Genomic and Generic Assay bin counts endpoints (cBioPortal#10964)

* Genomic and Generic Assay bin counts endpoints
dippindots · Sep 11, 2024 · 10c4a87 · 10c4a87
1 parent e2d502a
commit 10c4a87
Show file tree

Hide file tree

Showing 23 changed files with 1,298 additions and 79 deletions.
diff --git a/src/main/java/org/cbioportal/persistence/StudyViewRepository.java b/src/main/java/org/cbioportal/persistence/StudyViewRepository.java
@@ -9,6 +9,7 @@
 import org.cbioportal.model.CopyNumberCountByGene;
 import org.cbioportal.model.GenomicDataCountItem;
 import org.cbioportal.model.GenomicDataCount;
+import org.cbioportal.model.MolecularProfile;
 import org.cbioportal.model.PatientTreatment;
 import org.cbioportal.model.Sample;
 import org.cbioportal.model.SampleTreatment;
@@ -67,4 +68,10 @@ public interface StudyViewRepository {
     Map<String, Integer> getMutationCounts(StudyViewFilterContext studyViewFilterContext, GenomicDataFilter genomicDataFilter);
 
     List<GenomicDataCountItem> getMutationCountsByType(StudyViewFilterContext studyViewFilterContext, List<GenomicDataFilter> genomicDataFilters);
+
+    List<ClinicalDataCount> getGenomicDataBinCounts(StudyViewFilterContext studyViewFilterContext, List<String> filteredAttributes);
+
+    List<ClinicalDataCount> getGenericAssayDataBinCounts(StudyViewFilterContext studyViewFilterContext, List<String> filteredAttributes);
+
+    List<MolecularProfile> getGenericAssayProfiles();
 }
diff --git a/...ce/enums/ClinicalAttributeDataSource.java → ...oportal/persistence/enums/DataSource.java b/...ce/enums/ClinicalAttributeDataSource.java → ...oportal/persistence/enums/DataSource.java
@@ -1,11 +1,11 @@
 package org.cbioportal.persistence.enums;
 
-public enum ClinicalAttributeDataSource {
+public enum DataSource {
     PATIENT("PATIENT"),SAMPLE("SAMPLE");
 
     private final String value;
 
-    ClinicalAttributeDataSource(String value) {
+    DataSource(String value) {
         this.value = value;
     }
 

diff --git a/src/main/java/org/cbioportal/persistence/helper/StudyViewFilterHelper.java b/src/main/java/org/cbioportal/persistence/helper/StudyViewFilterHelper.java
@@ -1,44 +1,109 @@
 package org.cbioportal.persistence.helper;
 
+import org.cbioportal.model.MolecularProfile;
+import org.cbioportal.persistence.enums.DataSource;
 import org.cbioportal.web.parameter.ClinicalDataFilter;
+import org.cbioportal.web.parameter.CategorizedGenericAssayDataCountFilter;
+import org.cbioportal.web.parameter.CategorizedGenomicDataCountFilter;
 import org.cbioportal.web.parameter.CustomSampleIdentifier;
 import org.cbioportal.web.parameter.StudyViewFilter;
 import org.springframework.lang.NonNull;
 import org.springframework.lang.Nullable;
 
 import java.util.ArrayList;
+import java.util.EnumMap;
 import java.util.List;
+import java.util.Map;
 import java.util.Objects;
 
 public final class StudyViewFilterHelper {
     public static StudyViewFilterHelper build(@Nullable StudyViewFilter studyViewFilter,
-            @Nullable List<CustomSampleIdentifier> customDataSamples) {
+                                              @Nullable Map<DataSource, List<MolecularProfile>> genericAssayProfilesMap,
+                                              @Nullable List<CustomSampleIdentifier> customDataSamples) {
         if (Objects.isNull(studyViewFilter)) {
             studyViewFilter = new StudyViewFilter();
         }
+        if (Objects.isNull(genericAssayProfilesMap)) {
+            genericAssayProfilesMap = new EnumMap<>(DataSource.class);
+        }
         if (Objects.isNull(customDataSamples)) {
             customDataSamples = new ArrayList<>();
         }
-        return new StudyViewFilterHelper(studyViewFilter, customDataSamples);
+        return new StudyViewFilterHelper(studyViewFilter, genericAssayProfilesMap, customDataSamples);
     }
 
     private final StudyViewFilter studyViewFilter;
+    private final CategorizedGenomicDataCountFilter categorizedGenomicDataCountFilter;
+    private final CategorizedGenericAssayDataCountFilter categorizedGenericAssayDataCountFilter;
     private final List<CustomSampleIdentifier> customDataSamples;
 
