Skip to content

Commit

Permalink
update XML schema
Browse files Browse the repository at this point in the history
  • Loading branch information
@kermitt2
kermitt2 committed Feb 9, 2024
1 parent e1ecac9 commit bcce229
Show file tree
Hide file tree
Showing 8 changed files with 1,144 additions and 283 deletions.
297 changes: 162 additions & 135 deletions grobid-home/schemas/doc/Grobid_doc.html
View file

Large diffs are not rendered by default.

106 changes: 74 additions & 32 deletions grobid-home/schemas/dtd/Grobid.dtd
View file
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@

<!--
DTD generated from ODD source 2024-02-08T19:57:10Z. .
DTD generated from ODD source 2024-02-09T13:55:02Z. .
TEI Edition: P5 Version 4.7.0. Last updated on 16th November 2023, revision e5dd73ed0
TEI Edition Location: https://www.tei-c.org/Vault/P5/4.7.0/

Expand Down Expand Up @@ -297,6 +297,10 @@ scope %teidata.enumerated; #IMPLIED'>
hand %teidata.pointer; #IMPLIED'>
<!ENTITY % tei_att.written.attributes '
%tei_att.written.attribute.hand;'>
<!ENTITY % tei_att.breaking.attribute.break '
break %teidata.enumerated; #IMPLIED'>
<!ENTITY % tei_att.breaking.attributes '
%tei_att.breaking.attribute.break;'>
<!ENTITY % tei_att.cReferencing.attribute.cRef '
cRef %teidata.text; #IMPLIED'>
<!ENTITY % tei_att.cReferencing.attributes '
Expand Down Expand Up @@ -493,6 +497,18 @@ function %teidata.enumerated; #IMPLIED'>
sortKey %teidata.word; #IMPLIED'>
<!ENTITY % tei_att.sortable.attributes '
%tei_att.sortable.attribute.sortKey;'>
<!ENTITY % tei_att.edition.attribute.ed '
ed CDATA #IMPLIED'>

<!ENTITY % tei_att.edition.attribute.edRef '
edRef CDATA #IMPLIED'>
<!ENTITY % tei_att.edition.attributes '
%tei_att.edition.attribute.ed;
%tei_att.edition.attribute.edRef;'>
<!ENTITY % tei_att.spanning.attribute.spanTo '
spanTo %teidata.pointer; #IMPLIED'>
<!ENTITY % tei_att.spanning.attributes '
%tei_att.spanning.attribute.spanTo;'>
<!ENTITY % tei_att.timed.attribute.start '
start %teidata.pointer; #IMPLIED'>

