Pattern Ontology (PO) Sebastian Barzaghi (https://orcid.org/0000-0002-0799-1527) Angelo Di Iorio (https://orcid.org/0000-0002-6893-7452) Fabio Vitali (https://orcid.org/0000-0002-7562-5203) Silvio Peroni (https://orcid.org/0000-0003-0530-4305) 2026-09-21 The **_Pattern Ontology (PO)_** is an OWL 2 DL ontology that formally defines patterns for segmenting a document into atomic components, in order to be manipulated independently and re-flowed in different contexts.  Structural patterns allow representing document elements (such as sections, paragraphs, lists, metadata, and tables) through abstract, reusable primitives independent of specific markup syntaxes. ## Examples of use In the following subsections, we introduce some examples to showcase how to use PO. The prefixes that are used in all the examples provided below are defined as follows: @prefix : <http://sparontologies.net/example/> . @prefix po: <http://www.essepuntato.it/2008/12/pattern#> . @prefix xsd: <http://www.w3.org/2001/XMLSchema#> . ### Structural decomposition of a document section With PO, it is possible to represent document elements and how they manifest as patterns occurring within the document's stucture itself. For example, the following example shows how to use PO to model a section defined as a headed container and containing a heading block as its header and a body paragraph block containing inline text and atom elements. :section-1 a po:HeadedContainer ; po:hasName "section" ; po:containsAsHeader :heading-1 ; po:contains :paragraph-1 . :heading-1 a po:Block ; po:hasName "h1" ; po:contains :text-heading . :paragraph-1 a po:Block ; po:hasName "p" ; po:contains :text-span . :text-span a po:Inline ; po:hasName "span" ; po:contains :atom-text . :text-heading a po:Atom ; po:hasName "text" . :atom-text a po:Atom ; po:hasName "text" . ## Competency Questions PO can be used for answering several questions related to the atomic components that form a document accoridng to certain patterns. In the following subsections, some of them are introduced together with their respective SPARQL queries. The prefixes that are used in all the SPARQL queries provided below are defined as follows: PREFIX po: <http://www.essepuntato.it/2008/12/pattern#> ### CQ1 Which header elements are contained within a headed container? SELECT ?header WHERE { ?section a po:headedContainer ; po:containsAsHeader ?header . } ### CQ2 Which block elements are directly contained within a container? SELECT ?block WHERE { ?section po:contains ?block . ?block a po:Block . } The Pattern Ontology (PO) Di Iorio, A., Peroni, S., Poggi, F., Vitali, F. (2012). Afirst approach to the automatic recognition ofstructural patterns in XML documents. In Proceedingsof the 2012 ACM symposium on Document Engineering(DocEng 2012): 85-94. DOI: 10.1145/2361354.2361374 2013-07-16 2013-07-16 2026-09-21 TBA po http://purl.org/spar/po 1.4.3 documents document engineering scholarly publishing semantic publishing * **v1.4.3**: Added new documentation, including annotation properties (e.g., `bibo:doi`, `schema:codeRepository`, etc.), term properties (e.g., `vann:example`, `vs:term_status`, etc.), exemplary competency questions with their respective SPARQL queries, and a Grafoo diagram. contains as header The relation between a structured element, containing another generic element as part of its header, and that other element. :my-section po:containsAsHeader :my-header-block . atom element Literal text that is part of the document body. Similarly to `po:Inline`, elements following the 'atom' pattern can only be contained within block or inline elements (and consequently they also cannot be used as root elements of documents). It can contain textual content and no other elements. :my-atom a po:Atom ; po:isContainedBy :my-paragraph . non-structured element An element that cannot contain other elements. :my-leaf-node a po:NonStructured . container element An element concerning the structural organization of a document. Elements following this pattern contain no textual content and contain only elements compliant with the patterns: `po:Meta`, `po:Field`, `po:Block` and any subtype of `po:Container`. It is disjointed with the pattern `po:Popup`, although they share the same content model. :my-body a po:Container ; po:contains :section-1, :section-2 . pattern-based element An atomic component of a document that is able to capture a recurrent and well knwon structure of a document. Any entity belonging to this class is compliant with exactly one structural pattern. :my-element a po:Pattern . can contain homonymous elements An indication whether a hierarchical-based entity contains al least two entities that share the same name or not. :my-list po:canContainHomonymousElements true . false structured element An element that can contain other elements. :my-section a po:Structured ; po:contains :my-paragraph . headed-container element A container element whose