HTSJDK SamReader implementation using the API and example of its usage

Added ant build and updated HTSJDK

.gitignore

@@ -1,4 +1,7 @@
 bin
 .classpath
 .project
-
+.jar_tmp
+classes
+dist
+gatk-tools-java.version.properties

build.xml

@@ -0,0 +1,145 @@
+<?xml version="1.0"?>
+<project name="gatk-tools-java" basedir="." default="all">
+
+    <property name="src" value="src/main/java"/>
+    <property name="lib" value="lib"/>
+    <property name="dist" value="dist"/>
+    <property name="classes" value="classes"/>
+    <property name="jar_tmp" value=".jar_tmp"/>
+   
+    <property name="javac.target" value="1.6"/>
+    <property name="javac.debug" value="true"/>
+
+    <!-- Get GIT hash, if available, otherwise leave it blank.  -->
+    <exec executable="git" outputproperty="repository.revision" failifexecutionfails="true" errorproperty="">
+        <arg value="log"/>
+        <arg value="-1"/>
+        <arg value="--pretty=format:%H_%at"/>
+    </exec>
+    <property name="repository.revision" value=""/>
+    <property name="gatk-tools-java-version" value="1.0"/>
+    <property name="gatk-tools-java-version-file" value="gatk-tools-java.version.properties"/>
+   
+    <!-- VERSION PROPERTY --> 
+    <target name="write-version-property">
+        <propertyfile
+            file="${gatk-tools-java-version-file}"
+            comment="gatk-tools-java version">
+            <entry  key="gatk-tools-java-version" value="${gatk-tools-java-version}"/>
+        </propertyfile>
+    </target>
+
+    <!-- INIT -->
+    <target name="init" depends="write-version-property">
+        <path id="classpath">
+            <fileset dir="${lib}">
+                <include name="**/*.jar"/>
+            </fileset>
+        </path>
+    </target>
+
+    <!-- CLEAN -->
+    <target name="clean">
+        <delete dir="${classes}"/>
+        <delete dir="${dist}"/>
+        <delete dir="javadoc"/>
+        <delete dir="${jar_tmp}"/>
+    	  <delete file="${gatk-tools-java-version-file}"/> 	
+    </target>
+
+    <!-- COMPILE -->
+    <target name="compile" depends="init"
+            description="Compile files without cleaning">
+    	<compile-src includes="**/*.*"/>
+    </target>
+
+   
+    <target name="gatk-tools-java-jar" depends="compile"
+            description="Builds gatk-tools-java-${gatk-tools-java-version}.jar for inclusion in other projects">
+        <mkdir dir="${dist}"/>
+        <delete dir="${jar_tmp}"/>
+        <mkdir dir="${jar_tmp}"/>
+        <copy todir="${jar_tmp}">
+            <fileset dir="${classes}" includes="**/*"/>
+        </copy>
+        <unzip dest="${jar_tmp}">
+            <fileset dir="${lib}">
+                <include name="*.jar"/>
+            </fileset>
+        </unzip>
+        <jar destfile="${dist}/gatk-tools-java-${gatk-tools-java-version}.jar" compress="no">
+        	 <fileset dir="${jar_tmp}" includes="**/*.*"/>
+            <manifest>
+            	  <attribute name="Implementation-Title" value="@gatk-tools-java"/>
+                <attribute name="Implementation-Version" value="${gatk-tools-java-version}(${repository.revision})"/>
+            </manifest>
+        </jar>
+        <copy todir="${dist}">
+            <fileset dir="lib" includes="*.jar"/>
+        </copy>
+    </target>
+
+ 
+
+    <!-- ALL -->
+    <target name="all" depends="compile, gatk-tools-java-jar" description="Default build target">
+    </target>
+
+    <!-- ************************************************************************************** -->
+    <!-- ************************************************************************************** -->
+    <!-- Beginning of taskdefs that are used elsewhere in the build file                        -->
+    <!-- ************************************************************************************** -->
+    <!-- ************************************************************************************** -->
+
+    <!-- Compile source files specified by includes, from source root. Can specifically
+    include or exclude-->
+    <macrodef name="compile-src">
+        <attribute name="includes" default=""/>
+        <attribute name="excludes" default=""/>
+        <attribute name="destdir" default="${classes}"/>
+        <attribute name="compile.classpath" default="classpath"/>
+        <attribute name="compiler.args" default=""/>
+        <sequential>
+        <mkdir dir="${classes}"/>
+            <!-- unset the sourcepath attribute in order to compile only files explicitly specified and disable javac's default searching mechanism -->
+            <javac destdir="@{destdir}"
+                   optimize="${javac.opt}"
+                   debug="${javac.debug}"
+                   sourcepath=""
+                   srcdir="${src}"
+                   includes="@{includes}"
+                   excludes="@{excludes}"
+                   source="${javac.target}"
+                   target="${javac.target}"
+            	     includeantruntime="false">
+                <classpath refid="@{compile.classpath}"/>
+                <compilerarg line="@{compiler.args}" />
+            </javac>
+        </sequential>
+    </macrodef>
+
+    <macrodef name="compile-tests">
+        <attribute name="includes" default=""/>
+        <attribute name="excludes" default=""/>
+        <attribute name="compiler.args" default=""/>
+
+        <sequential>
+            <mkdir dir="${classes.test}"/>
+            <javac destdir="${classes.test}"
+                   optimize="${javac.opt}"
+                   debug="${javac.debug}"
+                   srcdir="${src.test.java}"
+                   includes="@{includes}"
+                   excludes="@{excludes}"
+                   source="${javac.target}"
+                   target="${javac.target}"
+                   includeantruntime="false">
+                <classpath>
+                    <path refid="classpath"/>
+                    <pathelement location="${classes}"/>
+                </classpath>
+                <compilerarg line="@{compiler.args}"/>
+            </javac>
+        </sequential>
+    </macrodef>
+</project>

