- Owned by the African Science community
- Distrbuted hosting on African territory
- Open source
- Open access
- Free for users
African research output should be owned and hosted on African territory. We therefore propose here an African smart platform and portal for preprint uploads as well as the aggregation of African research output including:
- Scholarly books
- Journal articles
- List of digital African scholarly journals
- …
The portal will be hosted by African research institutions in each of the 5 regions on the continent. To achieve this we will build a decentralized infrastructure with the following features:
- Core distributed data engine with intelligent aggregation and search models
- The data platfrom components are hosted by institutions in African regions (EastA, WestA, NorthA, CentralA)
- A master data intance can be replicated to a cloud service as a matter of backup.
- Number of distributed instances can be increased seamlesly to any country/region in Africa ensuring African ownership
- Utlizing open source distrubted data technologies, big data platforms and AI driven models
- be able to search content published online with drop down selection and/or free chouce of keywords
- specific to African scholarly content
- ...<please add more from end user prespective i.e. how the end user would like to see the applications interface running as well as other features like speed of search, retirved content dsiplay...etc.>
| Preprint Repository | Hub and Search Portal |
|---|---|
| AfricArxiv/preprint-repository | as described here |
| upload of preprint manuscripts, student reports, research proposals, registered reports/preregistrations, short communications, etc. | aggregating scholarly output from and about Africa via one interface inlc preprints form toher platforms, scholarly books, datasets for streamlined discoverability |
| based on PKP OPS (?) | Integrates via API of partner platforms but applying different search sechme atop of thier functions |
| DOI and CC-BY 4.0 attribution | Uses crowling and feature extraction for more accurate semantic context search |
| integration with Crossref, ORCID, … | Enbales much more search oprtions relvent to content pulish in relation to Africa as per the sources section below |
-
Direct publishing to the the hub by uploading the conent directly into the core data platfrom
-
Wikidata / Scholia - https://tools.wmflabs.org/scholia/faq
-
Open Knowledge Maps (uses BASE or PubMed) / possible to integrate the map?
- https://www.connecting-africa.net/index.htm
- https://info.africarxiv.org/african-digital-research-repositories/
- Search by region = Africa* - each country and by other specific identifiers
- Repositories
- https://v2.sherpa.ac.uk/view/repository_by_country/002.html // API requirements https://v2.sherpa.ac.uk/opendoar/api.html
- Scholarly Books
- Journals
- AJOL / https://www.ajol.info/
In the advanced smart search based , crawling and extrating key features related to Africa from sources enables semantic context search with more accurat results. This should take place in follwoing main categories of search criteria or combination of them:
-
Content by African authors / co-authors who are based out in Africa
-
Content by African authors / co-authors who are based outside Africa
-
Content by non-African authors / co-authors related to Africa
-
Content authored / co-authored by /f or African Institutions
-
Content authored / co-authored by / for one or more African region, city or country
-
Other non-location specific: date, time, topic, publisher, sponsers...etc.
- Languages commonly spoken on the continet
- Machine translations with options/overlay interface to manually improve the text online (see www.GTranslate.io) // GoogleTranspate API
- Language switch // editable by the users and admin
The system consists of three main tiers: The Core Smart Data Platform, Source Data Crowling and Ingestion, Search Intergation and Portal Applications
- The main engine to mainly store features of content by source publisher incluing the content uploaded directly to the platfrom
- Contains interuim data structires needed for the storing the semnatic features including local ID and source ID to link to the orginal source content when retreving results.
- Stores addtional metadata about the source content to optimize the search process while syncornize online with changes on the source system.
- Pre-cacluates and mainatians updated search metrics triggred by the crolwing and data ingestion process as well as caches the prevouise seaerch results for repeated searches.
- Responsible for connecting, parsining and restriving content feastures from external sources by applying the above search idintifers.
- Updates the core data platform continously with new contetn and changes on the exiting content on pre-confirgured time intervales.
- Conduct needed trnafomrations and canconclization on the exterated features and trigger updating interuim search data strtcures and metrics calculations.
- Run a librairy of APIs and Consumers for different appacroches of connectivty to external sources.
- Enbales the end users who would like to use the services to search, retrive, display and anlayze the results by providing search keywords and options.
- Connect and run the semantic context search module on the core data platfrom to retrevie most relvant content and get a list of source idintifiers
- reach out to the source systems to ensure the conent is still intact and recent then restive the content title and abstact as well as other metadata.
- Display the retrived content in different layouts and arrnagments based on customer selection including saving the search and search results for later use.
- Administeration backoffice application for AfricArxiv to be bale to mointor, configure, secure and operate the whole solution.
- embdedd Altmetric badge: https://api.altmetric.com/embeds.html
- embedd AiS badge on country level: https://africainscience.org/
- …
- …
- The solution should be devloped in incremental appraoch and capatize on contriantization technologies of rease of vertical and horniztal sclability.
- For the first development increment (Proof of Value phase), 5 servers - one at each location - is needed witht he following specifictions profile:
- Operating System: Ubuntu Linux
- CPU: Intel Xeon 8 Cores x64 3.x GHz
- Memory: 64 GB
- Disk: 2 TB SSD or All Flash
- Networking: Static IP / 10 Gb NIC and Server-grad Internt Connection for external access
- All other Data Center services
- We might adapte to lower specifictions for first phase if the above are not avilable.
- Identify at least one host institution in each African region that can host the AfricArXiv hub preferably with a data center that is decntly oprtionalixed and uses virtulaization or private cloud.
- Capciaty provject for coming 5 years should be provided down the road after completing the first phase so that sounding extimates can be made.
- The design and technical archiecture will be made for potetial scale out to other African hosting facilities with no need to rebuild or change the solution.
We can use public cloud hosted and runing within Africa for the baocve main nodes and at the same time we can use a Public Cloud hosted anywhare on the world as a master copy backup for high avilbaility. Some potential services:
- Microsoft Azure’s services // already in SA
- Amazon Web Services (AWS) // announced to open servers in Africa
- Google Cloud // announced to open servers in Africa
Contact us at [email protected]
