add OBVSG test with fix fo subfield ID length

gradle.properties

@@ -1,6 +1,6 @@
 group = org.xbib
 name = marc
-version = 1.0.13
+version = 1.0.14
 
 xbib-content.version = 1.0.7
 xbib-bibliographic-character-sets.version = 1.0.0

src/main/java/org/xbib/marc/Marc.java

@@ -30,7 +30,6 @@ import org.xbib.marc.io.BufferedSeparatorInputStream;
 import org.xbib.marc.io.BytesReference;
 import org.xbib.marc.io.Chunk;
 import org.xbib.marc.io.ChunkStream;
-import org.xbib.marc.io.InformationSeparator;
 import org.xbib.marc.label.RecordLabel;
 import org.xbib.marc.label.RecordLabelFixer;
 import org.xbib.marc.transformer.MarcTransformer;

src/main/java/org/xbib/marc/MarcField.java

@@ -467,7 +467,7 @@ public class MarcField implements Comparable<MarcField> {
         /**
          * Set subfield with help of record label information from raw data.
          * @param label the record label
-         * @param raw the raw data
+         * @param raw the subfield, including ID and separator
          * @return this builder
          */
         public Builder subfield(RecordLabel label, String raw) {
@@ -481,6 +481,27 @@ public class MarcField implements Comparable<MarcField> {
             return this;
         }
 
+        /**
+         * Set synthetic subfield with help of record label information from raw data.
+         * If the len of the subfield ID is zero or undefined, the dummy subfield ID is used.
+         * @param label the record label
+         * @param dummySubfieldId the dummy subfield ID
+         * @param value the subfield value
+         * @return this builder
+         */
+        public Builder subfield(RecordLabel label, String dummySubfieldId, String value) {
+            int len = label.getSubfieldIdentifierLength() - 1;
+            if (len <= 0) {
+                subfields.add(new Subfield(dummySubfieldId, value));
+                subfieldIds.add(dummySubfieldId);
+            } else if (value.length() >= len) {
+                String id = value.substring(0, len);
+                subfields.add(new Subfield(id, value.substring(len)));
+                subfieldIds.add(id);
+            }
+            return this;
+        }
+
         /**
          * Set a new data field with help of a record label from raw data.
          * @param label the record label

src/main/java/org/xbib/marc/MarcGenerator.java

@@ -40,6 +40,8 @@ import java.util.List;
  */
 public class MarcGenerator implements ChunkListener<byte[], BytesReference>, Closeable {
 
+    private static final String EMPTY = " ";
+
     private String format;
 
     private String type;
@@ -166,11 +168,12 @@ public class MarcGenerator implements ChunkListener<byte[], BytesReference>, Clo
                         int pos = recordLabel.getIndicatorLength();
                         builder.indicator(this.data.substring(0, pos));
                         if (pos < this.data.length()) {
-                            builder.subfield(" ", this.data.substring(pos));
+                            builder.subfield(recordLabel, EMPTY, this.data.substring(pos));
                         }
                     }
                 } else {
                     boolean found = false;
+                    // try more than one position
                     for (int offset = 1; offset < 5; offset++) {
                         if (directory.containsKey(position + offset)) {
                             position = position + offset;
@@ -202,7 +205,7 @@ public class MarcGenerator implements ChunkListener<byte[], BytesReference>, Clo
                 break;
             }
             case US: /* 1f */{
-                builder.subfield(recordLabel, this.data);
+                builder.subfield(recordLabel, EMPTY, this.data);
                 break;
             }
             default: {

src/test/java/org/xbib/marc/SWBTest.java

@@ -48,9 +48,7 @@ public class SWBTest {
             try (InputStream in = getClass().getResource(file).openStream()) {
                 Marc.Builder builder = Marc.builder()
                         .setInputStream(in)
-                        .setCharset(Charset.forName("UTF-8"))
-                        .setFormat("Marc21")
-                        .setType("Bibliographic");
+                        .setCharset(Charset.forName("UTF-8"));
                 for (MarcRecord marcRecord : builder.iterable()) {
                     count++;
                 }

src/test/java/org/xbib/marc/dialects/mab/MabTest.java

@@ -52,6 +52,10 @@ public class MabTest {
 
     private final StringBuilder sb = new StringBuilder();
 
