Merge pull request ga4gh#165 from david4096/newvann

New test for effect search, update schemas

ctk-domain/pom.xml

@@ -54,6 +54,18 @@
     </dependencies>
     <build>
         <plugins>
+            <plugin>
+                <artifactId>maven-clean-plugin</artifactId>
+                <version>3.0.0</version>
+                <configuration>
+                <filesets>
+                    <fileset>
+                        <directory>src/main/assertj-assertions</directory>
+                        <followSymlinks>false</followSymlinks>
+                    </fileset>
+                </filesets>
+                </configuration>
+            </plugin>
             <plugin>
                 <groupId>org.codehaus.mojo</groupId>
                 <artifactId>build-helper-maven-plugin</artifactId>

ctk-schemas/src/main/resources/avro/alleleAnnotationmethods.avdl

@@ -76,9 +76,13 @@ record SearchVariantAnnotationsRequest {
 //  union { null, array<string> } featureIds = null;
 
   /**
-  Only return variant annotations including these effects (SO terms). Exact
-  matching across all fields of the OntologyTerm is required. If null, return
-  all variant annotations.
+  This filter allows variant, transcript combinations to be extracted by effect
+  type(s).
+  Only return variant annotations including any of these effects and only return
+  transcript effects including any of these effects. Exact matching across all
+  fields of the Sequence Ontology OntologyTerm is required.
+  (A transcript effect may have multiple SO effects which will all be reported.)
+  If null, return all variant annotations.
   */
   union { null, array<org.ga4gh.models.OntologyTerm> } effects = null;
 

ctk-schemas/src/main/resources/avro/metadata.avdl

@@ -12,22 +12,36 @@ import idl "common.avdl";
   An ontology term describing an attribute. (e.g. the phenotype attribute
   'polydactyly' from HPO)
   */
-  record OntologyTerm {
-    /**
-    The source of the onotology term.
-    (e.g. `Ontology for Biomedical Investigation`)
-    */
-    string ontologySource;
-
-    /**
-    The ID defined by the external onotology source.
-    (e.g. `http://purl.obolibrary.org/obo/OBI_0001271`)
-    */
-    string id;
-
-    /** The name of the onotology term. (e.g. `RNA-seq assay`) */
-    union { null, string } name = null;
-  }
+
+record OntologyTerm {
+  /**
+  Ontology source identifier - the identifier, a CURIE (preferred) or
+  PURL for an ontology source e.g. http://purl.obolibrary.org/obo/hp.obo
+  It differs from the standard GA4GH schema's :ref:`id <apidesign_object_ids>`
+  in that it is a URI pointing to an information resource outside of the scope
+  of the schema or its resource implementation.
+  */
+  string id;
+
+  /**
+  Ontology term - the representation the id is pointing to.
+  */
+  union { null, string } term = null;
+
+  /**
+  Ontology source name - the name of ontology from which the term is obtained
+  e.g. 'Human Phenotype Ontology'
+  */
+  union { null, string } sourceName = null;
+
+  /**
+  Ontology source version - the version of the ontology from which the
+  OntologyTerm is obtained; e.g. 2.6.1.
+  There is no standard for ontology versioning and some frequently
+  released ontologies may use a datestamp, or build number.
+  */
+  union { null, string } sourceVersion = null;
+}
 
 /**
 An experimental preparation of a sample.
@@ -141,13 +155,15 @@ An analysis contains an interpretation of one or several experiments.
 information about the methodology used.
 */
 record Analysis {
-  /** The analysis UUID. This is globally unique. */
+
+  /**
+  Formats of id | name | description | accessions are described in the
+  documentation on general attributes and formats.
+  */
   string id;
 
-  /** The name of the analysis. */
   union { null, string } name = null;
 
-  /** A description of the analysis. */
   union { null, string } description = null;
 
   /**

cts-java/src/test/java/org/ga4gh/cts/api/TestData.java

@@ -57,6 +57,13 @@ private TestData() {
      */
     public static final String VARIANT_ANNOTATION_REFERENCE_NAME = "1";
 
+
+    /**
+     * The names of the variant annotation sets used for variant annotation in the standard test data.
+     */
+    public static final List<String> VARIANT_ANNOTATION_SET_NAMES =
+            Arrays.asList("WASH7P_annotation", "OR4F_annotation");
+
     /**
      * The names of known-good read groups.
      */
@@ -134,6 +141,12 @@ private TestData() {
      */
     public static final String ALIGNED_SEQUENCE_CONTENTS_PATTERN = "[ACTGN]+";
 
