Provide examples of usage in JSON-LD

[cmungall]

unfortunately a large fraction of the bioinformatics community will be put off by RDF.

Luckily it's very easy to use JSON-LD as a trojan horse here. The bioinformatics community needs a JSON successor to GFF3 and it would be easy to place FALDO in this place

[JervenBolleman]

That makes sense, do you suggest doing this in the README.md? and also should we have a default context at biohackathon.org?

[JervenBolleman]

The fig9. in the paper is in JSON-LD

[cmungall]

On 22 Aug 2017, at 23:25, JervenBolleman wrote: That makes sense, do you suggest doing this in the README.md?

Ideally one example in the README and then some additional examples / unit tests elsewhere... maybe even a readthedocs for FALDO. But assuming not infinite time to work on this, yes just the README.md for now

and also should we have a default context at biohackathon.org?

Great idea. How do things currently get published there? I don't see a gh-pages or docs folder.

…

-- You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: #33 (comment)

[cmungall]

The fig9. in the paper is in JSON-LD

Ah yes great, we can just dupe that for now

[JervenBolleman]

Regarding a degault context at biohackathon.org, it currently checks out the codebase every hour if I recall @ktym is the maintainer. I think it something we should be able to do something in Japan. I will at least prepare some code to generate an context.

I was looking into having tabs with examples in the README. Otherwise we can do more regarding documentation.

[JervenBolleman]

I was planning to commit this but some key issue means I can't commit right now. Will do later

A minimal JSON-LD example

We have a feature starting at a position 1 and ending on position 2 of a sequence that is not described here.

{
  "@context":{
    "@vocab": "http://biohackathon.org/resource/faldo#",
    "feature" : {
      "@type" : "@id",
      "@id" : "http://purl.uniprot.org/core/annotation"
    }
  },
  "feature":{ "@id":"my-feature-id",
             "begin":{"position":1},
             "end":{"position":2}}
}

This uses non of the extensive advanced capabilities of JSON-LD or FALDO. It is a minimal example to show that any
JSON producing protein/dna/rna data service can easily be turned into an RDF one using just a bit of JSON-LD @context.

[cmungall]

Great start! As you say, we can use additional features of json-ld to make it more appealing to developers: we can have the @id field in the main body simply be id (it's possible to push all the @ stuff into the context header or linked document).

I note also this example creates blank nodes for the Positions. I think this is OK, and we can imagine skolemizing or UUIDing as part of a separate translation step if this is perceived to be an issue.

[cmungall]

May want to coordinate the json-ld example with bioschemas, see:

BioSchemas/specifications#210

@JervenBolleman your example above is probably too underspecified, we would need at least a reference

[JervenBolleman]

@cmungall yes, the example is completely underspecified as I was thinking of some of the existing public JSON for genomic/protein information. Which are like

{
    "features": [ {"type": "SIGNAL",
        "category": "MOLECULE_PROCESSING",
        "begin": "1",
        "end": "17"}
    ]
}

Which are common but put 3 things into one object making it hard to separate them out with just the JSON-LD processing (Maybe framing can help but I am not an expert there).

[cmungall]

I don't think JSON-LD processing is sufficient to transform the above into a FALDO-compliant RDF graph (in which each position is its own node).

One possibility is to make FALDO more lenient and allow flatter representations, and define some kind of rule-based transform to entail the deeper representation.