Expand Down Expand Up @@ -540,11 +556,11 @@ to %teidata.word; #IMPLIED'>
<!ENTITY % model.hiLike_sequenceOptional "hi?">
<!ENTITY % model.hiLike_sequenceOptionalRepeatable "hi*">
<!ENTITY % model.hiLike_sequenceRepeatable "hi+">
<!ENTITY % model.emphLike "term |title">
<!ENTITY % model.emphLike_sequence "term, title">
<!ENTITY % model.emphLike_sequenceOptional "term?, title?">
<!ENTITY % model.emphLike_sequenceOptionalRepeatable "term*, title*">
<!ENTITY % model.emphLike_sequenceRepeatable "term+, title+">
<!ENTITY % model.emphLike "gloss |term |title">
<!ENTITY % model.emphLike_sequence "gloss, term, title">
<!ENTITY % model.emphLike_sequenceOptional "gloss?, term?, title?">
<!ENTITY % model.emphLike_sequenceOptionalRepeatable "gloss*, term*, title*">
<!ENTITY % model.emphLike_sequenceRepeatable "gloss+, term+, title+">
<!ENTITY % model.highlighted "%model.hiLike; |%model.emphLike;">
<!ENTITY % model.highlighted_sequence "%model.hiLike;, %model.emphLike;">
<!ENTITY % model.highlighted_sequenceOptional "%model.hiLike;?, %model.emphLike;?">
Expand Down Expand Up @@ -610,11 +626,11 @@ to %teidata.word; #IMPLIED'>
<!ENTITY % model.global.meta_sequenceOptional "link?">
<!ENTITY % model.global.meta_sequenceOptionalRepeatable "link*">
<!ENTITY % model.global.meta_sequenceRepeatable "link+">
<!ENTITY % model.milestoneLike "anchor">
<!ENTITY % model.milestoneLike_sequence "anchor">
<!ENTITY % model.milestoneLike_sequenceOptional "anchor?">
<!ENTITY % model.milestoneLike_sequenceOptionalRepeatable "anchor*">
<!ENTITY % model.milestoneLike_sequenceRepeatable "anchor+">
<!ENTITY % model.milestoneLike "lb |anchor">
<!ENTITY % model.milestoneLike_sequence "lb, anchor">
<!ENTITY % model.milestoneLike_sequenceOptional "lb?, anchor?">
<!ENTITY % model.milestoneLike_sequenceOptionalRepeatable "lb*, anchor*">
<!ENTITY % model.milestoneLike_sequenceRepeatable "lb+, anchor+">
<!ENTITY % model.gLike "">
<!ENTITY % model.gLike_sequence "">
<!ENTITY % model.gLike_sequenceOptional "">
Expand Down Expand Up @@ -665,11 +681,11 @@ to %teidata.word; #IMPLIED'>
<!ENTITY % model.lLike_sequenceOptional "">
<!ENTITY % model.lLike_sequenceOptionalRepeatable "">
<!ENTITY % model.lLike_sequenceRepeatable "">
<!ENTITY % model.pLike "p |formula |s |trash">
<!ENTITY % model.pLike_sequence "p, formula, s, trash">
<!ENTITY % model.pLike_sequenceOptional "p?, formula?, s?, trash?">
<!ENTITY % model.pLike_sequenceOptionalRepeatable "p*, formula*, s*, trash*">
<!ENTITY % model.pLike_sequenceRepeatable "p+, formula+, s+, trash+">
<!ENTITY % model.pLike "%model.segLike; |p |formula |trash">
<!ENTITY % model.pLike_sequence "%model.segLike;, p, formula, trash">
<!ENTITY % model.pLike_sequenceOptional "%model.segLike;?, p?, formula?, trash?">
<!ENTITY % model.pLike_sequenceOptionalRepeatable "%model.segLike;*, p*, formula*, trash*">
<!ENTITY % model.pLike_sequenceRepeatable "%model.segLike;+, p+, formula+, trash+">
<!ENTITY % model.stageLike "">
<!ENTITY % model.stageLike_sequence "">
<!ENTITY % model.stageLike_sequenceOptional "">
Expand Down Expand Up @@ -740,16 +756,16 @@ to %teidata.word; #IMPLIED'>
<!ENTITY % model.descLike_sequenceOptional "desc?">
<!ENTITY % model.descLike_sequenceOptionalRepeatable "desc*">
<!ENTITY % model.descLike_sequenceRepeatable "desc+">
<!ENTITY % model.quoteLike "">
<!ENTITY % model.quoteLike_sequence "">
<!ENTITY % model.quoteLike_sequenceOptional "">
<!ENTITY % model.quoteLike_sequenceOptionalRepeatable "">
<!ENTITY % model.quoteLike_sequenceRepeatable "">
<!ENTITY % model.attributable "">
<!ENTITY % model.attributable_sequence "">
<!ENTITY % model.attributable_sequenceOptional "">
<!ENTITY % model.attributable_sequenceOptionalRepeatable "">
<!ENTITY % model.attributable_sequenceRepeatable "">
<!ENTITY % model.quoteLike "quote">
<!ENTITY % model.quoteLike_sequence "quote">
<!ENTITY % model.quoteLike_sequenceOptional "quote?">
<!ENTITY % model.quoteLike_sequenceOptionalRepeatable "quote*">
<!ENTITY % model.quoteLike_sequenceRepeatable "quote+">
<!ENTITY % model.attributable "%model.quoteLike;">
<!ENTITY % model.attributable_sequence "%model.quoteLike;">
<!ENTITY % model.attributable_sequenceOptional "%model.quoteLike;?">
<!ENTITY % model.attributable_sequenceOptionalRepeatable "%model.quoteLike;*">
<!ENTITY % model.attributable_sequenceRepeatable "%model.quoteLike;+">
<!ENTITY % model.respLike "author |editor |meeting">
<!ENTITY % model.respLike_sequence "author, editor, meeting">
<!ENTITY % model.respLike_sequenceOptional "author?, editor?, meeting?">
Expand Down Expand Up @@ -830,11 +846,11 @@ to %teidata.word; #IMPLIED'>
<!ENTITY % model.pPart.data_sequenceOptional "%model.dateLike;?, %model.measureLike;?, %model.addressLike;?, %model.nameLike;?">
<!ENTITY % model.pPart.data_sequenceOptionalRepeatable "%model.dateLike;*, %model.measureLike;*, %model.addressLike;*, %model.nameLike;*">
<!ENTITY % model.pPart.data_sequenceRepeatable "%model.dateLike;+, %model.measureLike;+, %model.addressLike;+, %model.nameLike;+">
<!ENTITY % model.inter "%model.biblLike; |%model.labelLike; |%model.listLike;">
<!ENTITY % model.inter_sequence "%model.biblLike;, %model.labelLike;, %model.listLike;">
<!ENTITY % model.inter_sequenceOptional "%model.biblLike;?, %model.labelLike;?, %model.listLike;?">
<!ENTITY % model.inter_sequenceOptionalRepeatable "%model.biblLike;*, %model.labelLike;*, %model.listLike;*">
<!ENTITY % model.inter_sequenceRepeatable "%model.biblLike;+, %model.labelLike;+, %model.listLike;+">
<!ENTITY % model.inter "%model.biblLike; |%model.labelLike; |%model.listLike; |%model.attributable;">
<!ENTITY % model.inter_sequence "%model.biblLike;, %model.labelLike;, %model.listLike;, %model.attributable;">
<!ENTITY % model.inter_sequenceOptional "%model.biblLike;?, %model.labelLike;?, %model.listLike;?, %model.attributable;?">
<!ENTITY % model.inter_sequenceOptionalRepeatable "%model.biblLike;*, %model.labelLike;*, %model.listLike;*, %model.attributable;*">
<!ENTITY % model.inter_sequenceRepeatable "%model.biblLike;+, %model.labelLike;+, %model.listLike;+, %model.attributable;+">
<!ENTITY % model.common "%model.divPart; |%model.inter;">
<!ENTITY % model.common_sequence "%model.divPart;, %model.inter;">
<!ENTITY % model.common_sequenceOptional "%model.divPart;?, %model.inter;?">
Expand Down Expand Up @@ -1103,7 +1119,7 @@ ana CDATA #IMPLIED'>

