Added some generic read tag/expression fliters for use on numeric tags

src/main/java/org/broadinstitute/hellbender/cmdline/ReadFilterArgumentDefinitions.java

@@ -56,5 +56,11 @@ private ReadFilterArgumentDefinitions(){}
 
     public static final String INVERT_SOFT_CLIP_RATIO_FILTER = "invert-soft-clip-ratio-filter";
 
+    public static final String READ_FILTER_TAG = "read-filter-tag";
+    public static final String READ_FILTER_TAG_COMP = "read-filter-tag-comp";
+    public static final String READ_FILTER_TAG_OP = "read-filter-tag-op";
+    public static final String READ_FILTER_EXPRESSION_LONG_NAME = "read-filter-expression";
+
+
     public static final String MATE_TOO_DISTANT_LENGTH = "mate-too-distant-length";
 }

src/main/java/org/broadinstitute/hellbender/engine/filters/JexlExpressionReadTagValueFilter.java

@@ -0,0 +1,95 @@
+package org.broadinstitute.hellbender.engine.filters;
+
+import htsjdk.samtools.util.Lazy;
+import htsjdk.variant.variantcontext.VariantContextUtils;
+import org.apache.commons.jexl2.Expression;
+import org.apache.commons.jexl2.JexlContext;
+import org.apache.logging.log4j.LogManager;
+import org.apache.logging.log4j.Logger;
+import org.broadinstitute.barclay.argparser.Argument;
+import org.broadinstitute.barclay.help.DocumentedFeature;
+import org.broadinstitute.hellbender.cmdline.ReadFilterArgumentDefinitions;
+import org.broadinstitute.hellbender.tools.walkers.haplotypecaller.HaplotypeCallerEngine;
+import org.broadinstitute.hellbender.utils.help.HelpConstants;
+import org.broadinstitute.hellbender.utils.read.GATKRead;
+
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.LinkedList;
+import java.util.List;
+import java.util.function.BiFunction;
+
+/**
+ * Keep only reads that the attributes of meet a given set of jexl expressions
+ */
+@DocumentedFeature(groupName = HelpConstants.DOC_CAT_READFILTERS, groupSummary = HelpConstants.DOC_CAT_READFILTERS_SUMMARY,
+        summary = "Keep only reads that meet all given jexl expressions (on their attributes)")
+public final class JexlExpressionReadTagValueFilter extends ReadFilter {
+    private static final long serialVersionUID = 1L;
+    private static final Logger logger = LogManager.getLogger(JexlExpressionReadTagValueFilter.class);
+
+    @Argument(fullName=ReadFilterArgumentDefinitions.READ_FILTER_EXPRESSION_LONG_NAME, doc="One or more JEXL expressions used to filter", optional=false)
+    public List<String> filterExpressions = new ArrayList<>();
+
+    private Lazy<List<Expression>> jexlExprs = new Lazy<>(() -> {
+        List<Expression>        l = new LinkedList<>();
+        for ( String expr : filterExpressions ) {
+            final Expression    jexl =  VariantContextUtils.engine.get().createExpression(expr);
+            logger.info("created jexl: " + jexl);
+            l.add(jexl);
+        }
+        return l;
+    });
+
+    private static class GATKReadJexlContext implements JexlContext {
+
+        final private GATKRead        read;
+
+        GATKReadJexlContext(final GATKRead read) {
+            this.read = read;
+        }
+
+        @Override
+        public Object get(final String name) {
+            return read.getAttributeAsString(name);
+        }
+
+        @Override
+        public void set(final String name, final Object value) {
+            throw new IllegalArgumentException("setting attributes is not allowed");
+        }
+
+        @Override
+        public boolean has(final String name) {
+            return read.hasAttribute(name);
+        }
+    }
+
+    public JexlExpressionReadTagValueFilter() {
+    }
+
+    // convenience constructor for using a single jexl expression
+    public JexlExpressionReadTagValueFilter(final String jexlExpr) {
+        this.filterExpressions = Collections.singletonList(jexlExpr);
+    }
+
+    // convenience constructor for using a multiple jexl expressions
+    public JexlExpressionReadTagValueFilter(final List<String> jexlExprs) {
+        this.filterExpressions = jexlExprs;
+    }
+
+    @Override
+    public boolean test(final GATKRead read) {
+
+        // loop over expressions. At this point expressions are ANDed
+        for ( Expression expr : jexlExprs.get() ) {
+            Object v = expr.evaluate(new GATKReadJexlContext(read));
+            if (!v.equals(Boolean.TRUE)) {
+                return false;
+            }
+        }
+
+        // if here, all expressions matched
+        return true;
+    }
+}

src/main/java/org/broadinstitute/hellbender/engine/filters/ReadTagValueFilter.java