-    private StudyViewFilterHelper(@NonNull StudyViewFilter studyViewFilter,
-            @NonNull List<CustomSampleIdentifier> customDataSamples) {
+    private StudyViewFilterHelper(@NonNull StudyViewFilter studyViewFilter, 
+                                  @NonNull Map<DataSource, List<MolecularProfile>> genericAssayProfilesMap,
+                                  @NonNull List<CustomSampleIdentifier> customDataSamples) {
         this.studyViewFilter = studyViewFilter;
-
+        this.categorizedGenomicDataCountFilter = extractGenomicDataCountFilters(studyViewFilter);
+        this.categorizedGenericAssayDataCountFilter = extractGenericAssayDataCountFilters(studyViewFilter, genericAssayProfilesMap);
         this.customDataSamples = customDataSamples;
     }
 
     public StudyViewFilter studyViewFilter() {
         return studyViewFilter;
     }
+
+    public CategorizedGenomicDataCountFilter categorizedGenomicDataCountFilter() {
+        return categorizedGenomicDataCountFilter;
+    }
 
+    public CategorizedGenericAssayDataCountFilter categorizedGenericAssayDataCountFilter() {
+        return categorizedGenericAssayDataCountFilter;
+    }
+
     public List<CustomSampleIdentifier> customDataSamples() {
         return this.customDataSamples;
     }
+
+    private CategorizedGenomicDataCountFilter extractGenomicDataCountFilters(final StudyViewFilter studyViewFilter) {
+        if (studyViewFilter.getGenomicDataFilters() == null)
+        {
+            return CategorizedGenomicDataCountFilter.getBuilder().build();
+        }
+
+        CategorizedGenomicDataCountFilter.Builder builder = CategorizedGenomicDataCountFilter.getBuilder();
+
+        builder.setSampleNumericalGenomicDataFilters(studyViewFilter.getGenomicDataFilters().stream()
+            .filter(genomicDataFilter -> !genomicDataFilter.getProfileType().equals("cna") && !genomicDataFilter.getProfileType().equals("gistic"))
+            .toList());
+        builder.setSampleCategoricalGenomicDataFilters(studyViewFilter.getGenomicDataFilters().stream()
+            .filter(genomicDataFilter -> genomicDataFilter.getProfileType().equals("cna") || genomicDataFilter.getProfileType().equals("gistic"))
+            .toList());
+        return builder.build();
+    }
+
+    private CategorizedGenericAssayDataCountFilter extractGenericAssayDataCountFilters(final StudyViewFilter studyViewFilter, Map<DataSource, List<MolecularProfile>> genericAssayProfilesMap) {
+        if ((studyViewFilter.getGenericAssayDataFilters() == null || genericAssayProfilesMap.isEmpty()))
+        {
+            return CategorizedGenericAssayDataCountFilter.getBuilder().build();
+        }
+
+        CategorizedGenericAssayDataCountFilter.Builder builder = CategorizedGenericAssayDataCountFilter.getBuilder();
+
+        // TODO: Support patient level profiles and data filtering
+        List<String> sampleCategoricalProfileTypes = genericAssayProfilesMap.get(DataSource.SAMPLE)
+            .stream().filter(profile -> profile.getDatatype().equals("CATEGORICAL") || profile.getDatatype().equals("BINARY"))
+            .map(profile -> profile.getStableId().replace(profile.getCancerStudyIdentifier() + "_", ""))
+            .toList();
+        List<String> sampleNumericalProfileTypes = genericAssayProfilesMap.get(DataSource.SAMPLE)
+            .stream().filter(profile -> profile.getDatatype().equals("LIMIT-VALUE"))
+            .map(profile -> profile.getStableId().replace(profile.getCancerStudyIdentifier() + "_", ""))
+            .toList();
+        builder.setSampleNumericalGenericAssayDataFilters(studyViewFilter.getGenericAssayDataFilters().stream()
+            .filter(genericAssayDataFilter -> sampleNumericalProfileTypes.contains(genericAssayDataFilter.getProfileType()))
+            .toList());
+        builder.setSampleCategoricalGenericAssayDataFilters(studyViewFilter.getGenericAssayDataFilters().stream()
+            .filter(genericAssayDataFilter -> sampleCategoricalProfileTypes.contains(genericAssayDataFilter.getProfileType()))
+            .toList());
+        return builder.build();
+    }
 
