The Community Model supports the social cohesion across groups, by the understanding of their differences and recognizing what they have in common. The community model is responsible for storing information about explicit communities that users belong to. Additionally, it creates the implicit communities inferred from user interactions and it computes the metrics needed to define the similarity (and dissimilarity) among group of users. The Community Model will support the recommender system in the variety and serendipity to the recommendation results, that will not be oriented to the typically popular contents or based on providing similar contents to the users (the so called, filter bubble) but to the inter-group similarities and the intra-group differences.
-
Create a
.envfile in thedeployfolder using theenv.templatefile and following the instructions in it -
Deploy the Community Model using Docker:
docker-compose up
Follow the instructions in the deploy/env.template file to configure a development environment.
The API server is implemented using Node.js. Dependencies are available at package.json in the api-server folder. Main dependencies are:
Database is implemented using MongoDB
The CM API server is tested using Jest and supertest. To run the tests, mongodb must be launched, and run the following command in api-server folder:
npm test
Documentation for the Community Model is available at http://spice.fdi.ucm.es/
The content of this repository is distributed under Apache 2.0 License.