Implement migrator that creates/populates workflow_chain_set (file: migrator_from_X_to_PR9.py)

[brynnz22]

Migrator creates the workflow_chain_set and maps all WorkflowExecution subclasses to a workflow chain instance by the omics processing id in the was_informed_by slot.

• Migrator needs to come Before class name change from OmicsProcessing to DataGeneration as it refers to the omics_processing_set.

Address changes in PR #9

[eecavanna]

Here's the command I'm using to run the doctests in this migrator:

poetry run python -m doctest -v nmdc_schema/migrators/migrator_from_X_to_PR9.py

[eecavanna]

I'll focus on these two issues:

1. "No adapter was specified" message appears

Trying:
    m = Migrator()
Expecting nothing
No adapter was specified. Migration capability will be limited.

2. populate_workflow_chain doctest fails

Trying:
    m.populate_workflow_chain()
Expecting:
    # After execution, the workflow_chain_set collection should contain a document with the following data:
    {'id': 'nmdc:wfc-456', 'was_informed_by': 'nmdc:omcp-123', 'analyte_category': 'metagenome', 'type': 'nmdc:WorkflowChain', 'name': 'Workflow Chain for metagenome analysis related to Study 1'}
**********************************************************************
File "/nmdc-schema/nmdc_schema/migrators/migrator_from_X_to_PR9.py", line 169, in migrator_from_X_to_PR9.Migrator.populate_workflow_chain
Failed example:
    m.populate_workflow_chain()
Exception raised:
    Traceback (most recent call last):
      File "/usr/local/lib/python3.9/doctest.py", line 1334, in __run
        exec(compile(example.source, filename, "single",
      File "<doctest migrator_from_X_to_PR9.Migrator.populate_workflow_chain[4]>", line 1, in <module>
        m.populate_workflow_chain()
      File "/nmdc-schema/nmdc_schema/migrators/migrator_from_X_to_PR9.py", line 191, in populate_workflow_chain
        self.adapter.insert_document("workflow_chain_set", workflow_chain_doc)
    AttributeError: 'NoneType' object has no attribute 'insert_document'

[eecavanna]

The log from the most recent GitHub Actions failure says:

  RuntimeError

  Unable to find installation candidates for bioregistry (0.10.157)

The installation of a Python package failed. The failure is not related to the doctests.

I opened a ticket in nmdc-schema (microbiomedata#1932) about this failure. The failure is occurring in that repository also.

[brynnz22]

The log from the most recent GitHub Actions failure says:

  RuntimeError

  Unable to find installation candidates for bioregistry (0.10.157)

The installation of a Python package failed. The failure is not related to the doctests.

I opened a ticket in nmdc-schema (microbiomedata#1932) about this failure. The failure is occurring in that repository also.

Katherine has a PR up to solve this, so its being worked on.

[eecavanna]

Implementation of this migrator is currently blocked by the Minter (or, rather, "a Minter") not allowing people to mint IDs for WorkflowChains. I consider myself — @eecavanna — responsible for resolving that (with respect to this PR).

[brynnz22]

@eecavanna @turbomam this migrator is finally ready for a final review and can be merged into main afterwards. The workflochain ids have been pre-minted. The change with the project.makefile was because I edited it and then recently change it back to match main - so it should match main. Not sure why its showing those interesting changes.

[eecavanna]

I'm reviewing this now. I will make some changes to variable names and push a new commit today.

[eecavanna]

As @brynnz22 said, the lines that GitHub is showing as being added to the project.Makefile via this PR, already exist in that file on the tip of the main branch. I don't understand why GitHub is showing them as new lines in this PR. 🤷

[eecavanna]

Looks good to me! I am comfortable with this being merged into the main branch. Once it's merged in, it'll be easier for me to test "all Berkeley migrators" in series.

[eecavanna]

Hi @turbomam, once you merge this branch into main, will you also publish a new prerelease package to PyPI? I think publishing a new package to PyPI will unblock microbiomedata/nmdc-runtime#519 (comment) (I'm getting some path-related errors when installing the package directly from the GitHub repo as opposed to from PyPI, and I suspect it's because what's in the GitHub repo hasn't undergone whatever build steps GitHub Actions performs when building a package that will be published to PyPI).