95% PedReader

-- Passes significiant unit tests
-- Implicit sample creation for mom / dad when you create single samples
-- Continuing cleanup of Sample and SampleDataSource

public/java/src/org/broadinstitute/sting/gatk/samples/PedReader.java

@@ -160,13 +160,13 @@ public class PedReader {
         final List<String[]> splits = new ArrayList<String[]>(lines.size());
         for ( final String line : lines ) {
             if ( line.startsWith(commentMarker)) continue;
-            String[] parts = line.split("\\W+");
+            String[] parts = line.split("\\s+");
 
             if ( parts.length != nExpectedFields )
                 throw new UserException.MalformedFile(reader.toString(), "Bad PED line " + lineNo + ": wrong number of fields");
 
             if ( phenotypePos != -1 ) {
-                isQT = isQT || CATAGORICAL_TRAIT_VALUES.contains(parts[phenotypePos]);
+                isQT = isQT || ! CATAGORICAL_TRAIT_VALUES.contains(parts[phenotypePos]);
             }
 
             splits.add(parts);
@@ -211,12 +211,21 @@ public class PedReader {
                 }
             }
 
-            final Sample s = new Sample(familyID, sampleDB, individualID, paternalID, maternalID, sex, affection, quantitativePhenotype);
+            final Sample s = new Sample(individualID, sampleDB, familyID, paternalID, maternalID, sex, affection, quantitativePhenotype);
             samples.add(s);
             sampleDB.addSample(s);
             lineNo++;
         }
 
+        for ( final Sample sample : new ArrayList<Sample>(samples) ) {
+            Sample dad = maybeAddImplicitSample(sampleDB, sample.getPaternalID(), sample.getFamilyID(), Gender.MALE);
+            if ( dad != null ) samples.add(dad);
+
+            Sample mom = maybeAddImplicitSample(sampleDB, sample.getMaternalID(), sample.getFamilyID(), Gender.FEMALE);
+            if ( mom != null ) samples.add(mom);
+        }
+
+
         sampleDB.validate(samples);
         return samples;
     }
@@ -227,4 +236,13 @@ public class PedReader {
         else
             return string;
     }
+
+    private final Sample maybeAddImplicitSample(SampleDataSource sampleDB, final String id, final String familyID, final Gender gender) {
+        if ( id != null && sampleDB.getSample(id) == null ) {
+            Sample s = new Sample(id, sampleDB, familyID, null, null, gender, Affection.UNKNOWN, Sample.UNSET_QT);
+            sampleDB.addSample(s);
+            return s;
+        } else
+            return null;
+    }
 }

public/java/src/org/broadinstitute/sting/gatk/samples/Sample.java

@@ -117,8 +117,11 @@ public class Sample implements java.io.Serializable {
         return gender;
     }
 
-    public String getFamilyId() {
+    @Override
-        return familyID;
+    public String toString() {
+        return String.format("Sample %s fam=%s dad=%s mom=%s gender=%s affection=%s qt=%s props=%s",
+                getID(), getFamilyID(), getPaternalID(), getMaternalID(), getGender(), getAffection(),
+                getQuantitativePhenotype(), getExtraProperties());
     }
 
     // -------------------------------------------------------------------------------------
@@ -148,4 +151,34 @@ public class Sample implements java.io.Serializable {
     public boolean hasExtraProperty(String key) {
         return properties.containsKey(key);
     }
+
+    @Override
+    public int hashCode() {
+        return ID.hashCode();
+    }
+
+    @Override
+    public boolean equals(final Object o) {
+        if(o == null)
+            return false;
+        if(o instanceof Sample) {
+            Sample otherSample = (Sample)o;
+            return ID.equals(otherSample.ID) &&
+                    equalOrNull(familyID, otherSample.familyID) &&
+                    equalOrNull(paternalID, otherSample.paternalID) &&
+                    equalOrNull(maternalID, otherSample.maternalID) &&
+                    equalOrNull(gender, otherSample.gender) &&
+                    equalOrNull(quantitativePhenotype, otherSample.quantitativePhenotype) &&
+                    equalOrNull(affection, otherSample.affection) &&
+                    equalOrNull(properties, otherSample.properties);
+        }
+        return false;
+    }
+
+    private final static boolean equalOrNull(final Object o1, final Object o2) {
+        if ( o1 == null )
+            return o2 == null;
+        else
+            return o2 == null ? false : o1.equals(o2);
+    }
 }

public/java/src/org/broadinstitute/sting/gatk/samples/SampleDataSource.java

@@ -156,8 +156,8 @@ public class SampleDataSource {
         HashSet<Sample> familyMembers = new HashSet<Sample>();
 
