Create a Dataset from a list of files

[huard]

Hi, newbie here,

I'd like to create a Dataset inside the root Dataset with files of my choosing. The way I understand how the code works, I can either manually add File objects, which will appear in a flat list in the root Dataset, or a Dataset object, which will gather everything in that directory.

Is there a way to create a Dataset from a list of files? The motivation for this is to be able to assign authors and a license to that Dataset.

I'm able to create the correct JSON-LD by appending a Dataset using the ContextEntity, but it's not copying the files to the crate.

Thanks,

[simleo]

Suppose you have this layout on the filesystem:

mkdir -p exp/data
touch exp/paper.pdf
touch exp/data/results.csv
touch exp/data/diagram.svg

To build a crate where data is a Dataset, you can do:

from rocrate.rocrate import rocrate

crate = ROCrate()
crate.add_file("exp/paper.pdf")
crate.add_dataset("exp/data")
crate.write("exp_crate")

You will get the following crate layout:

exp_crate/
|-- data
|   |-- diagram.svg
|   `-- results.csv
|-- paper.pdf
`-- ro-crate-metadata.json

And the metadata will be:

{
    "@context": "https://w3id.org/ro/crate/1.2/context",
    "@graph": [
        {
            "@id": "./",
            "@type": "Dataset",
            "datePublished": "2026-07-17T08:00:26+00:00",
            "hasPart": [
                {
                    "@id": "paper.pdf"
                },
                {
                    "@id": "data/"
                }
            ]
        },
        {
            "@id": "ro-crate-metadata.json",
            "@type": "CreativeWork",
            "about": {
                "@id": "./"
            },
            "conformsTo": {
                "@id": "https://w3id.org/ro/crate/1.2"
            }
        },
        {
            "@id": "paper.pdf",
            "@type": "File"
        },
        {
            "@id": "data/",
            "@type": "Dataset"
        }
    ]
}

If you also want the directory contents to show up as File entities in the metadata, you have to add them with add_file:

crate.add_file("exp/data/results.csv", dest_path="data/results.csv")
crate.add_file("exp/data/diagram.svg", dest_path="data/diagram.svg")

[huard]

Yes, but all the files inside data/ are going to be included in the crate when I write it to disk. I would like to have explicit control on which files are included in the Dataset.

What I want is

      {
            "@id": "data/",
            "@type": "Dataset"
            "hasPart": [
                {
                    "@id": "data/results.csv",
                    "@type": "File"
                },
                            ]
        }

So instead of add_files appending files to the root dataset, it would be appending files to the data/ Dataset.

Does that make sense ?

[simleo]

You can add an empty Dataset and then populate it only with the files you want:

from rocrate.rocrate import ROCrate

crate = ROCrate()
crate.add_file("exp/paper.pdf")
data = crate.add_dataset(dest_path="data")
results = crate.add_file("exp/data/results.csv", dest_path="data/results.csv")
data.append_to("hasPart", results)
crate.write("exp_crate")

In this case, the crate layout is:

exp_crate/
|-- data
|   `-- results.csv
|-- paper.pdf
`-- ro-crate-metadata.json

And the metadata:

{
    "@context": "https://w3id.org/ro/crate/1.2/context",
    "@graph": [
        {
            "@id": "./",
            "@type": "Dataset",
            "datePublished": "2026-07-20T08:25:42+00:00",
            "hasPart": [
                {
                    "@id": "paper.pdf"
                },
                {
                    "@id": "data/"
                },
                {
                    "@id": "data/results.csv"
                }
            ]
        },
        {
            "@id": "ro-crate-metadata.json",
            "@type": "CreativeWork",
            "about": {
                "@id": "./"
            },
            "conformsTo": {
                "@id": "https://w3id.org/ro/crate/1.2"
            }
        },
        {
            "@id": "paper.pdf",
            "@type": "File"
        },
        {
            "@id": "data/",
            "@type": "Dataset",
            "hasPart": [
                {
                    "@id": "data/results.csv"
                }
            ]
        },
        {
            "@id": "data/results.csv",
            "@type": "File"
        }
    ]
}

[huard]

But if I run this example (version 0.10.0) and look into exp_crate/data, I see diagram.svg, which was never explicitly added to the crate.

[simleo]

But if I run this example (version 0.10.0) and look into exp_crate/data, I see diagram.svg, which was never explicitly added to the crate.

I can't reproduce this, maybe you have leftover files from a previous run. Try restarting from scratch.