ParlaMint-TEI.ana.rng

<?xml version="1.0" encoding="UTF-8"?>
<grammar ns="http://www.tei-c.org/ns/1.0"
         xmlns:xml="http://www.w3.org/XML/1998/namespace"
         xmlns="http://relaxng.org/ns/structure/1.0"
         xmlns:a="http://relaxng.org/ns/compatibility/annotations/1.0"
         datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes">

  <a:documentation>RelaxNG XML schema for ParlaMint corpora: schema
  for linguistically annotatated corpus component file.</a:documentation>
  
  <a:documentation>The linguistic annotation markup covers sentences, tokens, lemmas,
  morphosyntactic tags, UD morphological features, UD dependency relations, and named
  entities.</a:documentation>

  <include href="ParlaMint-TEI.rng">
    
    <define name="encodingDesc">
      <a:documentation>Encoding description linguistically annotated corpus compoment.</a:documentation>
      <element name="encodingDesc">
        <ref name="global.atts"/>
        <element name="projectDesc">
          <ref name="paras"/>
        </element>
        <ref name="tagsDecl"/>
	<a:documentation>Definition of listPrefix for MTed corpora.</a:documentation>
	<optional>
          <element name="listPrefixDef">
            <ref name="prefixDef"/>
          </element>
	</optional>
      </element>
    </define>
    
    <define name="seg">
      <a:documentation>The only element that contains analysed text is the segment. It can,
      however, contain transcription commentary.</a:documentation>
      <element name="seg">
        <ref name="global.atts"/>
        <!-- Might have a linguistic analysis -->
	<optional>
          <ref name="ana.att"/>
        </optional>
        <oneOrMore>
          <choice>
            <ref name="comment"/>
            <ref name="pb"/>
            <ref name="sentence"/>
          </choice>
        </oneOrMore>
      </element>
    </define>
    
  </include>
  
  <define name="sentence">
    <a:documentation>A sentence.</a:documentation>
    <element name="s">
      <ref name="global.atts"/>
      <!-- Might have a linguistic analysis -->
      <optional>
        <ref name="ana.att"/>
      </optional>
      <oneOrMore>
        <choice>
          <ref name="comment"/>
          <ref name="pb"/>
          <ref name="ner"/>
          <ref name="phr"/>
          <ref name="word"/>
          <ref name="punct"/>
        </choice>
      </oneOrMore>
      <optional>
	<ref name="syntax"/>
      </optional>
    </element>
  </define>

  <define name="phr">
    <a:documentation>Multi-Word Unit</a:documentation>
    <element name="phr">
      <ref name="global.atts"/>
      <attribute name="function">
        <data type="token"/>
      </attribute>
      <attribute name="ana">
        <data type="anyURI"/>
      </attribute>
      <a:documentation>Currently we only support phrases with semantic markers</a:documentation>
      <attribute name="type">
        <value>sem</value>
      </attribute>
      <ref name="mwe_tokens"/>
    </element>
  </define>
  
  <define name="ner">
    <a:documentation>Named entity element</a:documentation>
    <choice>
      <a:documentation>Standard name element</a:documentation>
      <element name="name">
        <ref name="global.atts"/>
        <ref name="ref.att"/>
        <optional>
          <attribute name="ana">
            <data type="anyURI"/>
          </attribute>
        </optional>
        <optional>
          <attribute name="type">
            <choice>
              <value>PER</value>
              <value>LOC</value>
              <value>ORG</value>
              <value>MISC</value>
            </choice>
          </attribute>
        </optional>
        <optional>
          <attribute name="subtype">
            <data type="token"/>
          </attribute>
        </optional>
        <ref name="ner_tokens"/>
      </element>
      <a:documentation>Additional content used by CZ:</a:documentation>
      <element name="date">
        <ref name="ner.atts"/>
        <optional>
          <ref name="temporal.atts"/>
        </optional>
        <ref name="ner_tokens"/>
      </element>
      <element name="time">
        <ref name="ner.atts"/>
        <ref name="ner_tokens"/>
      </element>
      <element name="num">
        <ref name="ner.atts"/>
        <ref name="ner_tokens"/>
      </element>
      <element name="unit">
        <ref name="ner.atts"/>
        <ref name="ner_tokens"/>
      </element>
      <element name="email">
        <ref name="ner.atts"/>
        <ref name="ner_tokens"/>
      </element>
      <element name="ref">
        <ref name="ner.atts"/>
        <ref name="ner_tokens"/>
      </element>
    </choice>
  </define>

  <define name="ner.atts">
    <a:documentation>Attributes for NEs.</a:documentation>
    <ref name="global.atts"/>
    <attribute name="ana">
      <data type="anyURI"/>
    </attribute>
  </define>
  
