SIB Next-steps after demonstration

[balessan]

URI example: https://api.community.startinblox.com/

• [ ] #38

  • [ ] Get numbers
  • [ ] Advancing on index creation/update/maintenance architecture
  • [ ] Suggestion: Usage of the activities for that

• [ ] Validate the efficiency of criteria intersection with all the users directory search fields

  • [ ] Name / Firstname - Job - City - Skills ?
  • [ ] What to do about cities ? Nothing for now probably.
  • [ ] So adapt the example on client side, but also add examples indexing for name/firstname and jobs
  • [ ] For jobs it is equivalent to cities

• [ ] Rationalize search strategies in the filterMixin (client/server/index)

  • [ ] Add the support for the rdf-type-fields and predicate-fields attribute on solid-form-search
  • Is it enough for proper configuration of the search request ?
  • [ ] Do we want to use a combination of server and index -> we want to fusion
  • [ ] Dynamic index generation as a first version

• [ ] Test at larger scale (*10 on current volume)

• [ ] Pursuing the Comunica integration in the client framework

• Missing features on indexes:

  • [ ] Pagination ?
  • [ ] Ordering ?

• [ ] Do we pass the index of the users containers as data-src: see suggestions below:

Client-side example:

<solid-form-search
   fields="skills"
   rdf-type-skills="foaf:user"
   predicate-skills="hasSkill"

/></solid-form-search>
<solid-display
   data-src="https://api.community.startinblox.com/users/"
   >
</solid-display>

Server-side example:

{
  "@context": "https://cdn.startinblox.com/owl/context.jsonld",
  "@type": [
       "dcat:Dataset",
       "ldp:Container",
       "idx:Entrypoint"
  ],
  "@id": "https://api.community.startinblox.com/users/",
  "idx:indexes": {
      "@id": "https://api.community.startinblox.com/users/indexes/"
  },
  "dcat:Resource": [
    {
      "@id": "https://api.community.startinblox.com/users/antoine/"
    },
    {
      "@id": "https://api.community.startinblox.com/users/balessan/"
    },
    {
      "@id": "https://api.community.startinblox.com/users/sylvain/"
    }
  ],
  "ldp:contains": [
    {
      "@id": "https://api.community.startinblox.com/users/antoine/"
    },
    {
      "@id": "https://api.community.startinblox.com/users/balessan/"
    },
    {
      "@id": "https://api.community.startinblox.com/users/sylvain/"
    }
  ]
}

Description of an index:

{
  "@context": "https://cdn.startinblox.com/owl/context.jsonld",
  "@type": [
       "dcat:Dataset",
       "ldp:Container",
       "idx:Entrypoint"
  ],
  "@id": "https://api.community.startinblox.com/users/",
  "idx:indexes": {
      "@id": "https://api.community.startinblox.com/users/indexes/"
  },
https://api.community.startinblox.com/users/indexes/

[lecoqlibre]

Technical deliveries:

• [ ] Write a RDF document library.
  • [ ] Make a module for our indexes.
  • [ ] Make a module for TypeIndex.
  • [ ] Make a module for a document store with abstraction (in memory, local storage, session-storage, indexed db).
  • [ ] Make a module for Solid CRUD operations.
  • [ ] Make a module to generate BGP SPARQL queries.
• [ ] Integrate the RDF document library into the framework.
  • [ ] Refactor index generation source code.
  • [ ] Make an interface to generate indexes.
• [ ] Write an indexer agent using the RDF document library.
• [ ] Write a BGP query planner.
• [ ] Fix Comunica named graph querying.

Into the SiB framework:

• [ ] Generate SPARQL BGP queries from the components.
• [ ] Integrate the BGP query planner.
• [ ] Make an abstraction layer of Comunica.
• [ ] Integrate Comunica.
• [ ] Execute queries thought the query planner and send the results to the components.

About the string indexes: The depth of the index branches must be fixed. Each branch must have the same depth regardless of the data.