src/main/java/com/google/cloud/genomics/gatk/common/GA4GHUrl.java

@@ -16,11 +16,12 @@
 package com.google.cloud.genomics.gatk.common;
 
 import java.net.URISyntaxException;
+import java.net.URL;
 
 /**
  * Represents a GA4GH reads resource as a URL in the form of
  * ga4gh://<base api path>/readsets/<readset>/<sequence>/[start-end],
- * e.g. ga4gh://www.googleapis.com/genomics/v1beta/reads/CLqN8Z3sDRDQldHJ_rTS9VE/1/
+ * e.g. ga4gh://www.googleapis.com/genomics/v1beta/readsets/CLqN8Z3sDRDQldHJ_rTS9VE/1/
  */
 public class GA4GHUrl {
   int rangeStart = 0;
@@ -53,6 +54,10 @@ public GA4GHUrl(String rootUrl,
     this.rangeEnd = rangeEnd;
   }
   
+  public GA4GHUrl(URL input) throws URISyntaxException {
+   this(input.toString().replace("https://", GA4GH_SCHEMA_PREFIX));
+  }
+  
   public GA4GHUrl(String input) throws URISyntaxException {
     if (!isGA4GHUrl(input)) {
       throw new URISyntaxException(input, "Schema is not ga4gh");
@@ -67,7 +72,7 @@ public GA4GHUrl(String input) throws URISyntaxException {
     String[] pathComponents = readsPath.split("/");
     if (pathComponents.length < 4) {
       throw new URISyntaxException(input,
-          "Expecting " + READS_PATH_COMPONENT +"readset/sequence/[range], got "
+          "Expecting " + READS_PATH_COMPONENT + "readset/sequence/[range], got "
               + readsPath);
     }
   

src/main/java/com/google/cloud/genomics/gatk/common/GenomicsApiDataSource.java

@@ -15,10 +15,15 @@
 */
 package com.google.cloud.genomics.gatk.common;
 
