remove table title from tabula extraction

kermitt2 · Apr 5, 2018 · 9ac30b1 · 9ac30b1
1 parent 69af68a
commit 9ac30b1
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 1 deletion.
diff --git a/grobid-core/src/main/java/org/grobid/core/data/Table.java b/grobid-core/src/main/java/org/grobid/core/data/Table.java
@@ -8,6 +8,7 @@
 import org.grobid.core.engines.Engine;
 import org.grobid.core.engines.config.GrobidAnalysisConfig;
 import org.grobid.core.layout.BoundingBox;
+import org.grobid.core.layout.GraphicObject;
 import org.grobid.core.layout.LayoutToken;
 import org.grobid.core.utilities.BoundingBoxCalculator;
 import org.grobid.core.utilities.LayoutTokensUtil;
@@ -42,6 +43,16 @@ public Table() {
     	content = new StringBuilder();
     	label = new StringBuilder();
     }
+
+    private double headerHeight = 0;
+
+    public void setHeaderHeight(double height) {
+    	headerHeight = height;
+    }
+
+    public double getHeaderHeight() {
+    	return headerHeight;
+    }
 
 	@Override
     public String toTEI(GrobidAnalysisConfig config) {
@@ -134,10 +145,11 @@ private String cleanString(String input) {
     private String[][] tabulaRes = null;
 
     public void tabulaExtract(File pdfFile) throws IOException {
+
     	PDDocument document = PDDocument.load(pdfFile);
     	ObjectExtractor objectExtractor = new ObjectExtractor(document);
     	technology.tabula.Page page = objectExtractor.extract(getPage());
-    	technology.tabula.Page pageArea = page.getArea((float)getY(), (float)getX(), (float)(getY()+getHeight()), (float)(getX()+getWidth()));
+    	technology.tabula.Page pageArea = page.getArea((float)(getY()+getHeaderHeight()), (float)getX(), (float)(getY()+getHeight()), (float)(getX()+getWidth()));
 
     	BasicExtractionAlgorithm bea = new BasicExtractionAlgorithm();
     	technology.tabula.Table table = bea.extract(pageArea).get(0);

diff --git a/grobid-core/src/main/java/org/grobid/core/engines/TableParser.java b/grobid-core/src/main/java/org/grobid/core/engines/TableParser.java
@@ -90,6 +90,8 @@ private Table getExtractionResult(List<LayoutToken> tokenizations, String result
             } else if (clusterLabel.equals(TBL_CONTENT)) {
                 table.appendContent(clusterContent);
                 table.getContentTokens().addAll(tokens);
+                BoundingBox contentBBox = BoundingBoxCalculator.calculateOneBox(tokens, true);
+                table.setHeaderHeight(contentBBox.getY()-bbox.getY());
             } else {
                 LOGGER.error("Warning: unexpected table model label - " + clusterLabel + " for " + clusterContent);
             }