Improve robustness of OSIS import

Do not fail if metadata fields contain line breaks or non-normalized whitespace.
schierlm · Nov 18, 2016 · 48d4dc9 · 48d4dc9
1 parent c55ee61
commit 48d4dc9
Showing 1 changed file with 3 additions and 3 deletions.
diff --git a/biblemulticonverter/src/main/java/biblemulticonverter/format/OSIS.java b/biblemulticonverter/src/main/java/biblemulticonverter/format/OSIS.java
@@ -86,13 +86,13 @@ public Bible doImport(File inputFile) throws Exception {
 			String titleDesc = xpath.evaluate("/osis/osisText/titlePage/description/text()", osisDoc);
 			MetadataBook mb = new MetadataBook();
 			if (!description.isEmpty())
-				mb.setValue(MetadataBookKey.description, description.trim());
+				mb.setValue(MetadataBookKey.description, description.replaceAll("[\r\n\t ]+", " ").trim());
 			if (!rights.isEmpty())
-				mb.setValue(MetadataBookKey.rights, rights.trim());
+				mb.setValue(MetadataBookKey.rights, rights.replaceAll("[\r\n\t ]+", " ").trim());
 			if (!date.isEmpty())
 				mb.setValue(MetadataBookKey.date, date);
 			if (!titleDesc.isEmpty())
-				mb.setValue("description@titlePage", titleDesc.trim());
+				mb.setValue("description@titlePage", titleDesc.replaceAll("[\r\n\t ]+", " ").trim());
 			mb.finished();
 			result.getBooks().add(mb.getBook());
 		}