29.05.2016 | 1. WHiSe Workshop / ESWC 2016, Anissaras - Crete

Data Repositories in the Humanities and the Semantic Web

Modelling, Linking, Visualising

Max Grüntgens, Torsten Schrade | @digicademy | digicademy

Download: https://github.com/digicademy/whise2016-xtriples |  View: http://digicademy.github.io/whise2016-xtriples

Table of Contents

1. Humanities' data repositories: Examples from the
   Academy of Sciences and Literature Mainz
2. Implicit and explicit semantics:
   bridging the gap between TEI-XML and RDF
3. The XTriples Webservice
4. XTriples use cases: harvesting, transforming and vizualising XML data
5. XTriples: Outlook & Roadmap

01

Humanities' data repositories

Examples from the Academy of Sciences and Literature Mainz

The German Academies of Sciences

Long-term research in the humanities and social sciences

Website of the Union of the German Academies of Sciences

• The "Academies' Programme" is currently the most comprehensive humanities research programme in Germany.
• It provides funding for long-term research projects, predominantly in the humanities but also in the social sciences.
• Within the Academies’ Programme, a total of over 900 researchers and administrative staff are currently working on 144 different projects across 199 locations. Projects can range from 12 to 24 years.
• The main focus of the programme is on compiling dictionaries, encyclopaedias and critical editions in the fields of theology, philosophy, history, literature and linguistics, art history and archaeology, inscriptions, onomastics, and musicology.
• Methods from the Digital Humanities increasingly become common practice in the research conducted by the academies. Projects accumulate a broad range of humanities research data that is often made available on specialised online platforms.

Example Data Repositories

Academy of Sciences and Literature | Mainz

Regesta Imperii Online (http://www.regesta-imperii.de)

• Database with ca. 173.000 Regestae of medieval charters from Charlemagne to Maximilian I covering 800 years of European medieval history
• Records available under CC-BY 4.0 license via REST-Interface
• TEI-XML according to the guidelines of the Charters Encoding Initiative (example)

Corpus Vitrearum Medii Aevi (http://www.corpusvitrearum.de)

• Digital archive with high-resolution TIFF images of medieval stained glass (CC-BY 4.0)
• Extensive XMP metadata embedded into the images
• Controlled vocabulary (Iconclass) to describe the iconography
• REST-Interface for harvesting images and metadata in XMP and JSON-LD (example)

German Inscriptions Online (http://www.inschriften.net)

• 17.000 catalogue numbers from the critial edition "Deutsche Inschriften" with 18.000 images, covering medieval and early modern inscriptions in the German speaking area
• Currently in the process of making the records available marked up in TEI-XML according to the EpiDoc standard
• REST-Interface for harvesting the XML data (example)

LOD for the Academies' projects

Great potential in connecting research data repositories

New connections, new questions

Questions that could be analysed using semantic web technologies

• Which historical persons occur across all repositories in which particular contexts?
• Which iconographically similar resources exist in the data sets and during what particular time frame?
• Are there geographical patterns in the occurrence of certain religous symbols (e.g. on gravestones etc.) that can be inferenced from the data?
• Can historical phenomena like floods, famine etc. be pinpointed by cross-analysing the data sets?

But the primary question is: How to get the semantic information out of TEI-XML?

02

Implicit and explicit semantics

Bridging the gap between TEI-XML and RDF

Implicit Semantics in TEI-XML

Metadata of a letter from Johann Wolfgang von Goethe

Link: Johann Wolfgang von Goethe to Samuel Thomas Soemmering (1793)

    <correspDesc key="686" cs:source="#SOE20">
        <correspAction type="sent">
            <persName ref="http://d-nb.info/gnd/118540238">
                Johann Wolfgang von Goethe
            </persName>
            <placeName ref="http://www.geonames.org/2812482">
                Weimar
            </placeName>
            <date when="1793-12-05">5.12.1793</date>
        </correspAction>
        [...]
    </correspDesc>

• Subject
• Predicate
• Object

The semantics are "implicit" in the sense that they can be contained in XML tag names, tag values, attribute names as well as values.

Explicit Semantics in RDF

The same information as expressed in RDF

    Goethe           is a              Person ;
                     sends             Letter .

    Letter           dates to          1793 ;
                     sent from         Weimar .

    Weimar           is a              City ;
                     has latitude      11.32 ;
                     has longitude     50.98 .

• Subject
• Predicate
• Object

Gathering additional data

Enriching the RDF model of Goethe's letter with LOD

1. 1
2. 2
3. 3
4. 4
5. 5
6. 6
7. 7

• correspSearch
• GND
• Geonames

How?

