CMPUT 402 - Project Summary

Transcript

1. CMPUT 402 – Project Summary Victor Guana guana@ualberta.ca http://victorguana.blogspot.ca
2. 2. The project The data that characterize researchers, and the heterogeneous repositories where this data is maintained, make the construction of a unified author profile a non-trivial endeavour. Gathering and exploring information about academic researchers is a challenging task.
3. 2. The project The project goal was to develop an aggregator service to collect and fuse evidence of author's research impact and social media. The envisioned platform will integrate 6 major social and research websites. U of A U of A U of A U of A IIT Indore IIT Indore IIT Indore Centralized Repository
4. 3. The Development Team 4 Teams 3 Teams Total of 17 developers geographically distributed around the globe. Edmonton (Canada) and Indore (India).
5. 2. The project architecture
6. 2. The project architecture Original Sites e.g. twitter, ArXive, Mendeley.
7. 2. The project architecture Some of the sites expose access REST APIs with security layers such as oAuth. Sites like Google Scholar don’t provide access APis. In the architecture we call these access APIs “Extraction Interfaces”
8. 2. The project architecture An adapter component was developed for every site. In the case of Google Scholar a web crawler was built in order to extract the information from Google’s original site.
9. 2. The project architecture The goal of the adapters is to: 1. Extract the site’s information and model it in a simplified way, removing irrelevant data, and restructuring its semantics so they are more user friendly.
10. 2. The project architecture The goal of the adapters is to: 2. Deal with the security layer handshakes with a unified set of credentials.
11. 2. The project architecture The goal of the adapters is to: 3.Expose a REST API with concrete and simplified methods that can be called from web applications or third- party standalone/mobile applications. We call this simplified API the “Summary Interface” of the adapter.
12. 2. The project architecture The adapters have been deployed in the RAC space virtual infrastructure provided by Cybera.
13. 2. The project architecture A component called the Centralized Repository was developed in order to aggregate, in a single repository, information about an author using a unified data model that fuses the information available in all the repositories.
14. 2. The project architecture The centralized repository exposes a REST API which allows loading an author to the central database, an access his/her information. We call this API the “Access Interface”
15. 2. The project architecture The centralized repository has also been deployed using Cybera’s RAC infrastructure.
16. 2. The project architecture All the summary interfaces allow cross-origin resource sharing (CORS) as to allow web applications to dynamically call the adapters “Summary Interfaces”
17. 3. Available Documentation All the adapters and the centralized repository have the following documentation: 1. ER Model of the simplified version of each site’s information. 2. Security and request quota restriction descriptions. 3. Summary information API description including request and response examples. 4. Summary information API test cases and expected exception paths. 5. CURL commands ready to be executed to exercise all the API endpoints.