version: '3'

tasks:
  main:
    desc: Powerhouse Museum Tutorial
    cmds:
      - task: refine
      - task: :check # check OpenRefine log for any warnings and exit on error
        vars: {PROJECT: '{{splitList ":" .TASK | first}}'} 

  refine:
    vars:
      PORT: 3336 # assign a different port for each project
      RAM: 2048M # maximum RAM for OpenRefine java heap space
      PROJECT: '{{splitList ":" .TASK | first}}'
    deps: # will be executed each run independent of up-to-date check
      - task: download
    cmds: # tasks prepended with ":" are defined in Taskfile.yml
      - task: :start
        vars: {PROJECT: '{{.PROJECT}}', PORT: '{{.PORT}}', RAM: '{{.RAM}}'}
      - task: import
        vars: {PROJECT: '{{.PROJECT}}', PORT: '{{.PORT}}'}
      - task: apply
        vars: {PROJECT: '{{.PROJECT}}', PORT: '{{.PORT}}'}
      - task: export
        vars: {PROJECT: '{{.PROJECT}}', PORT: '{{.PORT}}'}
      - task: stats
        vars: {PROJECT: '{{.PROJECT}}', PORT: '{{.PORT}}'}
      - task: :stop
        vars: {PROJECT: '{{.PROJECT}}', PORT: '{{.PORT}}'}
    sources:
      - input/**
      - config/**
    generates:
      - output/openrefine.log
      - output/{{.PROJECT}}.openrefine.tar.gz
    ignore_error: true # workaround to avoid an orphaned Java process on error https://github.com/go-task/task/issues/141

  download:
    cmds:
      - mkdir -p input config
      - wget --no-verbose -O input/phm-collection.tsv https://github.com/opencultureconsulting/openrefine-batch/raw/master/examples/powerhouse-museum/input/phm-collection.tsv
      - wget --no-verbose -O config/phm-transform.json https://github.com/opencultureconsulting/openrefine-batch/raw/master/examples/powerhouse-museum/config/phm-transform.json

  import:
    dir: input
    cmds:
      - | # import file
        ../../openrefine/client -P {{.PORT}} \
        --create phm-collection.tsv \
        --processQuotes false \
        --guessCellValueTypes true \
        --projectName {{.PROJECT}}
    ignore_error: true # workaround

  apply:
    dir: config
    cmds:
      - | # apply transformation rules
        ../../openrefine/client -P {{.PORT}} {{.PROJECT}} \
        --apply phm-transform.json
    ignore_error: true # workaround

  export:
    dir: output
    cmds:
      - | # export to file; use readlink to log full path to output file
        ../../openrefine/client -P {{.PORT}} {{.PROJECT}} \
        --output "$(readlink -m phm-results.tsv)"
    ignore_error: true # workaround

  stats:
    cmds:
      - ps -o start,etime,%mem,%cpu,rss -p $(lsof -t -i:{{.PORT}}) # print allocated system resources
    ignore_error: true # workaround

  default: # enable standalone execution (running `task` in project directory)
    cmds:
      - PROJECT="${PWD##*/}:main" && cd .. && task "$PROJECT"