// RULE: Feature can only belong to one Protocol in a DataSet.(see issue #1136)
private void checkFeatureCanOnlyBelongToOneProtocolForOneDataSet() {
// RULE: Feature can only belong to one Protocol in a DataSet. Check it (see issue #1136)
Iterable<DataSet> dataSets = dataService.findAll(DataSet.ENTITY_NAME, DataSet.class);
for (DataSet dataSet : dataSets) {
List<Protocol> dataSetProtocols =
ProtocolUtils.getProtocolDescendants(dataSet.getProtocolUsed(), true);
for (Protocol protocol : dataSetProtocols) {
for (ObservableFeature feature : protocol.getFeatures()) {
for (Protocol p : dataSetProtocols) {
if (!p.equals(protocol) && p.getFeatures().contains(feature)) {
String message =
String.format(
"An ObservableFeature can only belong to one Protocol but feature '%s' belongs to both '%s' and '%s'",
feature.getIdentifier(), p.getIdentifier(), protocol.getIdentifier());
throw new MolgenisValidationException(
Sets.newHashSet(
new ConstraintViolation(
message, feature.getIdentifier(), feature, null, null, 0)));
}
}
}
}
}
}
@RequestMapping(method = RequestMethod.GET)
public String init(
@RequestParam(value = "selectedDataSet", required = false) String selectedDataSetId,
Model model) {
Iterable<DataSet> allDataSets = dataService.findAll(DataSet.ENTITY_NAME, new QueryImpl());
List<DataSet> dataSets = new ArrayList<DataSet>();
for (DataSet dataSet : allDataSets) {
if (selectedDataSetId == null) selectedDataSetId = dataSet.getId().toString();
if (!dataSet.getProtocolUsed().getIdentifier().equals(PROTOCOL_IDENTIFIER))
dataSets.add(dataSet);
}
model.addAttribute("dataSets", dataSets);
List<String> mappedDataSets = new ArrayList<String>();
if (selectedDataSetId != null) {
model.addAttribute("selectedDataSet", selectedDataSetId);
Iterable<DataSet> it =
dataService.findAll(
DataSet.ENTITY_NAME, new QueryImpl().like(DataSet.IDENTIFIER, selectedDataSetId));
for (DataSet dataSet : it) {
if (dataSet
.getIdentifier()
.startsWith(SecurityUtils.getCurrentUsername() + "-" + selectedDataSetId)) {
String[] dataSetIds = dataSet.getIdentifier().toString().split("-");
if (dataSetIds.length > 1) mappedDataSets.add(dataSetIds[2]);
}
}
}
model.addAttribute("mappedDataSets", mappedDataSets);
return "EvaluationPlugin";
}
public static void setUp(DataService dataService) throws Exception {
features0 = new ArrayList<ObservableFeature>();
features1 = new ArrayList<ObservableFeature>();
allEntities = new ArrayList<Entity>();
subProtocols = new ArrayList<Protocol>();
subProtocols1 = new ArrayList<Protocol>();
categories = new ArrayList<Category>();
feature0 = new ObservableFeature();
feature0.setId(0);
feature0.setName("featureName" + 0);
feature0.setIdentifier("feature" + 0);
feature1 = new ObservableFeature();
feature1.setId(1);
feature1.setName("featureName" + 1);
feature1.setIdentifier("feature" + 1);
features0.add(feature0);
features1.add(feature1);
protocol0 = new Protocol();
protocol0.setDescription("description0");
protocol0.setIdentifier("identifier0");
protocol0.setId(0);
protocol0.setFeatures(features0);
protocol1 = new Protocol();
protocol1.setDescription("description1");
protocol1.setIdentifier("identifier1");
protocol1.setId(1);
protocol1.setFeatures(features1);
subProtocols1.add(protocol1);
protocol2 = new Protocol();
protocol2.setDescription("description2");
protocol2.setIdentifier("identifier2");
protocol2.setSubprotocols(subProtocols1);
protocol2.setId(2);
protocol3 = new Protocol();
protocol3.setDescription("description3");
protocol3.setIdentifier("identifier3");
protocol3.setSubprotocols(subProtocols1);
protocol3.setId(3);
allEntities.add(protocol0);
allEntities.add(protocol1);
allEntities.add(protocol2);
allEntities.add(protocol3);
subProtocols.add(protocol0);
protocolUsed = new Protocol();
protocolUsed.setDescription("protocolUsed_description");
protocolUsed.setIdentifier("protocolUsed_identifier");
protocolUsed.setId(100);
protocolUsed.setSubprotocols(subProtocols);
allEntities.add(protocolUsed);
dataset = new DataSet();
dataset.setId(0);
dataset.setIdentifier("dataset" + 0);
dataset.setName("datasetname" + 0);
dataset.setProtocolUsed(protocolUsed);
observationSet0 = new ObservationSet();
observationSet0.setId(0);
observationSet0.setPartOfDataSet(dataset);
observationSets0 = new ArrayList<Entity>();
observationSets0.add(observationSet0);
observationSet1 = new ObservationSet();
