fix --single-step step2_tool error

[huzech]

when use --single-step step2_tool parameter to run step in workflow meet this error

Traceback (most recent call last):
  File "/d/project/cwltool-main/cwltool/main.py", line 1181, in main
    ctool = choose_step(args, tool, loadingContext)
  File "/d/project/cwltool-main/cwltool/main.py", line 861, in choose_step
    extracted = get_step(tool, step_id, loading_context)
  File "/d/project/cwltool-main/cwltool/subgraph.py", line 225, in get_step
    name = outport.split("#")[-1].split("/")[-1]
AttributeError: 'CommentedMap' object has no attribute 'split'

[codecov]

Codecov Report

Merging #1597 (1bea0d7) into main (318876b) will decrease coverage by 0.03%.
The diff coverage is 100.00%.

❗ Current head 1bea0d7 differs from pull request most recent head 62745af. Consider uploading reports for the commit 62745af to get more accurate results

@@            Coverage Diff             @@
##             main    #1597      +/-   ##
==========================================
- Coverage   66.29%   66.26%   -0.04%     
==========================================
  Files          93       93              
  Lines       16503    16509       +6     
  Branches     4378     4380       +2     
==========================================
- Hits        10941    10940       -1     
- Misses       4408     4415       +7     
  Partials     1154     1154              

Impacted Files	Coverage Δ
cwltool/subgraph.py	80.60% <100.00%> (+0.35%)	⬆️
job.py	61.53% <0.00%> (-0.79%)	⬇️
subgraph.py	12.72% <0.00%> (-0.24%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 318876b...62745af. Read the comment docs.

[kinow]

mypy is failing. But it also worked when I tried a --single-step $param for a workflow.

$ cat /tmp/params.json 
{
  "file1": {
    "class": "File",
    "location": "/tmp/foo.txt"
  }
}
$ cwltool --debug --single-step step0 cwltool/schemas/v1.1/tests/count-lines11-extra-step-wf-noET.cwl /tmp/params.json
(...)
DEBUG Removing intermediate output directory /tmp/xrwwxp4z
DEBUG Removing intermediate output directory /tmp/7q408h7i
{
    "output": {
        "location": "file:///home/kinow/Development/python/workspace/cwltool/output",
        "basename": "output",
        "class": "File",
        "checksum": "sha1$da39a3ee5e6b4b0d3255bfef95601890afd80709",
        "size": 0,
        "path": "/home/kinow/Development/python/workspace/cwltool/output"
    }
}
INFO Final process status is success

[huzech]

Thank your test! I use list steps format in cwl, it's not work, I'll try to fix it with both list and object formt step

class: Workflow
cwlVersion: v1.0
id: subgraph_wrokflow
label: subgraph_wrokflow
inputs:
  - id: input
    type: File?
    default:
      class: File
      path: /d/project/cwlsubgraphtest/1.txt
  - id: txt
    type: string?
    default: step1
  - id: txt_1
    type: string?
    default: step2
  - id: txt_2
    type: string?
    default: step3
outputs:
  - id: output1
    outputSource:
      - step1_tool/output1
    type: File
  - id: output4
    outputSource:
      - step3_tool/output1
    type: File
  - id: output2
    outputSource:
      - step2_tool/output1
    type: File
steps:
  - id: step1_tool
    in:
      - id: input
        source: input
      - id: txt
        default: step1
        source: txt
    out:
      - id: output1
      - id: output2
    run: ./add_txt_to_file.cwl
    label: add_txt_to_file
  - id: step2_tool
    in:
      - id: input
        source: step1_tool/output2
      - id: txt
        default: step2
        source: txt_1
    out:
      - id: output1
      - id: output2
    run: ./add_txt_to_file.cwl
    label: add_txt_to_file
  - id: step3_tool
    in:
      - id: input
        source: step2_tool/output2
      - id: txt
        default: step3
        source: txt_2
    out:
      - id: output1
      - id: output2
    run: ./add_txt_to_file.cwl
    label: add_txt_to_file
requirements: []

[mr-c]

Thank you @huzech for this. Can you add your test case to the PR so we can confirm the fix and prevent regressions in the future?