A6S API: sparql Query API

This API uses named sparql queried to create a API adapter into the Governance Graph.

It's implemented as a microservice that maps URL sub-paths to named sparql queries.

The config/default.yaml is used to configure the chassis, specificially the 'sparqlclient' plugin.

The built-in "openapi" plugin is used to map /:query/ requests to the 'sparqlclient' operation.

Query Processing

The ./sparql/ folder contains a set of named sparql queries.

The API path is parsed to extract the named query.

The requests query parameters can be bound to named query variables.

Consider the query "this.sparql":

SELECT DISTINCT ?name ?value WHERE { ?this ?name ?value}

The "?this" variable can be bound at request time using the "?this" query parameter:

CURL http://localhost:7000/gg/this?this=http://example.com/

Special Variable: ?this

All query parameters are interpreted as string literals except for ?this which is always converted into a URI.

SparqlClient plugin

The sparqlclient is an opionated way to execute sparql queries:

1. translates a query name into a sparql filename
2. read sparql file from "queryPath" folder
3. forward the sparql query to the remote endpoint
4. format the results
5. handle errors

Options

The "sparqlclient" options can be specified as a global config or within relevent the "openapi" definition.

sparqlclient:
  endpoint: http://localhost:8080/rdf4j-server/repositories/example
  queryPath: "./sparql"
  ns:
    rdf: http://www.w3.org/1999/02/22-rdf-syntax-ns#
    rdfs: http://www.w3.org/2000/01/rdf-schema#
    dc: http://purl.org/dc/elements/1.1/
    eg: http://www.example.org/

The "endpoint" refers to an external sparql 1.x HTTP endpoint. The "queryPath" refers to the folder containing the sparql queries. The "ns" is a namespace map used as defaults for the sparql queries.

openapi plugin

The chassis is configured by an OpenAPI definition. The "operationId" fields must be mapped to a valid plugin.

The "operationId" can be used in more than one path - which is different from the Swagger/Open API specification.

openapi:
  paths:
    /healthz:
      get:
        chassis:
          operationId: heartbeat
    /swagger:
      get:
        chassis:
          operationId: apidocs

    /types:
      get:
        chassis:
          operationId: SPARQL
        params:
          query: rdfs_types
    "/:query":
      get:
        chassis:
          operationId: SPARQL

For example:

$ npm install && npm start
$ curl -v http://localhost:7001/types

This /types/ API is mapped to the rdfs_types.sparql file in the ./sparql/ folder.

Governance Graph

The govgraph.ttl file contains a a6s:Solution graph describing the OpenAPI and it's runtime components.

:api-service-sparql
    a               a6s:Solution;
    a               k8s:DockerImage;
    a               k8s:HelmChart;

    dc:title        "A6S APIs: Core Insights";
    k8s:name        "api-service-sparql";

    a6s:feature     <https://api.apigeeks.com/v1/gg/sparql/:query>;
    a6s:feature     <https://api.apigeeks.com/v1/gg/healthz/sparql>;
    a6s:feature     <https://api.apigeeks.com/v1/gg/openapi/sparql>;
.

Docker Demo

docker run --rm -p 8080:8080 yyz1989/rdf4j

curl -H "Content-Type: application/json" -X POST --data '{"this":"https://api.apigeeks.com/v1/gg/openapi/sparql/"}' http://localhost:7001/gg/this