@@ -0,0 +1,78 @@
+package org.broadinstitute.hellbender.engine.filters;
+
+import htsjdk.samtools.util.Lazy;
+import htsjdk.variant.variantcontext.VariantContextUtils;
+import org.apache.commons.jexl2.JexlContext;
+import org.apache.commons.jexl2.JexlEngine;
+import org.broadinstitute.barclay.argparser.Argument;
+import org.broadinstitute.barclay.help.DocumentedFeature;
+import org.broadinstitute.hellbender.cmdline.ReadFilterArgumentDefinitions;
+import org.broadinstitute.hellbender.utils.help.HelpConstants;
+import org.broadinstitute.hellbender.utils.read.GATKRead;
+
+import org.apache.commons.jexl2.Expression;
+
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.LinkedList;
+import java.util.List;
+import java.util.function.BiFunction;
+
+/**
+ * Keep only reads that contain a tag with a value that agrees with parameters as specified.
+ *
+ */
+@DocumentedFeature(groupName = HelpConstants.DOC_CAT_READFILTERS, groupSummary = HelpConstants.DOC_CAT_READFILTERS_SUMMARY,
+        summary = "Keep only reads that contains a tag with value that agrees with parameters")
+public final class ReadTagValueFilter extends ReadFilter {
+    private static final long serialVersionUID = 1L;
+
+    @Argument(fullName = ReadFilterArgumentDefinitions.READ_FILTER_TAG,
+            doc = "Look for this tag in read", optional=false)
+    public String readFilterTagName = null;
+
+    @Argument(fullName = ReadFilterArgumentDefinitions.READ_FILTER_TAG_COMP,
+            doc = "Compare value in tag to this value", optional=true)
+    public Float readFilterTagComp = 0F;
+
+    public enum Operator {
+        LESS((Float x, Float y) -> x < y),
+        LESS_OR_EQUAL((Float x, Float y) -> x <= y),
+        GREATER((Float x, Float y) -> x > y),
+        GREATER_OR_EQUAL((Float x, Float y) -> x >= y),
+        EQUAL(Float::equals),
+        NOT_EQUAL((Float x, Float y) -> !x.equals(y));
+
+        final BiFunction<Float, Float, Boolean> comp;
+
+        Operator(BiFunction<Float, Float, Boolean> comp) {
+            this.comp = comp;
+        }
+    }
+
+    @Argument(fullName = ReadFilterArgumentDefinitions.READ_FILTER_TAG_OP,
+            doc = "Compare value in tag to value with this operator. " +
+                    "If T is the value in the tag, OP is the operation provided, " +
+                    "and V is the value in read-filter-tag, then the " +
+                    "read will pass the filter iff T OP V is true.", optional = true)
+    public Operator readFilterTagOp = Operator.EQUAL;
+
+    public ReadTagValueFilter() {
+    }
+
+    // convenience constructor for using <tag operator value> form
+    public ReadTagValueFilter(final String tagName, final float tagValue, final Operator operator) {
+        this.readFilterTagName = tagName;
+        this.readFilterTagComp = tagValue;
+        this.readFilterTagOp = operator;
+    }
+
+    @Override
+    public boolean test(final GATKRead read) {
+
+        return read.hasAttribute(this.readFilterTagName) &&
+                this.readFilterTagComp != null &&
+                this.readFilterTagOp.comp.apply(read.getAttributeAsFloat(this.readFilterTagName),
+                        this.readFilterTagComp);
+    }
+}

src/main/java/org/broadinstitute/hellbender/utils/read/GATKRead.java

@@ -762,6 +762,16 @@ default Optional<SamPairUtil.PairOrientation> getPairOrientation() {
      */
     Integer getAttributeAsInteger( final String attributeName );
 
+    /**
+     * Retrieve the value of a particular attribute typed as an floatingpoint value.
+     *
+     * @param attributeName name of the attribute to retrieve
+     * @return integer value of the requested attribute, or {@code null} if the attribute is not present
+     * @throws GATKException.ReadAttributeTypeMismatch if the attribute
+     *         value cannot be typed as a float
+     */
+    Float getAttributeAsFloat( final String attributeName );
+
     /**
      * Retrieve the value of a particular attribute typed as a String.
      *

src/main/java/org/broadinstitute/hellbender/utils/read/SAMRecordToGATKReadAdapter.java