03

The XTriples Webservice

A generic webservice to extract RDF statements from XML

Basic principles of the webservice

1. Generic: Extract RDF statements out of any XML
2. Simple: Slim configuration based on statement patterns
3. Inclusive: Harvest data from other XML resources during extraction
4. Flexible: Return a broad array of RDF formats
5. RESTful: Communicate solely over HTTP

XTriples

Process diagramm & technical components

Simple XML configuration

Based on statement patterns with XPATH / XQuery expressions

<xtriples>
    <configuration>
        <vocabularies>
            <vocabulary prefix="###PREFIX###" uri="###NAMESPACE###"/>
            <vocabulary prefix="rdf" uri="http://www.w3.org/1999/02/22-rdf-syntax-ns#"/>
            [...] <!-- more vocabularies -->
        </vocabularies>
        <triples>
            <statement>
                <subject prefix="###PREFIX###">###XPATH###</subject>
                <predicate prefix="rdf">about</predicate>
                <object type="literal">###XPATH###</object>
            </statement>
            [...] <!-- more statements -->
        </triples>
    </configuration>
    <collection uri="http://xml.collection.somewhere/resources/index.xml">
        <resources uri="http://xml.collection.somewhere/resources/{###XPATH###}.xml"/>
    </collection>
    [...] <!-- more collections -->
</xtriples>

Example data and configuration

Crawling a collection and extracting FOAF statements

<gods>
    <god id="1">
        <name>
            <greek>Ἀφροδίτη</greek>
            <roman>Venus</roman>
            <english>Aphrodite</english>
        </name>
        <gender>female</gender>
        [...]
    </god>
    <god id="2">
        <name>
            <greek>Ἀπόλλων</greek>
            <roman>Apollon</roman>
            <english>Apollo</english>
        </name>
        <gender>male</gender>
        [...]
    </god>
    <god id="3">
        <name>
            <greek>Ἄρης</greek>
            <roman>Mars</roman>
            <english>Ares</english>
        </name>
        <gender>male</gender>
        [...]
    </god>
    [...]
</gods>

<xtriples>
    <configuration>
        <vocabularies>
            <vocabulary prefix="gods" uri="http://xtriples.spatialhumanities.de/examples/gods/"/>
            <vocabulary prefix="rdf" uri="http://www.w3.org/1999/02/22-rdf-syntax-ns#"/>
            <vocabulary prefix="rdfs" uri="http://www.w3.org/2000/01/rdf-schema#"/>
            <vocabulary prefix="foaf" uri="http://xmlns.com/foaf/0.1/"/>
        </vocabularies>
        <triples>
            <statement>
                <subject prefix="gods">/@id</subject>
                <predicate prefix="rdf">type</predicate>
                <object prefix="foaf" type="uri">Person</object>
            </statement>
            <statement>
                <subject prefix="gods">/@id</subject>
                <predicate prefix="rdfs">label</predicate>
                <object type="literal" lang="en">/name/english</object>
            </statement>
            <statement>
                <subject prefix="gods">/@id</subject>
                <predicate prefix="rdfs">label</predicate>
                <object type="literal" lang="gr">/name/greek</object>
            </statement>
        </triples>
    </configuration>
    <collection uri="http://xtriples.spatialhumanities.de/examples/gods/all.xml">
        <resource uri="{//god}"/>
    </collection>
</xtriples>

04

XTriples use cases

Harvesting, transforming and vizualising XML data

Use case 1: Corpus Vitrearum

Connecting the CVMA image archive with Iconclass and Europeana

CVMA - Stage 01

Applying CIDOC, SKOS and dcterms

XTriples: Configuration | Run this live: SVG | RDF/XML | Turtle

CVMA - Stage 02

Connecting Iconclass

XTriples: Configuration | Run this live: SVG | RDF/XML | Turtle

CVMA - Stage 03

Iconographically similar resources from Europeana

XTriples: Configuration | Run this live: SVG | RDF/XML | Turtle

XTriples: Configuration | Run this live: SVG | RDF/XML | Turtle

Use case 2: A network of letters

Combining correspSearch with the GND and Geonames

correspSearch - Stage 01

Goethe and his correspondents in correspSearch

XTriples: Configuration | Run this live: SVG | RDF/XML | Turtle

correspSearch - Stage 02

Extracting / enriching Goethe's letter network in correspSearch

XTriples: Configuration | Run this live: SVG | RDF/XML | Turtle

05

XTriples

Outlook & Roadmap

Features and improvements

Outlook & Roadmap

• Performance: Improve performance with caching routines, XQuery optimisations and tutorials on how to write efficient configurations (and also increase server resources...)
• Context: Introduce a <context> tag within <statement> that will allow setting/generating a context for each extracted triple
• Triple stores: Implement mechanisms that will allow POSTing the result of an XTriples extraction directly to a triple store (like Sesame etc.)
• "Joins": Introduce a "join" mechanism that will allow "joining" whole parts of external LOD resources (i.e. RDF-XML tree fragments) with extracted statements
• Validation: Implement a validation service for XTriples configuration files
• And of course any other useful feature you can think of...

Links, Software & Attribution

Links

• XTriples: http://xtriples.spatialhumanities.de/
• Apache any23: https://any23.apache.org/
• Redefer rdf2svg: http://rhizomik.net/html/redefer/rdf2svg-form/
• eXist XML database: http://exist-db.org/

Software used

• Impress.js (Presentation)
• Open Sans (Google Font)
• Magnific Popup (Lightbox)
• greuler (Graph Modelling)
• hightlight.js (Syntax Highlighting)

Attribution

• Link to this presentation: http://digicademy.github.io/whise2016-xtriples/
• License: CC-BY 4.0, Max Grüntgens and Torsten Schrade