1. 问题
    2. 我如何描述哪些参数必须和不能组合使用?
    3. 目标
    4. 学习如何使用记录来描述输入之间的关系。

    有时，一个基础工具有几个必须一起提供的参数(它们是相互依赖的)或几个不能一起提供的参数(它们是互斥的)。您可以使用记录和类型联合将参数组合在一起来描述这两个条件。
    record.cwl

    1. #!/usr/bin/env cwl-runner
    3. cwlVersion: v1.0
    4. class: CommandLineTool
    5. inputs:
    6.   dependent_parameters:
    7.     type:
    8.       type: record
    9.       name: dependent_parameters
    10.       fields:
    11.         itemA:
    12.           type: string
    13.           inputBinding:
    14.             prefix: -A
    15.         itemB:
    16.           type: string
    17.           inputBinding:
    18.             prefix: -B
    19.   exclusive_parameters:
    20.     type:
    21.       - type: record
    22.         name: itemC
    23.         fields:
    24.           itemC:
    25.             type: string
    26.             inputBinding:
    27.               prefix: -C
    28.       - type: record
    29.         name: itemD
    30.         fields:
    31.           itemD:
    32.             type: string
    33.             inputBinding:
    34.               prefix: -D
    35. outputs:
    36.   example_out:
    37.     type: stdout
    38. stdout: output.txt
    39. baseCommand: echo

    record-job1.yml

    1. dependent_parameters:
    2.   itemA: one
    3. exclusive_parameters:
    4.   itemC: three

    输出为

    1. $ cwl-runner record.cwl record-job1.yml
    2. Workflow error, try again with --debug for more information:
    3. Invalid job input record:
    4. record-job1.yml:1:1: the `dependent_parameters` field is not valid because
    5.                        missing required field `itemB`

    在这个例子上，不能提供itemA就不能提供itemB。

    1. record-job2.yml
    3. dependent_parameters:
    4.   itemA: one
    5.   itemB: two
    6. exclusive_parameters:
    7.   itemC: three
    8.   itemD: four

    输出为

    1. $ cwl-runner record.cwl record-job2.yml
    2. record-job2.yml:6:3: invalid field `itemD`, expected one of: 'itemC'
    3. [job record.cwl] /home/example$ echo \
    4.     -A \
    5.     one \
    6.     -B \
    7.     two \
    8.     -C \
    9.     three > /home/example/output.txt
    10. [job record.cwl] completed success
    11. {
    12.     "example_out": {
    13.         "location": "file:///home/example/11-records/output.txt",
    14.         "basename": "output.txt",
    15.         "class": "File",
    16.         "checksum": "sha1$329fe3b598fed0dfd40f511522eaf386edb2d077",
    17.         "size": 23,
    18.         "path": "/home/example/output.txt"
    19.     }
    20. }
    21. Final process status is success
    22. $ cat output.txt
    23. -A one -B two -C three

    第二个例子中，itemC和itemD是独立的，因此只有itemC被调用，itemD被忽略。
    record-job3.yml

    1. dependent_parameters:
    2.   itemA: one
    3.   itemB: two
    4. exclusive_parameters:
    5.   itemD: four

    输出为

    1. $ cwl-runner record.cwl record-job3.yml
    2. [job record.cwl] /home/example$ echo \
    3.     -A \
    4.     one \
    5.     -B \
    6.     two \
    7.     -D \
    8.     four > /home/example/output.txt
    9. [job record.cwl] completed success
    10. {
    11.     "example_out": {
    12.         "location": "file:///home/example/output.txt",
    13.         "basename": "output.txt",
    14.         "class": "File",
    15.         "checksum": "sha1$77f572b28e441240a5e30eb14f1d300bcc13a3b4",
    16.         "size": 22,
    17.         "path": "/home/example/output.txt"
    18.     }
    19. }
    20. Final process status is success
    21. $ cat output.txt
    22. -A one -B two -D four

    第三个例子中，只有itemD是被调用的。