+    /**
+     * Shows how to override erraneous subfield ID length label settings, which is notorious to MAB dialect.
+     * @throws Exception if test faul
+     */
     @Test
     public void testZDB() throws Exception {
         String s = "1217zdbtit.dat";

src/test/java/org/xbib/marc/dialects/mab/OBVSGTest.java

@@ -0,0 +1,64 @@
+package org.xbib.marc.dialects.mab;
+
+import org.junit.Test;
+import org.xbib.marc.Marc;
+import org.xbib.marc.MarcRecord;
+import org.xbib.marc.label.RecordLabel;
+
+import java.io.InputStream;
+import java.nio.charset.Charset;
+import java.util.concurrent.atomic.AtomicInteger;
+
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertTrue;
+
+/**
+ *
+ */
+public class OBVSGTest {
+
+    @Test
+    public void testMarcStream() throws Exception {
+        String[] files = {
+                "obvsg1.mab",
+                "obvsg2.mab",
+                "obvsg3.mab"
+        };
+        for (String file : files) {
+            AtomicInteger count = new AtomicInteger();
+            try (InputStream in = getClass().getResource(file).openStream()) {
+                Marc marc = Marc.builder()
+                        .setInputStream(in)
+                        .setCharset(Charset.forName("UTF-8"))
+                        .build();
+                marc.iso2709Stream().chunks().forEach(chunk -> {
+                    count.incrementAndGet();
+                });
+            }
+            assertTrue(count.get() > 0);
+        }
+    }
+
+    @Test
+    public void testMarcRecordIterable() throws Exception {
+        String[] files = {
+                "obvsg1.mab",
+                "obvsg2.mab",
+                "obvsg3.mab"
+        };
+        for (String file : files) {
+            int count = 0;
+            try (InputStream in = getClass().getResource(file).openStream()) {
+                Marc.Builder builder = Marc.builder()
+                        .setInputStream(in)
+                        .setCharset(Charset.forName("x-MAB"))
+                        .setRecordLabelFixer(recordLabel ->
+                                RecordLabel.builder().from(recordLabel).setSubfieldIdentifierLength(0).build());
+                for (MarcRecord marcRecord : builder.iterable()) {
+                    count++;
+                }
+            }
+            assertEquals(1, count);
+        }
+    }
+}

src/test/resources/org/xbib/marc/bad_leaders_10_11.mrc.xml

@@ -1,15 +1,15 @@
 <?xml version="1.0" encoding="UTF-8"?>
-<collection xmlns="info:lc/xmlns/marcxchange-v2" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="info:lc/xmlns/marcxchange-v2 http://www.loc.gov/standards/iso25577/marcxchange-2-0.xsd">
+<collection xmlns="info:lc/xmlns/marcxchange-v2" xsi:schemaLocation="info:lc/xmlns/marcxchange-v2 http://www.loc.gov/standards/iso25577/marcxchange-2-0.xsd" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
   <record format="MARC21" type="Bibliographic">
     <leader>01794     0000445   000 </leader>
     <controlfield tag="001">000800000020003000008020004200038035001200080005001700092008008100109010001700190016002200207029002200229035001200251035001600263040001800279042000800297043003000305049000900335050002200344072001200366110004800378245024900426246002200675260008500697300004100782504005100823650004300874650004300917650004300960650004301003650004101046610005901087700002001146700002501166938005301191994001201244910002601256991006601282</controlfield>
-    <datafield tag="260" ind1=" " ind2=" ">
+    <datafield ind2=" " ind1=" " tag="260">
       <subfield code="a">0772</subfield>
     </datafield>
-    <datafield tag="200" ind1=" " ind2=" ">
+    <datafield ind2=" " ind1=" " tag="200">
       <subfield code="a">90311033800.0</subfield>
     </datafield>
-    <datafield tag="051" ind1=" " ind2=" ">
+    <datafield ind2=" " ind1=" " tag="051">
       <subfield code="a">028s                                 01609cam  22004094a 4500                </subfield>
     </datafield>
   </record>