I'm writing a samplesheet schema for a nested data structure. Each top-level array entry has five possible fields:

id

,

pacbio_hifi

,

oxford_nanopore

,

illumina_hic

and

illumina_10x

. How can I specify that

id

and at least /one/ of the other sections must be specified for an entry to be valid?

{
    "$schema": "<https://json-schema.org/draft/2020-12/schema>",
    "$id": "<https://raw.githubusercontent.com/test/test/master/assets/schema_input.json>",
    "title": "Test schema for samplesheets",
    "description": "Schema for the file provided with params.data_input",
    "type": "array",
    "items": {
        "type": "object",
        "properties": {
            "id": {
                "type": "string",
            },
            "pacbio_hifi": {
                "type": "object",
                "properties": {
                },
                "required": ["reads"]
            },
            "oxford_nanopore": {
                "type": "object",
                "properties": {
                },
                "required": ["reads"]
            },
            "illumina_hic": {
                "type": "object",
                "properties": {
                },
                "required": ["reads"]
            },
            "illumina_10x": {
                "type": "object",
                "properties": {
                },
                "required": ["reads"]
            }
        }
    }
}

You can add

"required": ["id"]

in the

items

scope for that

But that would mean that a section with /only/

id

would be valid - but I want to enforce specification of at least one other field

Then you can also add something like this:

"anyOf": [
    {"required": ["pacbio_hifi"]},
    ...
]

Where

...

is a repeat of the required section above with another field 😛

Do I need to specify every possible combination of fields? Or do I just have to specify each possible required field once

Each field once should do the trick 😛

Thanks!

You're welcome, let me know if something is not working as expected

I will! I am succumbing to expectations and re-writing this samplesheet to operate as a list, rather than as an object. So I'm very excited to be able to use samplesheetToList() on it!

Glad to hear that 😉

"anyOf": [
            {
                "dependentRequired": {
                    "maternal_illumina_dataset": ["paternal_illumina_dataset"]
                }
            },
            {
                "dependentRequired": {
                    "paternal_illumina_dataset": ["maternal_illumina_dataset"]
                }
            }
        ]

Is that the right way to specify a reciprocal optional dependency?

"dependentRequired": {
                    "paternal_illumina_dataset": ["maternal_illumina_dataset"],
                    "maternal_illumina_dataset": ["paternal_illumina_dataset"]
                }

Is enough here, but it does exactly the same as your code

Ah, that's cleaner!

Also, better to not use nested parameters if you can

This is a case of nested samplesheets 😛

Exactly, this is to specify a JSON/YAML input to the pipeline describing all the available data for each sample

Ah ok 👍🏻