skip availability statement eval for e2e PMC set

kermitt2 · Oct 16, 2022 · 0bd482c · 0bd482c
1 parent 134ac7a
commit 0bd482c
Showing 1 changed file with 19 additions and 2 deletions.
diff --git a/grobid-trainer/src/main/java/org/grobid/trainer/evaluation/EndToEndEvaluation.java b/grobid-trainer/src/main/java/org/grobid/trainer/evaluation/EndToEndEvaluation.java
@@ -398,21 +398,38 @@ else if (sectionType == this.FULLTEXT) {
  citationsLabels.remove("doi");
  citationsLabels.remove("pmid");
  citationsLabels.remove("pmcid");
-
+ 
  List<FieldSpecification> toRemove = new ArrayList<>();
  if (citationsFields != null && citationsFields.size() > 0) {
  for(FieldSpecification citationsField : citationsFields) {
  if (citationsField.fieldName.equals("doi") || 
  citationsField.fieldName.equals("pmid") || 
  citationsField.fieldName.equals("pmcid"))
- toRemove.add(citationsField);
+  toRemove.add(citationsField);
  }
  }
  if (toRemove.size() > 0) {
  for(FieldSpecification citationsField : toRemove) {
  citationsFields.remove(citationsField);
  }
  }
+
+ fulltextLabels.remove("availability_stmt");
+
+ toRemove = new ArrayList<>();
+ if (fulltextFields != null && fulltextFields.size() > 0) {
+ for(FieldSpecification fulltextField : fulltextFields) {
+ if (fulltextField.fieldName.equals("availability_stmt")) {
+ // remove availability statements from PMC (not covered and it would make metrics not comparable over time)
+ toRemove.add(fulltextField);
+ }
+ }
+ }
+ if (toRemove.size() > 0) {
+ for(FieldSpecification fulltextField : toRemove) {
+ fulltextFields.remove(fulltextField);
+ }
+ }
  }
 
  File input = new File(xmlInputPath);