Modify schema to allow text structure under figDesc, so that we can a…

…dd sentence segmentation
kermitt2 · Aug 21, 2020 · a1468ae · a1468ae
1 parent f39a3bc
commit a1468ae
Show file tree

Hide file tree

Showing 6 changed files with 242 additions and 216 deletions.
diff --git a/grobid-home/schemas/doc/Grobid_doc.html b/grobid-home/schemas/doc/Grobid_doc.html
diff --git a/grobid-home/schemas/dtd/Grobid.dtd b/grobid-home/schemas/dtd/Grobid.dtd
@@ -1,6 +1,6 @@
 
 <!--
-DTD generated from ODD source 2020-08-20T01:39:39Z. .
+DTD generated from ODD source 2020-08-21T03:03:48Z. .
 TEI Edition: Version 4.1.0. Last updated on
 	19th August 2020, revision b414ba550
 TEI Edition Location: https://www.tei-c.org/Vault/P5/Version 4.1.0/
@@ -1291,7 +1291,7 @@ type %teidata.enumerated;  #IMPLIED >
  %tei_att.global.attributes;
  %tei_att.declaring.attributes; >
 <!--doc:(description of figure) contains a brief prose description of the appearance or content of a graphic figure, for use when documenting an image without displaying it. [14.4. Specific Elements for Graphic Images] -->
-<!ELEMENT figDesc %macro.limitedContent;>
+<!ELEMENT figDesc (#PCDATA|%model.limitedPhrase;|%model.inter;|%model.divLike;)*>
 <!ATTLIST figDesc xmlns CDATA "http://www.tei-c.org/ns/1.0">
 <!ATTLIST figDesc
  %tei_att.global.attributes; >

diff --git a/grobid-home/schemas/odd/Grobid.odd b/grobid-home/schemas/odd/Grobid.odd
@@ -183,6 +183,21 @@
                     </attList>
                 </classSpec>
 
+                <!-- we modify figDesc so that it can contain generic structural divisions -->
+                <elementSpec ident="figDesc" module="figures" mode="change">
+                    <content>
+                        <rng:zeroOrMore>
+                            <rng:choice>
+                                <rng:text/>
+                                <rng:ref name="model.limitedPhrase"/>
+                                <rng:ref name="model.inter"/>
+                                <rng:ref name="model.divLike"/>
+                            </rng:choice>
+                        </rng:zeroOrMore>
+                    </content>
+
+                </elementSpec>
+
                 <!-- removing elements not used in Grobid results -->
                 <elementSpec ident="ab" mode="delete" module="linking"/>
                 <elementSpec ident="abbr" mode="delete" module="core"/>
@@ -357,6 +372,7 @@
                 <elementSpec ident="unclear" mode="delete" module="core"/>
                 <!-- only s from the analysis module, in case we want the final result with sentence segmentation -->
                 <moduleRef key="analysis" except="c cl interp interpGrp m pc phr span spanGrp w"/>
+
             </schemaSpec>
         </body>
     </text>

diff --git a/grobid-home/schemas/rng/Grobid.rnc b/grobid-home/schemas/rng/Grobid.rnc
diff --git a/grobid-home/schemas/rng/Grobid.rng b/grobid-home/schemas/rng/Grobid.rng
@@ -5,7 +5,7 @@
          xmlns="http://relaxng.org/ns/structure/1.0"
          datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes"
          ns="http://www.tei-c.org/ns/1.0"><!--
-Schema generated from ODD source 2020-08-20T01:39:02Z. .
+Schema generated from ODD source 2020-08-21T03:03:09Z. .
 TEI Edition: Version 4.1.0. Last updated on
 	19th August 2020, revision b414ba550
 TEI Edition Location: https://www.tei-c.org/Vault/P5/Version 4.1.0/
@@ -5026,7 +5026,14 @@ Suggested values include: 1] label; 2] data</a:documentation>
    <define name="tei_figDesc">
       <element name="figDesc">
          <a:documentation xmlns:a="http://relaxng.org/ns/compatibility/annotations/1.0">(description of figure) contains a brief prose description of the appearance or content of a graphic figure, for use when documenting an image without displaying it. [14.4. Specific Elements for Graphic Images]</a:documentation>
-         <ref name="tei_macro.limitedContent"/>
+         <zeroOrMore>
+            <choice>
+               <text/>
+               <ref name="tei_model.limitedPhrase"/>
+               <ref name="tei_model.inter"/>
+               <ref name="tei_model.divLike"/>
+            </choice>
+         </zeroOrMore>
          <ref name="tei_att.global.attributes"/>
          <empty/>
       </element>

diff --git a/grobid-home/schemas/xsd/Grobid.xsd b/grobid-home/schemas/xsd/Grobid.xsd
@@ -3,7 +3,7 @@
   <xs:import namespace="http://www.isocat.org/ns/dcr" schemaLocation="dcr.xsd"/>
   <xs:import namespace="http://www.w3.org/XML/1998/namespace" schemaLocation="xml.xsd"/>
   <!--
-    Schema generated from ODD source 2020-08-20T01:39:31Z. .
+    Schema generated from ODD source 2020-08-21T03:03:40Z. .
     TEI Edition: Version 4.1.0. Last updated on
     	19th August 2020, revision b414ba550
     TEI Edition Location: https://www.tei-c.org/Vault/P5/Version 4.1.0/
@@ -4699,12 +4699,13 @@ Suggested values include: 1] label; 2] data</xs:documentation>
     <xs:annotation>
       <xs:documentation>(description of figure) contains a brief prose description of the appearance or content of a graphic figure, for use when documenting an image without displaying it. [14.4. Specific Elements for Graphic Images]</xs:documentation>
     </xs:annotation>
-    <xs:complexType>
-      <xs:complexContent>
-        <xs:extension base="tei:tei_macro.limitedContent">
-          <xs:attributeGroup ref="tei:tei_att.global.attributes"/>
-        </xs:extension>
-      </xs:complexContent>
+    <xs:complexType mixed="true">
+      <xs:choice minOccurs="0" maxOccurs="unbounded">
+        <xs:group ref="tei:tei_model.limitedPhrase"/>
+        <xs:group ref="tei:tei_model.inter"/>
+        <xs:group ref="tei:tei_model.divLike"/>
+      </xs:choice>
+      <xs:attributeGroup ref="tei:tei_att.global.attributes"/>
     </xs:complexType>
   </xs:element>
   <xs:attributeGroup name="tei_att.global.facs.attributes">