<!ENTITY % macro.limitedContent '(#PCDATA|%model.limitedPhrase;|%model.inter;)*' >

<!ENTITY % macro.phraseSeq '(#PCDATA|_DUMMY_model.gLike|_DUMMY_model.attributable|%model.phrase;|%model.global;)*' >
<!ENTITY % macro.phraseSeq '(#PCDATA|_DUMMY_model.gLike|%model.attributable;|%model.phrase;|%model.global;)*' >

<!ENTITY % macro.phraseSeq.limited '(#PCDATA|%model.limitedPhrase;|%model.global;)*' >

Expand Down Expand Up @@ -1498,6 +1514,15 @@ calendar CDATA #IMPLIED >
%tei_att.typed.attributes;
calendar CDATA #IMPLIED
value CDATA #IMPLIED >
<!--doc:(gloss) identifies a phrase or word used to provide a gloss or definition for some other word or phrase. [3.4.1. Terms and Glosses 22.4.1. Description of Components] -->
<!ELEMENT gloss %macro.phraseSeq;>
<!ATTLIST gloss xmlns CDATA "http://www.tei-c.org/ns/1.0">
<!ATTLIST gloss
%tei_att.global.attributes;
%tei_att.declaring.attributes;
%tei_att.typed.attributes;
%tei_att.pointing.attributes;
%tei_att.cReferencing.attributes; >
<!--doc:(graphic) indicates the location of a graphic or illustration, either forming part of a text, or providing an image of it. [3.10. Graphics and Other Non-textual Components 11.1. Digital Facsimiles] -->
<!ELEMENT graphic (%model.descLike;)*>
<!ATTLIST graphic xmlns CDATA "http://www.tei-c.org/ns/1.0">
Expand Down Expand Up @@ -1556,6 +1581,16 @@ scheme %teidata.pointer; #IMPLIED >
%tei_att.typed.attributes;
%tei_att.placement.attributes;
%tei_att.written.attributes; >
<!--doc:(line beginning) marks the beginning of a new (typographic) line in some edition or version of a text. [3.11.3. Milestone
Elements 7.2.5. Speech Contents] -->
<!ELEMENT lb EMPTY>
<!ATTLIST lb xmlns CDATA "http://www.tei-c.org/ns/1.0">
<!ATTLIST lb
%tei_att.global.attributes;
%tei_att.typed.attributes;
%tei_att.edition.attributes;
%tei_att.spanning.attributes;
%tei_att.breaking.attributes; >
<!--doc:contains information about a licence or other legal agreement applicable to the text. [2.2.4. Publication, Distribution, Licensing, etc.] -->
<!ELEMENT licence %macro.specialPara;>
<!ATTLIST licence xmlns CDATA "http://www.tei-c.org/ns/1.0">
Expand Down Expand Up @@ -1776,6 +1811,13 @@ age %teidata.enumerated; #IMPLIED >
<!ATTLIST publisher
%tei_att.global.attributes;
%tei_att.canonical.attributes; >
<!--doc:(quotation) contains a phrase or passage attributed by the narrator or author to some agency external to the text. [3.3.3. Quotation 4.3.1. Grouped Texts] -->
<!ELEMENT quote %macro.specialPara;>
<!ATTLIST quote xmlns CDATA "http://www.tei-c.org/ns/1.0">
<!ATTLIST quote
%tei_att.global.attributes;
%tei_att.typed.attributes;
%tei_att.notated.attributes; >
<!--doc:(ruby base) contains the base text annotated by a ruby gloss. [3.4.2. Ruby Annotations] -->
<!ELEMENT rb %macro.phraseSeq;>
<!ATTLIST rb xmlns CDATA "http://www.tei-c.org/ns/1.0">
Expand Down
Loading

0 comments on commit bcce229

Please sign in to comment.