support id containing whitespaces

[prpfialho]

Hi,

Using the mlcroissant Python library to load a Croissant JSON exported from Dataverse, for instance using pydataverse:
ds = mlcroissant.Dataset(api.get_dataset_export(dset_doi, export_format='croissant').url)

I get the following validation error:
The dataset contains a wrong @id: 'a file name.csv'. Note that currently we do not support @ids containing whitespaces (not even if URL-escaped).

Does it make sense to have an option for Dataverse to automatically fix these errors, for instance upon file upload?

I'm using croissant-0.1.5.jar in a locally installed version of Dataverse 6.6.

[pdurbin]

@prpfialho hi! Thanks for the bug report. For now, please see my comment on your other issue: #21.

[prpfialho]

@prpfialho hi! Thanks for the bug report. For now, please see my comment on your other issue: #21.

this was more like a feature request, while the other you mention was more like a bug, but I couldn't figure how to apply the github labels.

[pdurbin]

@prpfialho you don't have permission but I just added them for you. Any ideas for me on #21?

[pdurbin]

Thanks, I'm able to reproduce this with the following:

(venv) pdurbin@beamish exporter-croissant % cat requirements.txt 
mlcroissant==1.0.22
pyDataverse==0.3.5
(venv) pdurbin@beamish exporter-croissant % cat test.py 
from pyDataverse.api import NativeApi
import mlcroissant

base_url = 'https://dataverse.harvard.edu'
dset_doi = 'doi:10.7910/DVN/Z9UB1H'
api = NativeApi(base_url)
export = api.get_dataset_export(dset_doi, export_format='croissant')
croissant_url = export.url
ds = mlcroissant.Dataset(str(croissant_url))
(venv) pdurbin@beamish exporter-croissant % 
(venv) pdurbin@beamish exporter-croissant % python test.py 
WARNING:absl:WARNING: The JSON-LD `@context` is not standard. Refer to the official @context (e.g., from the example datasets in https://github.com/mlcommons/croissant/tree/main/datasets/1.0). The different keys are: {'samplingRate'}
Traceback (most recent call last):
  File "/Users/pdurbin/github/gdcc/exporter-croissant/test.py", line 9, in <module>
    ds = mlcroissant.Dataset(str(croissant_url))
  File "<string>", line 6, in __init__
  File "/Users/pdurbin/github/gdcc/exporter-croissant/venv/lib/python3.13/site-packages/mlcroissant/_src/datasets.py", line 89, in __post_init__
    self.metadata = Metadata.from_file(ctx=ctx, file=self.jsonld)
                    ~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pdurbin/github/gdcc/exporter-croissant/venv/lib/python3.13/site-packages/mlcroissant/_src/structure_graph/nodes/metadata.py", line 457, in from_file
    return cls.from_json(ctx=ctx, json_=json_)
           ~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pdurbin/github/gdcc/exporter-croissant/venv/lib/python3.13/site-packages/mlcroissant/_src/structure_graph/nodes/metadata.py", line 471, in from_json
    return cls.from_jsonld(ctx=ctx, jsonld=jsonld)
           ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pdurbin/github/gdcc/exporter-croissant/venv/lib/python3.13/site-packages/mlcroissant/_src/structure_graph/base_node.py", line 404, in from_jsonld
    return cls(
        ctx=ctx,
    ...<2 lines>...
        **kwargs,  # type: ignore[arg-type]
    )
  File "<string>", line 46, in __init__
  File "/Users/pdurbin/github/gdcc/exporter-croissant/venv/lib/python3.13/site-packages/mlcroissant/_src/structure_graph/nodes/metadata.py", line 352, in __post_init__
    raise ValidationError(node.ctx.issues.report())
mlcroissant._src.core.issues.ValidationError: Found the following 1 error(s) during the validation:
  -  The dataset contains a wrong `@id`: 'Online Appendix.pdf'. Note that currently we do not support `@id`s containing whitespaces (not even if URL-escaped).
(venv) pdurbin@beamish exporter-croissant % 

As of this writing, https://dataverse.harvard.edu is running Dataverse 6.7.1 with the Croissant 0.1.5 exporter.

The dataset under test is https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/Z9UB1H and in the screenshot below you can see that the file "Data (CSV)" has a space in it:

I reported the issue upstream:

• clarify where to put file paths (e.g ml-25m/ratings.csv) mlcommons/croissant#639 (comment)