     public boolean isCategoricalClinicalDataFilter(ClinicalDataFilter clinicalDataFilter) {
         var filterValue = clinicalDataFilter.getValues().getFirst();

diff --git a/src/main/java/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.java b/src/main/java/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.java
@@ -12,6 +12,7 @@
 import org.cbioportal.model.GenomicDataCount;
 import org.cbioportal.model.GenomicDataCountItem;
 import org.cbioportal.model.PatientTreatment;
+import org.cbioportal.model.MolecularProfile;
 import org.cbioportal.model.Sample;
 import org.cbioportal.model.SampleTreatment;
 import org.cbioportal.persistence.helper.AlterationFilterHelper;
@@ -65,4 +66,9 @@ public interface StudyViewMapper {
     Map<String, Integer> getMutationCounts(StudyViewFilterHelper studyViewFilterHelper, GenomicDataFilter genomicDataFilter);
 
     List<GenomicDataCountItem> getMutationCountsByType(StudyViewFilterHelper studyViewFilterHelper, List<GenomicDataFilter> genomicDataFilters);
+
+    List<ClinicalDataCount> getGenomicDataBinCounts(StudyViewFilterHelper studyViewFilterHelper, List<String> attributeIds);
+    List<ClinicalDataCount> getGenericAssayDataBinCounts(StudyViewFilterHelper studyViewFilterHelper, List<String> attributeIds);
+
+    List<MolecularProfile> getGenericAssayProfiles();
 }
diff --git a/src/main/java/org/cbioportal/persistence/mybatisclickhouse/StudyViewMyBatisRepository.java b/src/main/java/org/cbioportal/persistence/mybatisclickhouse/StudyViewMyBatisRepository.java
@@ -9,12 +9,13 @@
 import org.cbioportal.model.GenomicDataCountItem;
 import org.cbioportal.model.GenomicDataCount;
 import org.cbioportal.model.CopyNumberCountByGene;
+import org.cbioportal.model.MolecularProfile;
 import org.cbioportal.model.PatientTreatment;
 import org.cbioportal.model.Sample;
 import org.cbioportal.model.SampleTreatment;
 import org.cbioportal.model.StudyViewFilterContext;
 import org.cbioportal.persistence.StudyViewRepository;
-import org.cbioportal.persistence.enums.ClinicalAttributeDataSource;
+import org.cbioportal.persistence.enums.DataSource;
 import org.cbioportal.persistence.helper.AlterationFilterHelper;
 import org.cbioportal.persistence.helper.StudyViewFilterHelper;
 import org.cbioportal.web.parameter.ClinicalDataType;
@@ -24,6 +25,7 @@
 
 import java.util.ArrayList;
 import java.util.Collections;
+import java.util.EnumMap;
 import java.util.HashMap;
 import java.util.List;
 import java.util.Map;
@@ -33,9 +35,9 @@
 @Repository
 public class StudyViewMyBatisRepository implements StudyViewRepository {
 
-    private Map<ClinicalAttributeDataSource, List<ClinicalAttribute>> clinicalAttributesMap = new HashMap<>();
-
-
+    private Map<DataSource, List<ClinicalAttribute>> clinicalAttributesMap = new EnumMap<>(DataSource.class);
+    private Map<DataSource, List<MolecularProfile>> genericAssayProfilesMap = new EnumMap<>(DataSource.class);
+    
     private static final List<String> FILTERED_CLINICAL_ATTR_VALUES = Collections.emptyList();
     private final StudyViewMapper mapper;
 