+import com.google.api.client.auth.oauth2.Credential;
 import com.google.api.client.extensions.java6.auth.oauth2.VerificationCodeReceiver;
 import com.google.api.client.extensions.jetty.auth.oauth2.LocalServerReceiver;
 import com.google.api.client.googleapis.extensions.java6.auth.oauth2.GooglePromptReceiver;
+import com.google.api.client.googleapis.javanet.GoogleNetHttpTransport;
 import com.google.api.client.googleapis.json.GoogleJsonResponseException;
+import com.google.api.client.http.HttpRequest;
+import com.google.api.client.http.HttpRequestInitializer;
+import com.google.api.client.json.jackson2.JacksonFactory;
 import com.google.api.services.genomics.Genomics;
 import com.google.api.services.genomics.GenomicsScopes;
 import com.google.api.services.genomics.model.Dataset;
@@ -70,39 +75,58 @@ private Genomics getApi() throws GeneralSecurityException, IOException {
   }
   
   private Genomics initGenomicsApi() throws GeneralSecurityException, IOException {
-    File clientSecrets = new File(clientSecretsFilename);
-    if (!clientSecrets.exists()) {
-      throw new IOException(
-          "Client secrets file " + clientSecretsFilename + " does not exist."
-          + " Visit https://developers.google.com/genomics to learn how"
-          + " to install a client_secrets.json file.  If you have installed a client_secrets.json"
-          + " in a specific location, use --client_secrets_filename <path>/client_secrets.json.");
+    LOG.info("Initializing Genomics API for " + rootUrl);
+    if (!clientSecretsFilename.isEmpty()) {
+      File clientSecrets = new File(clientSecretsFilename);
+      if (!clientSecrets.exists()) {
+        throw new IOException(
+            "Client secrets file " + clientSecretsFilename + " does not exist."
+            + " Visit https://developers.google.com/genomics to learn how"
+            + " to install a client_secrets.json file.  If you have installed a client_secrets.json"
+            + " in a specific location, use --client_secrets_filename <path>/client_secrets.json.");
+      }
+      LOG.info("Using client secrets file " + clientSecretsFilename);
+      
+      VerificationCodeReceiver receiver = noLocalServer ? 
+          new GooglePromptReceiver() : new LocalServerReceiver();
+      GenomicsFactory genomicsFactory = GenomicsFactory
+              .builder("genomics_java_client")
+              .setScopes(SCOPES)
+              .setRootUrl(rootUrl)
+              .setServicePath("/")
+              .setVerificationCodeReceiver(Suppliers.ofInstance(receiver))
+              .build();
+      return genomicsFactory.fromClientSecretsFile(clientSecrets);
+    } else {
+      final Genomics.Builder builder = new Genomics
+          .Builder(
+              GoogleNetHttpTransport.newTrustedTransport(),
+              JacksonFactory.getDefaultInstance(),
+              new HttpRequestInitializer() {
+                @Override public void initialize(HttpRequest httpRequest) throws IOException {
+                  httpRequest.setReadTimeout(20000);
+                  httpRequest.setConnectTimeout(20000);
+                }
+              })
+          .setApplicationName("genomics_java_client")
+          .setRootUrl(rootUrl)
+          .setServicePath("/");
+        return builder.build();
     }
-    
-    VerificationCodeReceiver receiver = noLocalServer ? 
-        new GooglePromptReceiver() : new LocalServerReceiver();
-
-    GenomicsFactory genomicsFactory = GenomicsFactory
-        .builder("genomics_java_client")
-        .setScopes(SCOPES)
-        .setUserName("user" + SCOPES.size())
-        .setVerificationCodeReceiver(Suppliers.ofInstance(receiver))
-        .setRootUrl(rootUrl)
-        .setServicePath("/")
-        .build();
-
-    return genomicsFactory.fromClientSecretsFile(clientSecrets);
   }
   
   public ReadIteratorResource getReadsFromGenomicsApi(GA4GHUrl url) 
        throws IOException, GeneralSecurityException {
+    LOG.info("Getting reads from " + url);
     return getReadsFromGenomicsApi(url.getReadset(), 
         url.getSequence(), url.getRangeStart(), url.getRangeEnd());
   }
   
