Create an EnsureMappings() constraint

[adswa]

EnsureMapping only takes a mapping with a single key-value pair. A check for required keys in meta-add would need a similar constraint that can take multiple key-value pairs.

[adswa]

I struggle to piece this together from smaller constraints. I feel like I might just miss an important piece of information to get this going:

Here's an example input:

{'type': 'file',
 'dataset_version': 'c81cf6189e2a87d3a0ad3c19e3eb6e8f3c84c121',
 'path': 'someii.txt',
 'extractor_name': 'metalad_example_file',
 'extractor_version': '0.0.1',
 'extraction_parameter': {},
 'extraction_time': 1677600832.9462137,
 'agent_name': 'Adina Wagner',
 'agent_email': '[email protected]',
 'extracted_metadata': {'@id': 'datalad:SHA1-s14--45fccab8df4dfa8e7077260ad45771bd25090bef',
  'type': 'file',
  'path': 'some.txt',
  'content_byte_size': 0,
  'comment': 'example file extractor executed at 1677600832.9461908'}}

The output needs to stay like this, but I require validation that a certain set of keys are included in the dict.
I could have EnsureTupleOf extract and return the keys from this dict, and EnsureContainsSet (proposed in #284) could validate that they indeed contain all required keys like this:

In [36]: from datalad_next.constraints import EnsureTupleOf, NoConstraint, EnsureContainsSet

In [37]: required_keys = (
    ...:         "type",
    ...:         "extractor_name",
    ...:         "extractor_version",
    ...:         "extraction_parameter",
    ...:         "extraction_time",
    ...:         "agent_name",
    ...:         "agent_email",
    ...:         "dataset_id",
    ...:         "dataset_version",
    ...:         "extracted_metadata")

In [38]: metadata = {"type": "file", "dataset_id": "fe46d177-d92a-4f0b-beac-8655e1aec75d", "dataset_version": "c81cf6189e2a
    ...: 87d3a0ad3c19e3eb6e8f3c84c121", "path": "some.txt", "extractor_name": "metalad_example_file", "extractor_version":
    ...: "0.0.1", "extraction_parameter": {}, "extraction_time": 1678198812.9966538, "agent_name": "Adina Wagner", "agent_e
    ...: mail": "[email protected]", "extracted_metadata": {"@id": "datalad:SHA1-s14--45fccab8df4dfa8e7077260ad45771
    ...: bd25090bef", "type": "file", "path": "some.txt", "content_byte_size": 0, "comment": "example file extractor execut
    ...: ed at 1678198812.9966106"}}

In [39]: EnsureContainsSet(*required_keys)(EnsureTupleOf(item_constraint=NoConstraint())(metadata))
Out[39]: 
('type',
 'dataset_id',
 'dataset_version',
 'path',
 'extractor_name',
 'extractor_version',
 'extraction_parameter',
 'extraction_time',
 'agent_name',
 'agent_email',
 'extracted_metadata')

However, at this point the value is keys-only. Do you see a way for me to perform such checks and preserve the JSON/dict structure?

[adswa]

NB: One perfectly fine outcome of this exploration could also be that a usecase like this is out of scope for the parameter constraint validation system

[adswa]

See also datalad/datalad-gooey#311