+
+    /**
+     * The call set name is used to test searching by call set name functionality.
+     */
+    public static final String CALL_SET_NAME = "HG00096";
+
     /**
      * Return the ID of the compliance dataset on the server being tested.
      * By default this is the value of {@link #DEFAULT_DATASET_ID}, but

cts-java/src/test/java/org/ga4gh/cts/api/Utils.java

@@ -571,4 +571,25 @@ public static String getVariantAnnotationSetId(Client client) throws AvroRemoteE
         final List<VariantAnnotationSet> variantAnnotationSets = getAllVariantAnnotationSets(client);
         return variantAnnotationSets.get(0).getId();
     }
+
+
+    /**
+     * Given a name return the variant annotation set corresponding to that name. When that name
+     * is not found returns the first annotation set found.
+     * @param client the connection to the server
+     * @param name the string name of the annotation set
+     * @return a {@link VariantAnnotationSet} with the requested name
+     * @throws AvroRemoteException if the server throws an exception or there's an I/O error
+     */
+    public static VariantAnnotationSet getVariantAnnotationSetByName(Client client, String name) throws AvroRemoteException {
+
+        // get all compliance variant annotation sets
+        final List<VariantAnnotationSet> variantAnnotationSets = getAllVariantAnnotationSets(client);
+        for (VariantAnnotationSet vas : variantAnnotationSets) {
+            if (vas.getName().equals(name)) {
+                return vas;
+            }
+        }
+        return variantAnnotationSets.get(0);
+    }
 }

cts-java/src/test/java/org/ga4gh/cts/api/reads/ReadsSearchIT.java

@@ -38,34 +38,23 @@ public class ReadsSearchIT implements CtkLogs {
      * @throws AvroRemoteException if there's a communication problem or server exception ({@link GAException})
      */
     @Test
-    public void searchReadsWithUninterestingRangeProducesZeroReads() throws AvroRemoteException {
-
+    public void searchRangeWithNoReadsReturnsZeroResults() throws AvroRemoteException {
         final String refId = Utils.getValidReferenceId(client);
 
-        final long emptyRangeStart = 150;
-        final long emptyRangeEnd = 160;
-
-        final SearchReadGroupSetsRequest req =
-                SearchReadGroupSetsRequest.newBuilder()
-                                          .setDatasetId(TestData.getDatasetId())
-                                          .build();
-        final SearchReadGroupSetsResponse resp = client.reads.searchReadGroupSets(req);
-
-        final List<ReadGroupSet> readGroupSets = resp.getReadGroupSets();
+        final long emptyRangeStart = 0; // is this range actually empty?
+        final long emptyRangeEnd = 100;
 
         final SearchReadsRequest srReq =
                 SearchReadsRequest.newBuilder()
-                                  .setReferenceId(refId)
-                                  .setReadGroupIds(aSingle(Utils.getReadGroupId(client)))
-                                  .setStart(emptyRangeStart)
-                                  .setEnd(emptyRangeEnd)
-                                  .build();
+                        .setReferenceId(refId)
+                        .setReadGroupIds(aSingle(Utils.getReadGroupId(client)))
+                        .setStart(emptyRangeStart)
+                        .setEnd(emptyRangeEnd)
+                        .build();
         final SearchReadsResponse srResp = client.reads.searchReads(srReq);
 
         final List<ReadAlignment> alignments = srResp.getAlignments();
-        assertThat(alignments).isNotEmpty();
-
-        assertThat(alignments).doesNotContain(Utils.nullReadAlignment);
+        assertThat(alignments).isEmpty();
     }
 
     /**

cts-java/src/test/java/org/ga4gh/cts/api/variantAnnotation/VariantAnnotationSetsSearchIT.java

@@ -91,7 +91,7 @@ public void checkVariantAnnotationAnalysis() throws AvroRemoteException {
         // Check the Analysis record within the VariantAnnotationSet matches the test data
         final String name        = "compliance1";
         final String description = "variant annotation test data";
-        final String created     = "2015-11-18";
+        final String created     = "2015-11-18T00:00:00.000000Z";
         final String software    = "SnpEff";
 
         assertThat(vSet.getAnalysis().getName()).isEqualTo(name);

cts-java/src/test/java/org/ga4gh/cts/api/variantAnnotation/VariantAnnotationsPagingIT.java

@@ -39,7 +39,8 @@ public void checkPagingOneByOneThroughVariantAnnotations() throws AvroRemoteExce
         final long end = 11008;
 
         // Find a compliance variantAnnotationSet id.
-        final String variantAnnotationSetId = Utils.getVariantAnnotationSetId(client);
+
+        final String variantAnnotationSetId = Utils.getVariantAnnotationSetByName(client, TestData.VARIANT_ANNOTATION_SET_NAMES.get(0)).getId();
         final List<VariantAnnotation> listOfVariantAnnotations = Utils.getAllVariantAnnotationsInRange(client, variantAnnotationSetId, start, end);
 
         // We will remove VariantAnnotations from this Set and assert at the end that we have zero

cts-java/src/test/java/org/ga4gh/cts/api/variantAnnotation/VariantAnnotationsSearchIT.java