         for (Sample sample : samples.values()) {
-            if (sample.getFamilyId() != null) {
+            if (sample.getFamilyID() != null) {
-                if (sample.getFamilyId().equals(familyId))
+                if (sample.getFamilyID().equals(familyId))
                     familyMembers.add(sample);
             }
         }
@@ -172,7 +172,7 @@ public class SampleDataSource {
      */
     public Set<Sample> getChildren(Sample sample) {
         HashSet<Sample> children = new HashSet<Sample>();
-        for (Sample familyMember : getFamily(sample.getFamilyId())) {
+        for (Sample familyMember : getFamily(sample.getFamilyID())) {
             if (familyMember.getMother() == sample || familyMember.getFather() == sample) {
                 children.add(familyMember);
             }

public/java/test/org/broadinstitute/sting/gatk/samples/PedReaderUnitTest.java

@@ -33,6 +33,7 @@ import org.testng.annotations.Test;
 import java.io.StringReader;
 import java.util.Arrays;
 import java.util.EnumSet;
+import java.util.HashSet;
 import java.util.List;
 
 /**
@@ -47,6 +48,7 @@ public class PedReaderUnitTest extends BaseTest {
     private class PedReaderTest extends TestDataProvider {
         public String fileContents;
         public List<Sample> expectedSamples;
+        EnumSet<PedReader.MissingPedFields> missing;
 
         private PedReaderTest(final String name, final List<Sample> expectedSamples, final String fileContents) {
             super(PedReaderTest.class, name);
@@ -55,6 +57,19 @@ public class PedReaderUnitTest extends BaseTest {
         }
     }
 
+    private class PedReaderTestMissing extends TestDataProvider {
+        public String fileContents;
+        public List<Sample> expectedSamples;
+        EnumSet<PedReader.MissingPedFields> missing;
+
+        private PedReaderTestMissing(final String name, EnumSet<PedReader.MissingPedFields> missing, final List<Sample> expectedSamples, final String fileContents) {
+            super(PedReaderTest.class, name);
+            this.fileContents = fileContents;
+            this.expectedSamples = expectedSamples;
+            this.missing = missing;
+        }
+    }
+
 //     Family ID
 //     Individual ID
 //     Paternal ID
@@ -100,9 +115,9 @@ public class PedReaderUnitTest extends BaseTest {
 
         new PedReaderTest("multipleUnrelated",
                 Arrays.asList(
-                    new Sample("s1", "fam1", null, null, Gender.MALE,   Affection.AFFECTED),
+                    new Sample("s1", "fam1", null, null, Gender.MALE,   Affection.UNAFFECTED),
-                    new Sample("s2", "fam2", null, null, Gender.FEMALE, Affection.UNAFFECTED)),
+                    new Sample("s2", "fam2", null, null, Gender.FEMALE, Affection.AFFECTED)),
-                String.format("%s\n%s",
+                String.format("%s%n%s",
                         "fam1 s1 0 0 1 1",
                         "fam2 s2 0 0 2 2"));
 
@@ -111,7 +126,7 @@ public class PedReaderUnitTest extends BaseTest {
                     new Sample("kid", "fam1", "dad", "mom", Gender.MALE,   Affection.AFFECTED),
                     new Sample("dad", "fam1", null, null,   Gender.MALE,   Affection.UNAFFECTED),
                     new Sample("mom", "fam1", null, null,   Gender.FEMALE, Affection.AFFECTED)),
-                String.format("%s\n%s\n%s",
+                String.format("%s%n%s%n%s",
                         "fam1 kid dad mom 1 2",
                         "fam1 dad 0   0   1 1",
                         "fam1 mom 0   0   2 2"));
@@ -121,14 +136,14 @@ public class PedReaderUnitTest extends BaseTest {
                     new Sample("kid", "fam1", "dad", "mom", Gender.MALE,   Affection.AFFECTED),
                     new Sample("dad", "fam1", null, null,   Gender.MALE,   Affection.UNKNOWN),
                     new Sample("mom", "fam1", null, null,   Gender.FEMALE, Affection.UNKNOWN)),
-                "fam1 kid dad mom 1 1");
+                "fam1 kid dad mom 1 2");
 
         new PedReaderTest("partialTrio",
                 Arrays.asList(
                     new Sample("kid", "fam1", "dad", "mom", Gender.MALE,   Affection.AFFECTED),
                     new Sample("dad", "fam1", null, null,   Gender.MALE,   Affection.UNAFFECTED),
                     new Sample("mom", "fam1", null, null,   Gender.FEMALE, Affection.UNKNOWN)),
-                String.format("%s\n%s",
+                String.format("%s%n%s",
                         "fam1 kid dad mom 1 2",
                         "fam1 dad 0   0   1 1"));
 