@@ -630,6 +630,27 @@ else if ( attributeValue instanceof Integer ) {
         }
     }
 
+    @Override
+    public Float getAttributeAsFloat( final String attributeName ) {
+        ReadUtils.assertAttributeNameIsLegal(attributeName);
+        final Object attributeValue = samRecord.getAttribute(attributeName);
+
+        if ( attributeValue == null ) {
+            return null;
+        }
+        else if ( attributeValue instanceof Float ) {
+            return (Float)attributeValue;
+        }
+        else {
+            try {
+                return Float.parseFloat(attributeValue.toString());
+            }
+            catch ( NumberFormatException e ) {
+                throw new GATKException.ReadAttributeTypeMismatch(attributeName, "integer", e);
+            }
+        }
+    }
+
     @Override
     public String getAttributeAsString( final String attributeName ) {
         ReadUtils.assertAttributeNameIsLegal(attributeName);

src/test/java/org/broadinstitute/hellbender/engine/filters/JexlExpressionReadTagValueFilterUnitTest.java

@@ -0,0 +1,89 @@
+package org.broadinstitute.hellbender.engine.filters;
+
+import org.broadinstitute.hellbender.GATKBaseTest;
+import org.broadinstitute.hellbender.utils.read.GATKRead;
+import org.testng.Assert;
+import org.testng.annotations.DataProvider;
+import org.testng.annotations.Test;
+
+import java.util.Arrays;
+import java.util.Iterator;
+import java.util.LinkedList;
+import java.util.List;
+
+public class JexlExpressionReadTagValueFilterUnitTest extends GATKBaseTest {
+
+    @Test(dataProvider= "JexlExpressionReadTagValueFilterDataProvider")
+    public void testJexlExpressionReadTagValueFilter(final String cigarString,
+                                      final Object[] attrsNameAndValue,
+                                      final String[] jexlExpr,
+                                      final boolean expectedResult,
+                                       final Class<? extends Throwable> expectedException) {
+
+        final JexlExpressionReadTagValueFilter filter;
+
+        // test different constructors here as well
+        if ( jexlExpr.length == 1 ) {
+            filter = new JexlExpressionReadTagValueFilter(jexlExpr[0]);
+        } else {
+            filter = new JexlExpressionReadTagValueFilter(Arrays.asList(jexlExpr));
+        }
+
+        final GATKRead read = ReadTagValueFilterUnitTest.buildSAMRead(cigarString, attrsNameAndValue);
+        try {
+            Assert.assertEquals(filter.test(read), expectedResult, cigarString);
+            Assert.assertNull(expectedException);
+        } catch (Throwable e) {
+            if ( expectedException == null || !expectedException.isInstance(e) )
+                throw e;
+        }
+    }
+
+    @DataProvider(name = "JexlExpressionReadTagValueFilterDataProvider")
+    public Iterator<Object[]> jexlExpressionReadTagValueFilterDataProvider() {
+        final List<Object[]> result = new LinkedList<>();
+
+        result.add(new Object[] {
+                "100M",                             // cigar
+                new Object[] {"TM", 1.0f},          // attributes
+                new String[] {"TM == 1.0"},         // jexl expressions
+                Boolean.TRUE,                       // expected
+                null                                // expected exception
+        });
+
+        result.add(new Object[] {
+                "100M",                             // cigar
+                new Object[] {"TM", 1.0f},          // attributes
+                new String[] {"TM < 1.0"},          // jexl expressions
+                Boolean.FALSE,                      // expected
+                null                                // expected exception
+        });
+
+        result.add(new Object[] {
+                "100M",                             // cigar
+                new Object[] {"TM", 1.0f},           // attributes
+                new String[] {"NO_SUCH < 1.0"},     // jexl expressions
+                Boolean.FALSE,                      // expected
+                IllegalArgumentException.class      // expected exception
+        });
+
+        result.add(new Object[] {
+                "100M",                             // cigar
+                new Object[] {"TM", 1.0f, "TA", 2.0f},  // attributes
+                new String[] {"TM >= 1.0", "TA <= 2.0"},  // jexl expressions
+                Boolean.TRUE,                      // expected
+                null                                // expected exception
+        });
+
+        result.add(new Object[] {
+                "100M",                             // cigar
+                new Object[] {"TM", 1.0f, "TA", 2.0f},  // attributes
+                new String[] {"TM >= 1.0", "TA < 2.0"},  // jexl expressions
+                Boolean.FALSE,                      // expected
+                null                                // expected exception
+        });
+        return result.iterator();
+    }
+
+
+}