content model needs to begin with one or more block elements (the heading). :my-section a po:HeadedContainer ; po:containsAsHeader :my-title ; po:contains :my-para . contains The relation between a structured element containing another generic element, and that other element. :my-container po:contains :my-block . block element An element that organises the document content as a sequence of nestable elements and text nodes. In particular, elements following this pattern can contain text and other elements compliant with the patterns `po:Inline`, `po:Atom`, `po:Milestones` and `po:Popup`. Furthermore, it is a requirement that block elements are contained only by container and popup elements. :my-paragraph a po:Block ; po:isContainedBy :my-section ; po:contains :my-span . is contained by An element that is contained by another structured element. :my-block po:isContainedBy :my-container . true non-textual element An element that cannot have textual content in its content model. :my-element-container a po:NonTextual . has name The name of a particular element. :my-paragraph po:hasName "p" . field element Literal metadata or text that is not really part of the document body. Its main difference with `po:Meta` is that `po:Field` can contain textual content. :my-author-name a po:Field ; po:isContainedBy :my-author-header . inline element An element that organises the document content as a sequence of nestable elements and text nodes, and that can contain other elements compliant with the same pattern. This class has the same use and content model of the pattern Block, but differing primarily because a) inline elements can contain other elements compliant with the same pattern while block elements cannot, and b) inline elements must always be contained by other block or inline elements and by no other type of element. These constraints also imply that inline elements cannot be used as root elements of documents and that the class Block is disjoint with the class Inline (i.e., a markup element cannot be a block and an inline at the same time). :my-emphasis a po:Inline ; po:isContainedBy :my-paragraph ; po:contains :my-atom-text . meta element An element that contains neither other elements nor textual content. Contrarily to the pattern `po:Milestone`, which was meant to describe markup elements that impact the document because of their location, the main feature of its disjoint sibling is the mere existence, independently from the position it has within the document. `po:Meta` elements convey metadata information about the document or part of it, independently of where they are. Thus, meta elements can be contained only within container elements. :my-doc-meta a po:Meta ; po:isContainedBy :my-doc-header . mixed element An element that can contain other elements and text nodes. :my-paragraph a po:Mixed ; po:contains :link . false textual element An element that can have textual content in their content models. :paragraph-text a po:Textual . milestone element An elements that contains neither other elements nor textual content. :my-page-break a po:Milestone ; po:isContainedBy :my-paragraph . bucket element An element that can contain other elements but no text nodes. :my-bucket a po:Bucket ; po:contains :item-1, :item-2 . Similarly to `po:Inline`, elements following the `po:Milestone` pattern can only be contained within block or inline elements (and consequently they also cannot be used as root elements of documents). The distinctive characteristic of the pattern `po:Milestone` is the location it assumes within the document. table element A container element that must contain only homogeneous elements (but they can be repeated). :my-list a po:Table ; po:canContainHeteronymousElements false ; po:canContainHomonymousElements true . flat element An element that can contain text nodes but no elements. :my-flat a po:Flat ; po:hasName "text" . popup element A bucket element that is only present within block and inline elements. Popup elements are used whenever complex structures need to be placed within content elements such as paragraphs. :my-footnote a po:Popup ; po:isContainedBy :my-paragraph ; po:contains :my-footnote-block . true can contain heteronymous elements An indication whether a hierarchical-based entity contains al least two entities having different names or not. :my-record po:canContainHeteronymousElements true . marker element An element that can contain neither text nodes nor elements. :my-break-tag a po:Marker ; po:hasName "br" . record element A container element that can only contain heterogeneous and non repeatable elements. :my-author-record a po:Record ; po:canContainHeteronymousElements true ; po:canContainHomonymousElements false . true false true is contained by as header The relation between an element, contained by another structured element as part of its header, and that other element. :my-header-block po:isContainedByAsHeader :my-section . false