changes to handle grobid

changes to handle grobid and warn when the extractors in the xml file doesnt belong to"svm header parse", "grobid" or "parscit"
SeerLabs · Nov 25, 2019 · 0f886cc · 0f886cc
1 parent 2611627
commit 0f886cc
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 5 deletions.
diff --git a/src/java/edu/psu/citeseerx/ingestion/datamappers/BatchMapper.java b/src/java/edu/psu/citeseerx/ingestion/datamappers/BatchMapper.java
@@ -90,11 +90,14 @@ public static edu.psu.citeseerx.domain.Document map(Element root)
                 FileInfoMapper.map(doc, child);
             }
             if (child.getName().equals("algorithm")) {
-                if (child.getAttributeValue("name").equals("SVM HeaderParse")) {
+                if (child.getAttributeValue("name").equals("SVM HeaderParse") || child.getAttributeValue("name").contains("GROBID")) {
                     ParsHedMapper.map(doc, child);
-                }
-                if (child.getAttributeValue("name").equals("ParsCit")) {
+                }else if (child.getAttributeValue("name").equals("ParsCit")) {
                     ParscitMapper.map(doc, child);
+                }else {
+                    System.err.println("WARNING : Algorithm name attribute is not what " +
+                            "was expected: found "+child.getName()+
+                            ", expected SVM HeaderParse, ParsCit or GROBID");
                 }
             }
         }

diff --git a/src/java/edu/psu/citeseerx/ingestion/datamappers/ParsHedMapper.java b/src/java/edu/psu/citeseerx/ingestion/datamappers/ParsHedMapper.java
@@ -79,10 +79,10 @@ public static Document map(String xml)
     public static void map(Document doc, Element root)
     throws MappingException {
 
-        if (!root.getAttributeValue("name").equals(ALG_NAME)) {
+        if (!root.getAttributeValue("name").equals(ALG_NAME) || !root.getAttributeValue("name").contains("GROBID")) {
             throw new MappingException("Root name attribute is not what " +
                     "was expected: found "+root.getAttributeValue("name")+
-                    ", expected "+ALG_NAME);
+                    ", expected "+ALG_NAME + "or GROBID");
         }
         buildDoc(doc, root);