@@ -101,14 +103,19 @@ public List<GenomicDataCount> getMolecularProfileSampleCounts(StudyViewFilterCon
     }
 
     public StudyViewFilterHelper createStudyViewFilterHelper(StudyViewFilterContext studyViewFilterContext) {
-        return StudyViewFilterHelper.build(studyViewFilterContext.studyViewFilter(), studyViewFilterContext.customDataFilterSamples());    
+        return StudyViewFilterHelper.build(studyViewFilterContext.studyViewFilter() , getGenericAssayProfilesMap(), studyViewFilterContext.customDataFilterSamples());    
     }
 
     @Override
     public List<ClinicalAttribute> getClinicalAttributes() {
         return mapper.getClinicalAttributes();
     }
 
+    @Override
+    public List<MolecularProfile> getGenericAssayProfiles() {
+        return mapper.getGenericAssayProfiles();
+    }
+
     @Override
     public Map<String, ClinicalDataType> getClinicalAttributeDatatypeMap() {
         if (clinicalAttributesMap.isEmpty()) {
@@ -118,11 +125,11 @@ public Map<String, ClinicalDataType> getClinicalAttributeDatatypeMap() {
         Map<String, ClinicalDataType> attributeDatatypeMap = new HashMap<>();
 
         clinicalAttributesMap
-            .get(ClinicalAttributeDataSource.SAMPLE)
+            .get(DataSource.SAMPLE)
             .forEach(attribute -> attributeDatatypeMap.put(attribute.getAttrId(), ClinicalDataType.SAMPLE));
 
         clinicalAttributesMap
-            .get(ClinicalAttributeDataSource.PATIENT)
+            .get(DataSource.PATIENT)
             .forEach(attribute -> attributeDatatypeMap.put(attribute.getAttrId(), ClinicalDataType.PATIENT));
 
         return attributeDatatypeMap;
@@ -199,18 +206,34 @@ public List<SampleTreatment> getSampleTreatments(StudyViewFilterContext studyVie
     public int getTotalSampleTreatmentCount(StudyViewFilterContext studyViewFilterContext) {
         return mapper.getTotalSampleTreatmentCounts(createStudyViewFilterHelper(studyViewFilterContext));
     }
+
+    @Override
+    public List<ClinicalDataCount> getGenomicDataBinCounts(StudyViewFilterContext studyViewFilterContext, List<String> attributeIds) {
+        return mapper.getGenomicDataBinCounts(createStudyViewFilterHelper(studyViewFilterContext), attributeIds);
+    }
 
+    @Override
+    public List<ClinicalDataCount> getGenericAssayDataBinCounts(StudyViewFilterContext studyViewFilterContext, List<String> attributeIds) {
+        return mapper.getGenericAssayDataBinCounts(createStudyViewFilterHelper(studyViewFilterContext), attributeIds);
+    }
+
     private void buildClinicalAttributeNameMap() {
         clinicalAttributesMap = this.getClinicalAttributes()
             .stream()
-            .collect(Collectors.groupingBy(ca -> ca.getPatientAttribute() ? ClinicalAttributeDataSource.PATIENT : ClinicalAttributeDataSource.SAMPLE));
+            .collect(Collectors.groupingBy(ca -> ca.getPatientAttribute().booleanValue() ? DataSource.PATIENT : DataSource.SAMPLE));
     }
-
-    private Map<ClinicalAttributeDataSource, List<ClinicalAttribute>> getClinicalAttributeNameMap() {
-        if (clinicalAttributesMap.isEmpty()) {
-            buildClinicalAttributeNameMap();
+
+    private void buildGenericAssayProfilesMap() {
+        genericAssayProfilesMap = this.getGenericAssayProfiles()
+            .stream()
+            .collect(Collectors.groupingBy(ca -> ca.getPatientLevel().booleanValue() ? DataSource.PATIENT : DataSource.SAMPLE));
+    }
+
+    private Map<DataSource, List<MolecularProfile>> getGenericAssayProfilesMap() {
+        if (genericAssayProfilesMap.isEmpty()) {
+            buildGenericAssayProfilesMap();
         }
-        return clinicalAttributesMap;
+        return genericAssayProfilesMap;
     }
 
