Merge pull request #838 from kermitt2/prioritize_crossref_author_meta

prefer author meta from consolidation , crossref is considered more r…
kermitt2 · Oct 15, 2022 · 134ac7a · 134ac7a
2 parents 197ecd0 + 6abbd4b
commit 134ac7a
Showing 1 changed file with 6 additions and 4 deletions.
diff --git a/grobid-core/src/main/java/org/grobid/core/data/BiblioItem.java b/grobid-core/src/main/java/org/grobid/core/data/BiblioItem.java
@@ -3978,6 +3978,9 @@ public static void injectIdentifiers(BiblioItem destination, BiblioItem source)
 
  /**
  * Correct fields of the first biblio item based on the second one and the reference string
+ * 
+ * @param bib extracted from document
+ * @param bibo fetched from metadata provider (biblioglutton, crossref..)
  */
  public static void correct(BiblioItem bib, BiblioItem bibo) {
  //System.out.println("correct: \n" + bib.toTEI(0));
@@ -4130,8 +4133,8 @@ else if (bibo.getFullAuthors().size() == 1) {
  // should we also check the country ? affiliation?
  if (StringUtils.isNotBlank(auto.getMiddleName()) && (StringUtils.isBlank(aut.getMiddleName())))
  aut.setMiddleName(auto.getMiddleName());
- if (StringUtils.isNotBlank(auto.getORCID()) && (StringUtils.isBlank(aut.getORCID())))
-  aut.setORCID(auto.getORCID());
+ // crossref is considered more reliable than PDF annotations
+ aut.setORCID(auto.getORCID());
  }
  }
  }
@@ -4175,8 +4178,6 @@ else if (bibo.getFullAuthors().size() == 1) {
  aut.setTitle(aut2.getTitle());
  if (StringUtils.isBlank(aut.getSuffix()))
  aut.setSuffix(aut2.getSuffix());
- if (StringUtils.isBlank(aut.getORCID()))
- aut.setORCID(aut2.getORCID());
  if (StringUtils.isBlank(aut.getEmail()))
  aut.setEmail(aut2.getEmail());
  if(!CollectionUtils.isEmpty(aut2.getAffiliations()))
@@ -4189,6 +4190,7 @@ else if (bibo.getFullAuthors().size() == 1) {
  aut.setMarkers(aut2.getMarkers());
  if (!CollectionUtils.isEmpty(aut2.getLayoutTokens())) 
  aut.setLayoutTokens(aut2.getLayoutTokens());
+ // crossref is considered more reliable than PDF annotations, so ORCIDs are not overwritten
  break;
  } 
  }