datasources
(TODO: Document how the datasources system works and how it can be configured. Show the steps for a simple datasource integration.)
The idea is to allow any REST or SPARQL API to be added to a Conzept system and render topics in the sidebar.
Features:
- Make adding a new datasource as easy as possible, but still flexible in terms of the rendered output.
- Make the API search results show up in the same style and results paging mechanics, as the other datasources.
- Let the topic results be augmented from Wikidata (if a Qid is provided for the topic).
- Have auto-completion work too (if wanted and possible).
- Allow the user to choose which datasources to use.
use cases
- Allow for adding extra “data sources” to a Conzept system, so other concepts - beyond Wikipedia and Wikidata - can also be presented.
- Any organization that wants to use the Conzept user-interface for showing and augmenting their own knowledge base.
status
- The datasources system is now working (configuration example: src/data/datasources.js) and has been tested with various datasources (with REST and SPARQL access).
- See: datasource issues.
- Future improvements:
- Add global search options like: sort-key (relevance, date-asc, date-desc) and media-types filter (text, image, audio, video, 3d, etc.).
- Indicate when a datasource is “unreachable”.
- Add hooks for pre- and post-processing of the API data.
- Add hooks to enrich the API data (with Wikipedia and Wikidata information)
- Research support for self-hosted, data files used as a datasource: JSON, Parquet, CSV, RDF
adding a new datasource
- Add the new datasource to the datasource-configuration file here: src/data/datasources.js
- Add a new datasource directory (using the name of the datasource, eg. “gleif”)
- Add a new JS file in that directory, implementing the required API access functions, example: src/datasources/gleif/index.js
×
>
<