Representing HCA Bundles as JSON-LD allows them to be easily converted to RDF and queried using SPARQL. This software, when combined with a graph or triplet store can provide semantic data models that allow for an expressive query interface.
The "bundle_to_rdf" module presents a method for converting HCA bundles to RDF ttl files.
import requests
import bundle_to_rdf
DSS_URL = "https://dss.dev.data.humancellatlas.org/v1"
bundle_uuid = "4be0071d-b36e-4414-a7ee-7b879f60be7c"
r = requests.get("{}/bundles/{}?replica=aws".format(DSS_URL, bundle_uuid))
bundle = r.json()
file_name = bundle_to_rdf.bundle_to_rdf(bundle)
print(file_name)
A file is created in the current working directory with a filename following the pattern:
{bundle_uuid}.ttl.
This file can then be loaded into a triplet store of your choosing. For an example of usage with BlazeGraph,
check out the example-usage.ipynb.
This module should make as few assumptions about the DSS data model as possible. Future developments would use this module as part of a service to either create RDF that can be queried directly, or loaded into a graph store like Blazegraph or Amazon Neptune.
- Add unit tests, the module currently relies on active connections to the DSS
- Improve example notebook with biological use cases
- Demonstrate using the SPARQL and RDF interface to demonstrate tabular representations
- Integrate into a microservice for generating RDF from bundles on the fly
- Please add any issues you experience to our Github issue tracker!
- This software is Open Source, Apache licensed, send in your changes!