keep xml for late serialization

kermitt2 · Aug 7, 2023 · a627fbe · a627fbe
1 parent 4b88b9c
commit a627fbe
Show file tree

Hide file tree

Showing 3 changed files with 11 additions and 6 deletions.
diff --git a/grobid-core/src/main/java/org/grobid/core/data/Funding.java b/grobid-core/src/main/java/org/grobid/core/data/Funding.java
@@ -37,6 +37,9 @@ public class Funding {
 
     private List<LayoutToken> layoutTokens = new ArrayList<>();
 
+    private Date start = null;
+    private Date end = null;
+
     public Funder getFunder() {
         return this.funder;
     }

diff --git a/grobid-core/src/main/java/org/grobid/core/engines/Engine.java b/grobid-core/src/main/java/org/grobid/core/engines/Engine.java
@@ -3,6 +3,8 @@
 import org.apache.commons.lang3.tuple.Pair;
 import org.apache.commons.lang3.tuple.Triple;
 
+import nu.xom.Element;
+
 import org.grobid.core.data.Affiliation;
 import org.grobid.core.data.BibDataSet;
 import org.grobid.core.data.BiblioItem;
@@ -1087,13 +1089,13 @@ public String processFundingAcknowledgement(String text, GrobidAnalysisConfig co
         StringBuilder result = new StringBuilder();
 
         try {
-            Pair<String, Triple<List<Funding>,List<Person>,List<Affiliation>>> localResult = 
+            Pair<Element, Triple<List<Funding>,List<Person>,List<Affiliation>>> localResult = 
                 parsers.getFundingAcknowledgementParser().processing(text, config);
 
             if (localResult == null || localResult.getLeft() == null) 
                 result.append(text);
             else
-                result.append(localResult.getLeft()); 
+                result.append(localResult.getLeft().toXML()); 
 
         } catch (final Exception exp) {
             throw new GrobidException("An exception occured while running Grobid funding-acknowledgement model.", exp);

diff --git a/grobid-core/src/main/java/org/grobid/core/engines/FundingAcknowledgementParser.java b/grobid-core/src/main/java/org/grobid/core/engines/FundingAcknowledgementParser.java
@@ -59,7 +59,7 @@ protected FundingAcknowledgementParser() {
         super(GrobidModels.FUNDING_ACKNOWLEDGEMENT);
     }
 
-    private Pair<String, Triple<List<Funding>,List<Person>,List<Affiliation>>>
+    private Pair<Element, Triple<List<Funding>,List<Person>,List<Affiliation>>>
         processing(List<LayoutToken> tokenizationFunding, GrobidAnalysisConfig config) {
         if (tokenizationFunding == null || tokenizationFunding.size() == 0)
             return null;
@@ -82,7 +82,7 @@ protected FundingAcknowledgementParser() {
      * For convenience, a processing method taking a raw string as input. 
      * Tokenization is done with the default Grobid analyzer triggered by the identified language. 
      **/
-    public Pair<String, Triple<List<Funding>,List<Person>,List<Affiliation>>> processing(String text,
+    public Pair<Element, Triple<List<Funding>,List<Person>,List<Affiliation>>> processing(String text,
                                GrobidAnalysisConfig config) {
         text = UnicodeUtil.normaliseText(text);
         List<LayoutToken> tokenizationFunding = GrobidAnalyzer.getInstance().tokenizeWithLayoutToken(text);
@@ -103,7 +103,7 @@ public Pair<String, Triple<List<Funding>,List<Person>,List<Affiliation>>> proces
      * extracted normalized entities. These entities are referenced by the inline 
      * annotations with the usual @target attribute pointing to xml:id. 
      */
-    private Pair<String, Triple<List<Funding>,List<Person>,List<Affiliation>>>
+    private Pair<Element, Triple<List<Funding>,List<Person>,List<Affiliation>>>
             getExtractionResult(List<LayoutToken> tokenizations, String result) {
         List<Funding> fundings = new ArrayList<>();
         List<Person> persons = new ArrayList<>();
@@ -338,7 +338,7 @@ public Pair<String, Triple<List<Funding>,List<Person>,List<Affiliation>>> proces
 
         Triple<List<Funding>,List<Person>,List<Affiliation>> entities = Triple.of(fundings, persons, affiliations);
 
-        return Pair.of(curParagraph.toXML(), entities);
+        return Pair.of(curParagraph, entities);
     }
 
     /**