  <define name="mwe_tokens">
    <a:documentation>Annotations that can appear in a MWE.</a:documentation>
    <oneOrMore>
      <choice>
        <ref name="word"/>
        <ref name="punct"/>
      </choice>
    </oneOrMore>
  </define>

  <define name="ner_tokens">
    <a:documentation>Annotations that can appear in a NE.</a:documentation>
    <oneOrMore>
      <choice>
        <ref name="word"/>
        <ref name="punct"/>
        <ref name="comment"/>
        <ref name="pb"/>
        <ref name="ner"/>
      </choice>
    </oneOrMore>
  </define>

  <define name="tokens">
    <a:documentation>A series of tokens.</a:documentation>
    <oneOrMore>
      <choice>
        <ref name="word"/>
        <ref name="punct"/>
      </choice>
    </oneOrMore>
  </define>

  <define name="word">
    <a:documentation>Definition of word.</a:documentation>
    <element name="w">
      <a:documentation>cf. https://clarin-eric.github.io/parla-clarin/#sec-ananorm</a:documentation>
      <choice>
        <a:documentation>Simple word:</a:documentation>
        <group>
          <ref name="word.atts"/>
	  <ref name="normalized-space.string"/>
        </group>
        <a:documentation>Word with several normalised words:</a:documentation>
        <group>
          <ref name="word-optional.atts"/>
          <!-- This doesn't work: error: group of "string" or "data" element -->
          <!--ref name="normalized-space.string"/-->
          <text/>
          <oneOrMore>
            <element name="w">
              <ref name="word.atts"/>
              <empty/>
            </element>
          </oneOrMore>
        </group>
        <a:documentation>Normalised word with several surface words:</a:documentation>
        <group>
          <ref name="word.atts"/>
          <oneOrMore>
            <element name="w">
	      <ref name="normalized-space.string"/>
            </element>
          </oneOrMore>
          <empty/>
        </group>
      </choice>
    </element>
  </define>

  <define name="punct">
    <a:documentation>A punctuation element.</a:documentation>
    <element name="pc">
      <ref name="token.atts"/>
      <ref name="non-empty.string"/>
    </element>
  </define>

  <define name="syntax">
    <a:documentation>Dependency syntax as a link group (inside a sentence).</a:documentation>
    <element name="linkGrp">
      <attribute name="targFunc">
        <value>head argument</value>
      </attribute>
      <attribute name="type">
        <value>UD-SYN</value>
      </attribute>
      <oneOrMore>
        <element name="link">
          <attribute name="ana">
            <data type="anyURI"/>
          </attribute>
          <attribute name="target">
            <ref name="anyURIs"/>
          </attribute>
          <empty/>
        </element>
      </oneOrMore>
    </element>
  </define>

  <define name="token.atts">
    <a:documentation>Attributes of token elements.</a:documentation>
    <optional>
      <ref name="id.att"/>
    </optional>
    <ref name="lang.att"/>
    <optional>
      <attribute name="join">
        <value>right</value>
      </attribute>
    </optional>
    <optional>
      <attribute name="norm">
	<ref name="non-empty.string"/>
      </attribute>
    </optional>
    <attribute name="msd">
      <ref name="non-empty.string"/>
    </attribute>
    <optional>
      <attribute name="ana">
        <ref name="anyURIs"/>
      </attribute>
    </optional>
    <optional>
      <attribute name="pos">
	<ref name="non-empty.string"/>
      </attribute>
    </optional>
    <optional>
      <attribute name="function">
	<!-- This is different from TEI, where it is a name token (which can't contain comma!?) -->
	<!--ref name="normalized-space.string"/-->
        <data type="token"/>
      </attribute>
    </optional>
  </define>

  <define name="word.atts">
    <a:documentation>Obligatory attributes of word.</a:documentation>
    <ref name="token.atts"/>
    <attribute name="lemma">
      <ref name="normalized-space.string"/>
    </attribute>
  </define>

  <define name="word-optional.atts">
    <a:documentation>Optional attributes of word.</a:documentation>
    <optional>
      <ref name="id.att"/>
    </optional>
    <ref name="lang.att"/>
    <optional>
      <attribute name="join">
        <value>right</value>
      </attribute>
    </optional>
    <optional>
      <a:documentation>UD PoS and morph.features, e.g. "UPosTag=NUM|NumForm=Digit":</a:documentation>
      <attribute name="msd">
	<ref name="non-empty.string"/>
      </attribute>
    </optional>
    <optional>
      <attribute name="ana">
        <ref name="anyURIs"/>
      </attribute>
    </optional>
    <optional>
      <attribute name="lemma">
	<ref name="normalized-space.string"/>
      </attribute>
    </optional>
  </define>

</grammar>