Skip to content

Latest commit

 

History

History
210 lines (171 loc) · 5.66 KB

DTS-Presentation.md

File metadata and controls

210 lines (171 loc) · 5.66 KB
<style> pre { font-size: 14px!important; margin:-5PX; padding:5px; } </style>

Distributed Text Services

A presentation for the Pelagios Text Reuse Working Group

Plan

  • (Really) Quick history
  • Objectives
  • Team
  • Principles
  • Current status
  • Implementations

(Really) Quick history

  • Started in a workshop about Canonical Text Services (CTS) future in 2015
  • Alternative to CTS to fit more needs, more technical realities.
  • Ups and downs, real speed up over the last academic year

Objectives

  • Open group for contributions, not a one-person project
  • Scalable API : does well for 10 texts as much as 1M
  • Linked Open Data
    • no access barrier
    • controlled vocabularies
  • Identifier agnostic
  • CTS retro compatible
  • Focused on serving and retrieval for the first lot
    • Search might come later ?
  • URI scheme should be flexible (not everybody can have reroute)
  • No catalogue hierarchy should be forced upon implementers
    • Freedom of catalogue architecture (details, depth, etc.)

Team

  • Bridget Almas
  • Hugh Cayless
  • Thibault Clérice
  • Jeffrey C. Witt
  • Vincent Jolivet
  • Pietro Luzzo
  • Emmanuelle Morlock
  • Jonathan Robie
  • Matteo Romanello
  • James Tauber
  • and the list goes on and I probably forgot people here.

Core of 4 to 5 peoples meeting every two weeks for the past year. Latinists and hellenists but other periods and languages as well

Principles

Three main functions turned into endpoint :

  • Collection browsing (Collection Endpoint)
    • JSON-LD, Hydra-based answers, DC as core metadata vocabulary
    • Restful
  • Navigation Endpoint
    • JSON (endpoint to be in discussion soon)
  • Document Endpoint
    • XML-TEI

Principles (2)

  • Endpoints do not need to be all available.
  • Endpoints should be usable without the need to browse from the top down
    • I know you have this text, please just give me its available pages.
    • I know your Victor Hugo's Misérables has paragraph 5 from chapter 3. Give it to me.
  • Freedom of other output type for the document endpoints (you can allow people to get plain text). But you have to serve TEI XML.

Collection Endpoint

{
    "@context": {
        "@vocab": "https://www.w3.org/ns/hydra/core#",
        "dc": "http://purl.org/dc/terms/",
        "dts": "https://w3id.org/dts/api#"
    },
    "@id": "general",
    "@type": "Collection",
    "totalItems": "2",
    "title": "Collection Générale de l'École Nationale des Chartes",
    "dts:dublincore": {
        "dc:publisher": ["École Nationale des Chartes", "https://viaf.org/viaf/167874585"],
        "dc:title": [
            {"fr" : "Collection Générale de l'École Nationale des Chartes"}
        ]
    },
    "member": [
        {
             "@id" : "cartulaires",
             "title" : "Cartulaires",
             "description": "Collection de cartulaires d'Île-de-France et de ses environs",
             "@type" : "Collection",
             "totalItems" : "10"
        },
        {
             "@id" : "lasciva_roma",
             "title" : "Lasciva Roma",
             "description": "Collection of primary sources of interest in [...]",
             "@type" : "Collection",
             "totalItems" : "1"
        },
        {
             "@id" : "lettres_de_poilus",
             "title" : "Correspondance des poilus",
             "description": "Collection de lettres de poilus entre 1917 et 1918",
             "@type" : "Collection",
             "totalItems" : "10000"
        }
    ]
}

Navigation endpoint

Warning: definitely a draft, not discussed in depth

{
    "@context": {
        "passage": "https://w3id.org/dts/api#/#passage"
    },
    "@base": "/dts/api/document/",
    "@id":"urn:cts:greekLit:tlg0012.tlg001.opp-grc5",
    "passage": ["1", "2", "3"]
}

Document Endpoint

<?xml version="1.0" encoding="UTF-8"?>
<TEI xmlns="http://www.tei-c.org/ns/1.0">
   <teiHeader>
      <fileDesc>
         <titleStmt>
            <title>bgu.11.2029</title>
         </titleStmt>
         <publicationStmt>
            <authority>Duke Collaboratory for Classics Computing (DC3)</authority>
            <idno type="filename">bgu.11.2029</idno>
            <idno type="ddb-perseus-style">0001;11;2029</idno>
            <idno type="ddb-hybrid">bgu;11;2029</idno>
            <idno type="HGV">9566</idno>
            <idno type="TM">9566</idno>
            <availability>
               <p>© Duke Databank of Documentary Papyri. This work is licensed under a 
               <ref type="license" target="http://creativecommons.org/licenses/by/3.0/">Creative 
               Commons Attribution 3.0 License</ref>.</p>
            </availability>
         </publicationStmt>
      </fileDesc>
   </teiHeader>
   <dts:fragment xmlns:dts="https://w3id.org/dts/api#">
    <lb n="1"/><expan>τετελ<ex>ώνηται</ex></expan> 
    <expan>δι<ex>ὰ</ex></expan> <expan>πύλ<ex>ης</ex></expan> Διονυσιάδος 
   </dts:fragment>
</TEI>

Current status

  • Collection endpoint (= catalogue browsing) is close to a good first draft. Some discussion remaining, mostly due to first implementation giving us feedback on this.
  • Navigation endpoint to be the next one to be looked at, should be shorted than collection (hopefully)
  • Document API already discussed a lot, not finalized

Implementations

  • SCTA by Jeffrey C Witt
  • Beta maṣāḥǝft by Pietro Luzzo
  • Capitains
    • web client, (Python)
    • web server, (Python)
    • local client, (Python, hopefully JS ?)
    • guidelines to encode for the server part (TEI + proprietary metadata scheme)

Thanks

If I forgot something, we have more than 3 members of DTS here, maybe they'd like to add something ?