     @Override

diff --git a/src/main/java/org/cbioportal/service/StudyViewColumnarService.java b/src/main/java/org/cbioportal/service/StudyViewColumnarService.java
@@ -45,6 +45,10 @@ public interface StudyViewColumnarService {
     List<GenomicDataCountItem> getCNACountsByGeneSpecific(StudyViewFilter studyViewFilter, List<GenomicDataFilter> genomicDataFilters);
 
     List<GenomicDataCountItem> getMutationCountsByGeneSpecific(StudyViewFilter studyViewFilter, List<GenomicDataFilter> genomicDataFilters);
+
+    List<ClinicalDataCountItem> getGenomicDataBinCounts(StudyViewFilter studyViewFilter, List<String> filteredAttributes);
+
+    List<ClinicalDataCountItem> getGenericAssayDataBinCounts(StudyViewFilter studyViewFilter, List<String> filteredAttributes);
 
     List<GenomicDataCountItem> getMutationTypeCountsByGeneSpecific(StudyViewFilter studyViewFilter, List<GenomicDataFilter> genomicDataFilters);
 }
diff --git a/src/main/java/org/cbioportal/service/impl/StudyViewColumnarServiceImpl.java b/src/main/java/org/cbioportal/service/impl/StudyViewColumnarServiceImpl.java
@@ -81,6 +81,16 @@ public SampleTreatmentReport getSampleTreatmentReport(StudyViewFilter studyViewF
         return treatmentCountReportService.getSampleTreatmentReport(createContext(studyViewFilter));
     }
 
+    @Override
+    public List<ClinicalDataCountItem> getGenomicDataBinCounts(StudyViewFilter studyViewFilter, List<String> filteredAttributes) {
+        return generateDataCountItemsFromDataCounts(studyViewRepository.getGenomicDataBinCounts(createContext(studyViewFilter), filteredAttributes));
+    }
+
+    @Override
+    public List<ClinicalDataCountItem> getGenericAssayDataBinCounts(StudyViewFilter studyViewFilter, List<String> filteredAttributes) {
+        return generateDataCountItemsFromDataCounts(studyViewRepository.getGenericAssayDataBinCounts(createContext(studyViewFilter), filteredAttributes));
+    }
+
     public List<CopyNumberCountByGene> getCnaGenes(StudyViewFilter studyViewFilter) {
         return alterationCountService.getCnaGenes(createContext(studyViewFilter));
     }
@@ -97,14 +107,7 @@ public Map<String, ClinicalDataType> getClinicalAttributeDatatypeMap() {
 
     @Override
     public List<ClinicalDataCountItem> getClinicalDataCounts(StudyViewFilter studyViewFilter, List<String> filteredAttributes) {
-        return studyViewRepository.getClinicalDataCounts(createContext(studyViewFilter), filteredAttributes)
-            .stream().collect(Collectors.groupingBy(ClinicalDataCount::getAttributeId))
-            .entrySet().parallelStream().map(e -> {
-                ClinicalDataCountItem item = new ClinicalDataCountItem();
-                item.setAttributeId(e.getKey());
-                item.setCounts(e.getValue());
-                return item;
-            }).collect(Collectors.toList());
+        return generateDataCountItemsFromDataCounts(studyViewRepository.getClinicalDataCounts(createContext(studyViewFilter), filteredAttributes));
     }
 
     @Override
@@ -156,8 +159,16 @@ private StudyViewFilterContext createContext(StudyViewFilter studyViewFilter) {
         List<CustomSampleIdentifier> customSampleIdentifiers = customDataFilterUtil.extractCustomDataSamples(studyViewFilter);
         return new StudyViewFilterContext(studyViewFilter, customSampleIdentifiers);
     }
-
-
+
+    private List<ClinicalDataCountItem> generateDataCountItemsFromDataCounts(List<ClinicalDataCount> dataCounts) {
+            return dataCounts.stream().collect(Collectors.groupingBy(ClinicalDataCount::getAttributeId))
+            .entrySet().parallelStream().map(e -> {
+                ClinicalDataCountItem item = new ClinicalDataCountItem();
+                item.setAttributeId(e.getKey());
+                item.setCounts(e.getValue());
+                return item;
+            }).toList();
+    }
 
     public static List<CaseListDataCount> mergeCaseListCounts(List<CaseListDataCount> counts) {
         Map<String, List<CaseListDataCount>> countsPerListType = counts.stream()