"
        - "tiedbeam"
      retries: 3
      timeout: 900
  singlebeam_merge:
    stage: 10
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.merge
      parameters:
        event_number: *event-number
        site: canfar
        parameters: ["baseband_pipeline", "singlebeam"]
      tags:
        - "lowDM"
        - "singlebeam"
        - "merge"
      timeout: 600
  singlebeam_analysis:
    stage: 11
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.waterfall
      parameters:
        event_number: *event-number
        site: canfar
      tags:
        - "lowDM"
        - "singlebeam"
        - "analysis"
        - "waterfall"
      timeout: 600
  finalize:
    stage: 12
    work:
      pipeline: "baseband-nmulyk"
      function: baseband_analysis.pipelines.cluster.cluster_cli.finalize
      parameters:
        event_number: *event-number
        site: canfar
      tags:
        - "lowDM"
        - "finalize"
      timeout: 600
  unstage_data:
    stage: 12
    if: 'always' # should always run even if an earlier stage fails... in practice, doesn't seem to always be the case
    work:
      pipeline: "baseband-nmulyk-datatrail"
      command:
        - "datatrail"
        - "clear"
        - "chime.event.baseband.raw"
        - *event-str
        - "-vvv"
        - "-f" # force
      tags:
        - "lowDM"
        - "unstage-data"
      priority: 5