W3C PROV

 

Luc Moreau

Co-chair of W3C Provenance Working Group

Warning: everything in this presentation is a DRAFT.

Thanks to Paul Groth, Tim Lebo, Paolo Missier, James Cheney, and the entire W3C Provenance Working Group

Talk Outline

• The PROV family of Documents
• The PROV Conceptual Data Model
• The PROV Notation and PROV XML
• PROV Constraints
• Validating provenance

A Definition of Provenance

Layered Model

Layered Model

Layered Model

Layered Model

Layered Model

Normative/Non Normative Documents

Recommendation (Blue), Note (Orange), Non WG Material (Green)

PROV Data Model Core

PROV extended model: n-ary relations, subtypes and subrelations, further relations.

PROV Data Model Components

Relations at a Glance

Relations at a Glance (Secondary Elements)

PROV-DM Types and Relations

Component 1: Entities and Activities

Entity

Activity

Start

Component 2: Derivations

Component 3: Agents and Responsibility

Component 4: Bundles

Component 5: Alternates

Component 6: Collections

The PROV Family: The PROV Notation

PROV-N: The PROV Notation

• A notation aimed at Human Consumption

  entity(tr:WD-prov-dm-20111215, [ prov:type='pr:RecsWD' ])
  

• Used in mapping data model to RDF (and XML)
• Used in providing semantics to data model

Example in PROV-N (sample from primer)

document
    prefix prov <http://www.w3.org/ns/prov#>

    entity(ex:article,[dcterms:title = "Crime rises in cities" %% xsd:string])

    activity(ex:correct,2012-03-31T09:21:00.000+01:00,2012-04-01T15:21:00.000+01:00)

    used(ex:compose,ex:dataSet1,-)

    wasGeneratedBy(ex:chart1,ex:compile,2012-03-02T10:30:00.000Z)

    agent(ex:derek,[prov:type="prov:Person" %% xsd:string, foaf:givenName = "Derek" %% xsd:string, 
                    foaf:mbox = "<mailto:derek@example.org>" %% xsd:string])

    wasAssociatedWith(ex:illustrate,ex:derek,-)

    actedOnBehalfOf(ex:derek,ex:chartgen,ex:compose)

    wasDerivedFrom(ex:dataSet2, ex:dataSet1,[prov:type='prov:Revision'])

    specializationOf(ex:articleV2,ex:article)

    alternateOf(ex:articleV2,ex:articleV1)
endDocument

Example in PROV-XML (sample from primer)

<prov:document xmlns:prov="http://www.w3.org/ns/prov#" xmlns:ns2="http://openprovenance.org/collection#" xmlns:ns3="http://openprovenance.org/validation#" xmlns:foaf="http://xmlns.com/foaf/0.1/" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:dcterms="http://purl.org/dc/terms/" xmlns:ex="http://example/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
  <prov:entity prov:id="ex:article">
    <dcterms:title xsi:type="xsd:string">Crime rises in cities</dcterms:title>
  </prov:entity>

  <prov:activity prov:id="ex:correct">
    <prov:startTime>2012-03-31T09:21:00.000+01:00</prov:startTime>
    <prov:endTime>2012-04-01T15:21:00.000+01:00</prov:endTime>
  </prov:activity>

  <prov:used>
    <prov:activity prov:ref="ex:compose"/>
    <prov:entity prov:ref="ex:dataSet1"/>
  </prov:used>

  <prov:wasGeneratedBy>
    <prov:time>2012-03-02T10:30:00.000Z</prov:time>
    <prov:entity prov:ref="ex:chart1"/>
    <prov:activity prov:ref="ex:compile"/>
  </prov:wasGeneratedBy>

  <prov:agent prov:id="ex:derek">
    <prov:type xsi:type="xsd:string">prov:Person</prov:type>
    <foaf:givenName xsi:type="xsd:string">Derek</foaf:givenName>
    <foaf:mbox xsi:type="xsd:string"><mailto:derek@example.org></foaf:mbox>
  </prov:agent>

  <prov:wasAssociatedWith>
    <prov:activity prov:ref="ex:illustrate"/>
    <prov:agent prov:ref="ex:derek"/>
  </prov:wasAssociatedWith>

  <prov:actedOnBehalfOf>
    <prov:subordinate prov:ref="ex:derek"/>
    <prov:responsible prov:ref="ex:chartgen"/>
    <prov:activity prov:ref="ex:compose"/>
  </prov:actedOnBehalfOf>

  <prov:wasDerivedFrom>
    <prov:generatedEntity prov:ref="ex:dataSet2"/>
    <prov:usedEntity prov:ref="ex:dataSet1"/>
    <prov:type xsi:type="xsd:QName">prov:Revision</prov:type>
  </prov:wasDerivedFrom>

  <prov:specializationOf>
    <prov:specializedEntity prov:ref="ex:articleV1"/>
    <prov:generalEntity prov:ref="ex:article"/>
  </prov:specializationOf>

</prov:document>

The PROV Family: PROV Constraints

Time

• Time is critical in the context of provenance
• Time can help corroborate provenance claims
• For instance, if an entity is claimed to be obtained by transforming another, then the latter must have existed before the former. If it is not the case, then there is something wrong with such a provenance claim
• Minimal time assumptions since provenance can be used in many contexts (standalone system, web, space)

Events

e1 precedes e2: reflexive, transitive relation.

e1 strictly precedes e2: irreflexive, transitive relation.

Constraints

Constraints (2)

Constraints(3) SW Technologies

PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX spin: <http://spinrdf.org/spin#>
PREFIX prov: <http://www.w3.org/ns/prov#>
 
CONSTRUCT {
   :_b0
      a spin:ConstraintViolation;
      spin:violationRoot ?this;
      rdfs:label "Qualified start time disagrees with Activity start time.";
   .
}
WHERE {
   ?this a prov:Activity;
        prov:wasStartedAt   ?t_1;
        prov:qualifiedStart ?start; 
   .
   ?start 
      a prov:Start;
      prov:atTime ?t_2;
   .
   FILTER(?t_1 != ?t_2)
}

:paulRule prov:wasAttributedTo <http://data.semanticweb.org/person/paul-groth> 

Constraints: Challenge for Participants

Question:

• What is the best SW technology to implement PROV constraints?

Participate:

• The PROV Working Group will issue a call for implementation circa December 6.
• Submit an implementation report
• Tell us which constraints you can validate
• Help us build a corpus of examples

Validator Demo

Primer in turtle

Derivation from e2 to e1

Derivation from e2 to e1 and from e1 to e2 (xml) (ttl).

Conclusion: Specifications

Conclusion: Participate

• Markup web pages with provenance and/or generate provenance
• Write applications/programs that generate/consume provenance
• Implement validation constraints
• Contribute examples of provenance
• Submit an implementation report for W3C PROV Working Group
• Feedback to public-prov-comments@w3.org

Further Material

Example

Example

Example (reports)

Example (activities)

Example (agent)

Example (plan)

Example (requests)

Example (usage)

Example (generation)

Example (derivation)

Example (association)