   public ReadIteratorResource getReadsFromGenomicsApi(String readsetId, 
       String sequenceName, int sequenceStart, int sequenceEnd) 
           throws IOException, GeneralSecurityException {
+    LOG.info("Getting readset " + readsetId + ", sequence " + sequenceName + 
+        ", start=" + sequenceStart + ", end=" + sequenceEnd);
     final Genomics stub = getApi();
     // TODO(iliat): implement API retries and using access key for public
     // datasets

src/main/java/com/google/cloud/genomics/gatk/htsjdk/GA4GHQueryInterval.java

@@ -0,0 +1,82 @@
+package com.google.cloud.genomics.gatk.htsjdk;
+
+import htsjdk.samtools.util.CoordMath;
+import htsjdk.samtools.SAMRecord;
+
+/**
+ * Similar to HTSJDK's QueryInterval but allows specifying sequence name 
+ * (as opposed to index in the header) and adds ability to check if a given read 
+ * matches the interval.
+ */
+public class GA4GHQueryInterval {
+  private String sequence;
+  private int start;
+  private int end;
+   
+  public enum ReadPositionConstraint {
+    OVERLAPPING,
+    CONTAINED,
+    START_AT
+  }
+  private ReadPositionConstraint readPositionConstraint;
+
+  public GA4GHQueryInterval(String sequence, int start, int end,
+      ReadPositionConstraint readPositionConstraint) {
+    super();
+    this.sequence = sequence;
+    this.start = start;
+    this.end = end;
+    this.readPositionConstraint = readPositionConstraint;
+  }
+  
+  public String getSequence() {
+    return sequence;
+  }
+  
+  public void setSequence(String sequence) {
+    this.sequence = sequence;
+  }
+  
+  public int getStart() {
+    return start;
+  }
+  
+  public void setStart(int start) {
+    this.start = start;
+  }
+  
+  public int getEnd() {
+    return end;
+  }
+  
+  public void setEnd(int end) {
+    this.end = end;
+  }
+  
+  public ReadPositionConstraint getReadPositionConstraint() {
+    return readPositionConstraint;
+  }
+
+  public void setReadPositionConstraint(ReadPositionConstraint readPositionConstraint) {
+    this.readPositionConstraint = readPositionConstraint;
+  }
+  
+  /** 
+   * Returns true iff the read specified by the record matches the interval
+   * given the interval's constraints and the read position.
+   */
+  public boolean matches(SAMRecord record) {
+    int myEnd = end == 0 ? Integer.MAX_VALUE : end;
+    switch (readPositionConstraint) {
+      case OVERLAPPING:
+        return CoordMath.overlaps(start, myEnd, 
+            record.getAlignmentStart(), record.getAlignmentEnd());
+      case CONTAINED:
+        return CoordMath.encloses(start, myEnd,
+            record.getAlignmentStart(), record.getAlignmentEnd());
+      case START_AT:
+        return start == record.getAlignmentStart();
+    }
+    return false;
+  }
+}

src/main/java/com/google/cloud/genomics/gatk/htsjdk/GA4GHReaderFactory.java

@@ -0,0 +1,25 @@
+package com.google.cloud.genomics.gatk.htsjdk;
+
+import htsjdk.samtools.CustomReaderFactory;
+import htsjdk.samtools.SamReader;
+
+import java.net.URL;
+import java.util.logging.Logger;
+/**
+ * HTSJDK CustomReaderFactory implementation.
+ * Returns a SamReader that reads data from GA4GH API.
+ */
+public class GA4GHReaderFactory implements CustomReaderFactory.ICustomReaderFactory {
+  private static final Logger LOG = Logger.getLogger(GA4GHReaderFactory.class.getName());
+  
+  @Override
+  public SamReader open(URL url) {
+    try {
+      return new GA4GHSamReader(url);
+    } catch (Exception ex) {
+      LOG.warning("Error creating SamReader " + ex.toString());
+      return null;
+    }
+  }
+
+}