@@ -8,19 +8,12 @@
 import org.ga4gh.cts.api.TestData;
 import org.ga4gh.cts.api.Utils;
 import org.ga4gh.methods.*;
-import org.ga4gh.models.Variant;
-import org.ga4gh.models.VariantSet;
-import org.ga4gh.models.VariantAnnotationSet;
-import org.ga4gh.models.VariantAnnotation;
-import org.ga4gh.models.TranscriptEffect;
-import org.ga4gh.models.HGVSAnnotation;
-import org.ga4gh.models.AlleleLocation;
-import org.ga4gh.models.AnalysisResult;
-import org.ga4gh.models.Impact;
+import org.ga4gh.models.*;
 import org.junit.Test;
 import org.junit.experimental.categories.Category;
 
 import java.net.HttpURLConnection;
+import java.util.Arrays;
 import java.util.List;
 import java.util.function.Consumer;
 
@@ -72,7 +65,8 @@ private void checkAllTranscriptEffects(List<VariantAnnotation> variantAnnotation
     public void checkSearchingVariantAnnotations() throws AvroRemoteException {
 
         // Obtain a VariantAnnotationSet from the compliance dataset.
-        final String variantAnnotationSetId = Utils.getVariantAnnotationSetId(client);
+        final String variantAnnotationSetId =
+                Utils.getVariantAnnotationSetByName(client, TestData.VARIANT_ANNOTATION_SET_NAMES.get(0)).getId();
 
 
         // Seek variant annotation records for the extracted VariantAnnotationSet.
@@ -186,7 +180,8 @@ public void checkTranscriptEffects() throws AvroRemoteException {
     public void checkSearchingSingleVariantAnnotation() throws AvroRemoteException {
 
         // Obtain a VariantAnnotationSet from the compliance dataset.
-        final String variantAnnotationSetId = Utils.getVariantAnnotationSetId(client);
+        final String variantAnnotationSetId =
+                Utils.getVariantAnnotationSetByName(client, TestData.VARIANT_ANNOTATION_SET_NAMES.get(1)).getId();
         final long single_start = 69540;
         final long single_end = 69541;
 
@@ -235,7 +230,8 @@ public void checkSearchingSingleVariantAnnotation() throws AvroRemoteException {
     public void checkVariantAnnotationVariants() throws AvroRemoteException {
 
         // Obtain a VariantAnnotationSet from the compliance dataset.
-        final String variantAnnotationSetId = Utils.getVariantAnnotationSetId(client);
+        final String variantAnnotationSetId =
+                Utils.getVariantAnnotationSetByName(client, TestData.VARIANT_ANNOTATION_SET_NAMES.get(0)).getId();
 
         // Seek variant annotation records for the extracted VariantAnnotationSet.
         final SearchVariantAnnotationsRequest req =
@@ -247,7 +243,7 @@ public void checkVariantAnnotationVariants() throws AvroRemoteException {
                                                .build();
 
         final SearchVariantAnnotationsResponse resp = client.variantAnnotations.searchVariantAnnotations(req);
-
+        assertThat(resp.getVariantAnnotations()).isNotEmpty();
         //Check a valid variant id is returned for each record.
         for( VariantAnnotation variantAnn : resp.getVariantAnnotations() ){
             final String  variantId = variantAnn.getVariantId();
@@ -256,6 +252,40 @@ public void checkVariantAnnotationVariants() throws AvroRemoteException {
             Variant v = client.variants.getVariant(variantId);
             assertThat(v).isNotNull();
             assertThat(v.getId()).isEqualTo(variantId);
+            assertThat(v.getStart()).isGreaterThanOrEqualTo(start);
+        }
+    }
+
+    /**
+     * Check whether filtering by effect ID finds the expected annotations.
+     *
+     *@throws AvroRemoteException if there's a communication problem or server exception ({@link GAException})
+     */
+    @Test
+    public void checkFilteringByEffectTerm() throws AvroRemoteException {
+
+        final OntologyTerm term = new OntologyTerm("SO:0001819", "synonymous_variant", "source", "0");
+        // Obtain a VariantAnnotationSet from the compliance dataset.
+        final String variantAnnotationSetId =
+                Utils.getVariantAnnotationSetByName(client, TestData.VARIANT_ANNOTATION_SET_NAMES.get(1)).getId();
+
+        // Seek variant annotation records for the extracted VariantAnnotationSet.
+        final SearchVariantAnnotationsRequest req =
+                SearchVariantAnnotationsRequest.newBuilder()
+                        .setVariantAnnotationSetId(variantAnnotationSetId)
+                        .setReferenceName(TestData.VARIANT_ANNOTATION_REFERENCE_NAME)
+                        .setStart(0)
+                        .setEnd(100000)
+                        .setEffects(Arrays.asList(term))
+                        .build();
+
+        final SearchVariantAnnotationsResponse resp = client.variantAnnotations.searchVariantAnnotations(req);
+
+        // There are 7 records with the "synonymous_variant" term in the test data.
+        assertThat(resp.getVariantAnnotations()).hasSize(7);
+        //Check the transcript effect field is not empty for each annotation.
+        for( VariantAnnotation variantAnn : resp.getVariantAnnotations() ){
+            assertThat(variantAnn.getTranscriptEffects()).isNotEmpty();
         }
     }