Extracting data from a Web API¤
This tutorial demonstrates how you can build a Knowledge Graph based on input data from a Web API. The tutorial is based on the GitHub API (v3), which we will use to fetch repository data of a certain organization and create a Knowledge Graph from the response.
The complete tutorial is available as a project file. You can import this project:
In order to get familiar with the API, simply fetch an example response with this command:
The HTTP Get request retrieves all repositories of a GitHub organization named vocol.
The JSON response includes the data for all repositories (mobivoc, vocol, …). You can also download the response file here: repos.json.
1 Register a Web API¤
Click the Create button (top right) in the data integration workspace and select the type REST request.
Define a Label, Description and the URL of the Web API. Example input:
2 Create a JSON parser¤
As we are only interested in the HTTP Message Body which holds the JSON repository data, we first have to parse the body from the entire HTTP response.
Click the Create button (top right) in the data integration workspace and select the type Parse JSON.
Define a Label, a Description, and the Input path. All other fields can keep the default settings. The default input path is always:
3 Create a JSON Dataset¤
To create a JSON-to-RDF-mapping within Corporate Memory, we have to first register an example response from the API (repos.json). Based on the schema of the response, we can then define step-by-step the mappings, which are used to build the Knowledge Graph.
Click the Create button (top right) in the data integration workspace and select the type JSON.
Upload the JSON file repos.json (API response) as a Dataset into Corporate Memory.
4 Create a Knowledge Graph¤
The Knowledge Graph will be used to integrate all data coming from one or more APIs. The Knowledge Graph receives RDF triples from the defined Transformations for each API.
Click the Create button (top right) in the data integration workspace and select the type Knowledge Graph.
Provide the Knowledge Graph with a Label and Description, as well as the following (example) Graph URI:
5 Create a Transformation¤
In order to transform the input data from the API, in our example structured in JSON, we have to define a mapping to create RDF triples which are then written to the Knowledge Graph.
Click the Create button (top right) in the data integration workspace and select the type Transform.
Provide the Transformation with a Label and Description, configure the Input Dataset (Repos.json) and the Output Dataset (Repository Knowledge Graph).
Click the Mapping Editor button in the previously defined Transformation.
In the following screenshots, we provide an example mapping for the data received by the GitHub API. For more complex mappings, we recommend the Tutorial Lift data from JSON and XML sources.
6 Create a Workflow¤
To build a workflow that combines all the elements we previously built, we now define a workflow for (1) requesting the data from the GitHub API, (2) parsing the HTTP response we receive, (3) transforming the JSON data into RDF triples and finally (4) writing the RDF triples into the Knowledge Graph.
Click the Create button (top right) in the data integration workspace and select the type Workflow.
Provide the Transformation with a Label and a Description.
Click the Workflow Editor button in the menu of the created workflow.
Drag and drop the different items into the Workflow Editor and combine them with one another (see example screenshot). Save the workflow, and press the run symbol to execute the workflow.
Validate the result by clicking on the Workflow Report tab and see the result of your execution. In this example, 15x repositories were found from the GitHub API request.