filter out fp crossref results

kermitt2 · Aug 13, 2023 · 2d1e9ed · 2d1e9ed
1 parent a28e2a1
commit 2d1e9ed
Show file tree

Hide file tree

Showing 4 changed files with 47 additions and 31 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
@@ -203,6 +203,19 @@ public boolean isValid() {
             return false;
     }
 
+    public boolean isNonEmptyFunding() {
+        if (grantNumber != null || 
+            grantName != null || 
+            projectFullName != null || 
+            projectAbbreviatedName != null || 
+            programFullName != null || 
+            programAbbreviatedName != null ||
+            url != null)
+            return true;
+        else
+            return false;
+    }
+
     public String toString() {
         StringBuilder builder = new StringBuilder();
         if (funder != null)

diff --git a/grobid-core/src/main/java/org/grobid/core/document/TEIFormatter.java b/grobid-core/src/main/java/org/grobid/core/document/TEIFormatter.java
@@ -191,7 +191,8 @@ public StringBuilder toTEIHeader(BiblioItem biblio,
                 // inject funding ref in the funder entries
                 String referenceString = "";
                 for(Funding funderFunding : entry.getValue()) {
-                    referenceString += " #" + funderFunding.getIdentifier();
+                    if (funderFunding.isNonEmptyFunding())
+                        referenceString += " #" + funderFunding.getIdentifier();
                 }
 
                 if (funderPiece != null) {

diff --git a/grobid-core/src/main/java/org/grobid/core/engines/FullTextParser.java b/grobid-core/src/main/java/org/grobid/core/engines/FullTextParser.java
@@ -2465,6 +2465,33 @@ private void toTEI(Document doc,
 
             List<String> annexStatements = new ArrayList<>();
 
+            // acknowledgement is in the back
+            StringBuilder acknowledgmentStmt = getSectionAsTEI("acknowledgement", "\t\t\t", doc, SegmentationLabels.ACKNOWLEDGEMENT,
+                teiFormatter, resCitations, config);
+
+            if (acknowledgmentStmt.length() > 0) {
+                MutablePair<Element, MutableTriple<List<Funding>,List<Person>,List<Affiliation>>> localResult = 
+                    parsers.getFundingAcknowledgementParser().processingXmlFragment(acknowledgmentStmt.toString(), config);
+
+                if (localResult != null && localResult.getLeft() != null) {
+                    String local_tei = localResult.getLeft().toXML();
+                    local_tei = local_tei.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
+                    //tei.append(local_tei);
+                    annexStatements.add(local_tei);
+                }
+                else {
+                    //tei.append(acknowledgmentStmt);
+                    annexStatements.add(acknowledgmentStmt.toString());
+                }
+
+                if (localResult != null && localResult.getRight() != null && localResult.getRight().getLeft() != null) {
+                    List<Funding> localFundings = localResult.getRight().getLeft();
+                    if (localFundings.size()>0) {
+                        fundings.addAll(localFundings);
+                    }
+                }
+            }
+
             // funding in header
             StringBuilder fundingStmt = new StringBuilder();
             if (StringUtils.isNotBlank(resHeader.getFunding())) {
@@ -2529,33 +2556,6 @@ private void toTEI(Document doc,
                     annexStatements.add(fundingStmt.toString());
                 }
 
-                if (localResult != null && localResult.getRight() != null && localResult.getRight().getLeft() != null) {
-                    List<Funding> localFundings = localResult.getRight().getLeft();
-                    if (localFundings.size()>0) {
-                        fundings.addAll(localFundings);
-                    }
-                }
-            }
-
-			// acknowledgement is in the back
-            StringBuilder acknowledgmentStmt = getSectionAsTEI("acknowledgement", "\t\t\t", doc, SegmentationLabels.ACKNOWLEDGEMENT,
-                teiFormatter, resCitations, config);
-
-            if (acknowledgmentStmt.length() > 0) {
-                MutablePair<Element, MutableTriple<List<Funding>,List<Person>,List<Affiliation>>> localResult = 
-                    parsers.getFundingAcknowledgementParser().processingXmlFragment(acknowledgmentStmt.toString(), config);
-
-                if (localResult != null && localResult.getLeft() != null) {
-                    String local_tei = localResult.getLeft().toXML();
-                    local_tei = local_tei.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
-                    //tei.append(local_tei);
-                    annexStatements.add(local_tei);
-                }
-                else {
-                    //tei.append(acknowledgmentStmt);
-                    annexStatements.add(acknowledgmentStmt.toString());
-                }
-
                 if (localResult != null && localResult.getRight() != null && localResult.getRight().getLeft() != null) {
                     List<Funding> localFundings = localResult.getRight().getLeft();
                     if (localFundings.size()>0) {
@@ -2577,7 +2577,8 @@ private void toTEI(Document doc,
 
             tei.append("\t\t\t<listOrg type=\"funding\">\n");
             for(Funding funding : fundings) {
-                tei.append(funding.toTEI(3));
+                if (funding.isNonEmptyFunding())
+                    tei.append(funding.toTEI(3));
             }
             tei.append("\t\t\t</listOrg>");
 

diff --git a/grobid-core/src/main/java/org/grobid/core/utilities/Consolidation.java b/grobid-core/src/main/java/org/grobid/core/utilities/Consolidation.java
@@ -588,9 +588,10 @@ public void onSuccess(List<Funder> res) {
                         for(Funder oneRes : res) {
                             /* 
                               Glutton integrates its own post-validation, so we can skip post-validation in GROBID when it is used as 
-                              consolidation service.  
+                              consolidation service. However, with CrossRef, post-validation is mandatory to control false positives.  
                             */
-                            results.add(oneRes);
+                            if (oneRes.getFullName() != null && oneRes.getFullName().toLowerCase().equals(funder.getFullName().toLowerCase()))
+                                results.add(oneRes);
                         }
                     }
                 }