@@ -141,7 +156,7 @@ public class PedReaderUnitTest extends BaseTest {
                     new Sample("mom", "fam1", "granddad2", "grandma2", Gender.FEMALE, Affection.AFFECTED),
                     new Sample("granddad2", "fam1", null, null,        Gender.MALE,   Affection.UNKNOWN),
                     new Sample("grandma2",  "fam1", null, null,        Gender.FEMALE,   Affection.UNKNOWN)),
-                String.format("%s\n%s\n%s",
+                String.format("%s%n%s%n%s",
                         "fam1 kid dad       mom      1 2",
                         "fam1 dad granddad1 grandma1 1 1",
                         "fam1 mom granddad2 grandma2 2 2"));
@@ -151,7 +166,7 @@ public class PedReaderUnitTest extends BaseTest {
                 Arrays.asList(
                     new Sample("s1", "fam1", null, null, Gender.MALE,   Affection.QUANTITATIVE, 1.0),
                     new Sample("s2", "fam2", null, null, Gender.FEMALE, Affection.QUANTITATIVE, 10.0)),
-                String.format("%s\n%s",
+                String.format("%s%n%s",
                         "fam1 s1 0 0 1 1",
                         "fam2 s2 0 0 2 10.0"));
 
@@ -159,7 +174,7 @@ public class PedReaderUnitTest extends BaseTest {
                 Arrays.asList(
                     new Sample("s1", "fam1", null, null, Gender.MALE,   Affection.UNKNOWN, Sample.UNSET_QT),
                     new Sample("s2", "fam2", null, null, Gender.FEMALE, Affection.QUANTITATIVE, 10.0)),
-                String.format("%s\n%s",
+                String.format("%s%n%s",
                         "fam1 s1 0 0 1 -9",
                         "fam2 s2 0 0 2 10.0"));
 
@@ -167,7 +182,7 @@ public class PedReaderUnitTest extends BaseTest {
                 Arrays.asList(
                     new Sample("s1", "fam1", null, null, Gender.MALE,   Affection.QUANTITATIVE, 1.0),
                     new Sample("s2", "fam2", null, null, Gender.FEMALE, Affection.QUANTITATIVE, 10.0)),
-                String.format("%s\n%s",
+                String.format("%s%n%s",
                         "fam1 s1 0 0 1 1",
                         "fam2 s2 0 0 2 10"));
 
@@ -179,7 +194,7 @@ public class PedReaderUnitTest extends BaseTest {
         PedReader reader = new PedReader();
         SampleDataSource sampleDB = new SampleDataSource();
         List<Sample> readSamples = reader.parse(new StringReader(myFileContents), missing, sampleDB);
-        Assert.assertEquals(test.expectedSamples, readSamples, "Parsed incorrect number of samples");
+        Assert.assertEquals(new HashSet<Sample>(test.expectedSamples), new HashSet<Sample>(readSamples), "Parsed incorrect number of samples");
     }
 
     @Test(enabled = true, dataProvider = "readerTest")
@@ -189,13 +204,32 @@ public class PedReaderUnitTest extends BaseTest {
 
     @Test(enabled = true, dataProvider = "readerTest", dependsOnMethods = "testPedReader")
     public void testPedReaderWithComments(PedReaderTest test) {
-        runTest(test, "#comment\n" + test.fileContents, EnumSet.noneOf(PedReader.MissingPedFields.class));
+        runTest(test, String.format("#comment%n%s", test.fileContents), EnumSet.noneOf(PedReader.MissingPedFields.class));
     }
 
-    @Test(enabled = true, dataProvider = "readerTest", dependsOnMethods = "testPedReader")
+    @DataProvider(name = "readerTestMissing")
+    public Object[][] createPEDFilesWithMissing() {
+        new PedReaderTestMissing("trioMissingFam", EnumSet.of(PedReader.MissingPedFields.NO_FAMILY_ID),
+                Arrays.asList(
+                    new Sample("kid", null, "dad", "mom", Gender.MALE,   Affection.AFFECTED),
+                    new Sample("dad", null, null, null,   Gender.MALE,   Affection.UNAFFECTED),
+                    new Sample("mom", null, null, null,   Gender.FEMALE, Affection.AFFECTED)),
+                String.format("%s%n%s%n%s",
+                        "kid dad mom 1 2",
+                        "dad 0   0   1 1",
+                        "mom 0   0   2 2"));
+
+        return PedReaderTestMissing.getTests(PedReaderTestMissing.class);
+    }
+
+    @Test(enabled = true, dataProvider = "readerTestMissing", dependsOnMethods = "testPedReader")
     public void testPedReaderWithMissing(PedReaderTest test) {
-        // todo -- test MISSING by splicing strings
+//        public enum MissingPedFields {
-        //runTest(test, "#comment\n" + test.fileContents, EnumSet.noneOf(PedReader.MissingPedFields.class));
+//            NO_FAMILY_ID,
+//            NO_PARENTS,
+//            NO_SEX,
+//            NO_PHENOTYPE
+//        }
+//        runTest(test, sliceContents(0, test.fileContents), EnumSet.of(PedReader.MissingPedFields.NO_FAMILY_ID));
     }
-
 }