observationSet1.setId(1);
ObservedValue observedValue0 = new ObservedValue();
observedValue0.setId(0);
observedValue0.setObservationSet(observationSet0);
observedValue0.setValue(new Value());
observedValue0.setFeature(feature0);
observedValue0.setObservationSet(observationSet0);
observedValues0 = new ArrayList<ObservedValue>();
observedValues0.add(observedValue0);
ObservedValue observedValue1 = new ObservedValue();
observedValue1.setId(1);
observedValue1.setObservationSet(observationSet1);
Value v1 = new Value();
v1.setId(1);
observedValue1.setValue(v1);
observedValue1.setFeature(feature1);
observedValue1.setObservationSet(observationSet1);
observedValues1 = new ArrayList<ObservedValue>();
observedValues1.add(observedValue1);
observationSets0 = new ArrayList<Entity>();
observationSets0.add(observationSet0);
category0 = new Category();
category0.setId(0);
category0.setIdentifier("category" + 0);
categories.add(category0);
List<DataSet> datasets = new ArrayList<DataSet>();
datasets.add(dataset);
when(dataService.findAllAsList(
DataSet.ENTITY_NAME, new QueryImpl().eq(DataSet.IDENTIFIER, "dataset1")))
.thenReturn(Arrays.<Entity>asList(dataset));
when(dataService.findAllAsList(
ObservedValue.ENTITY_NAME,
new QueryImpl().eq(ObservedValue.OBSERVATIONSET, observationSet0)))
.thenReturn(Arrays.<Entity>asList(observedValue0));
when(dataService.findAllAsList(
ObservationSet.ENTITY_NAME, new QueryImpl().eq(ObservationSet.PARTOFDATASET, 0)))
.thenReturn(Arrays.<Entity>asList(observationSet0));
}
@RequestMapping(
value = "/verify",
method = RequestMethod.POST,
headers = "Content-Type=multipart/form-data")
public void verify(
@RequestParam(value = "selectedDataSet", required = false) String selectedDataSetId,
@RequestParam Part file,
HttpServletResponse response,
Model model)
throws IOException {
EntitySource reader = null;
ExcelWriter<Entity> excelWriterRanks = null;
try {
if (selectedDataSetId != null) {
String origFileName = FileUploadUtils.getOriginalFileName(file);
File uploadFile = fileStore.store(file.getInputStream(), origFileName);
response.setContentType("application/vnd.ms-excel");
response.addHeader(
"Content-Disposition",
"attachment; filename=" + getCsvFileName(file.getName() + "-ranks"));
excelWriterRanks = new ExcelWriter<Entity>(response.getOutputStream());
excelWriterRanks.addCellProcessor(new LowerCaseProcessor(true, false));
Writable<Entity> sheetWriterRank = null;
Writable<Entity> sheetWriterRankStatistics = null;
Writable<Entity> sheetWriteBiobankRanks = null;
Writable<Entity> sheetWriteSpssInput = null;
reader = new ExcelEntitySourceFactory().create(uploadFile);
Repository<? extends Entity> inputSheet = reader.getRepositoryByEntityName("Sheet1");
List<String> biobankNames = new ArrayList<String>();
for (AttributeMetaData attr : inputSheet.getAttributes()) {
biobankNames.add(attr.getName());
}
String firstColumn = biobankNames.get(0);
biobankNames.remove(0);
// First column has to correspond to the selected dataset
DataSet ds = dataService.findOne(DataSet.ENTITY_NAME, Integer.parseInt(selectedDataSetId));
if (ds.getName().equalsIgnoreCase(firstColumn)) {
Map<String, Map<String, List<String>>> maunalMappings =
new HashMap<String, Map<String, List<String>>>();
for (Entity row : inputSheet) {
String variableName = row.getString(firstColumn);
if (!maunalMappings.containsKey(variableName))
maunalMappings.put(variableName, new HashMap<String, List<String>>());
for (String biobank : biobankNames) {
if (row.get(biobank) != null) {
String mappingString = row.get(biobank).toString();
if (!maunalMappings.containsKey(variableName)) {
maunalMappings.put(variableName, new HashMap<String, List<String>>());
}
if (!maunalMappings.get(variableName).containsKey(biobank.toLowerCase())) {
maunalMappings
.get(variableName)
.put(biobank.toLowerCase(), new ArrayList<String>());
}
maunalMappings
.get(variableName)
.get(biobank.toLowerCase())
.addAll(Arrays.asList(mappingString.split(",")));
}
}
}
List<String> lowerCaseBiobankNames = new ArrayList<String>();
for (String element : biobankNames) {
lowerCaseBiobankNames.add(element.toLowerCase());
}
List<DataSet> dataSets =
dataService.findAllAsList(
DataSet.ENTITY_NAME, new QueryImpl().in(DataSet.NAME, lowerCaseBiobankNames));
lowerCaseBiobankNames.add(0, firstColumn.toLowerCase());
sheetWriterRank = excelWriterRanks.createWritable("result", lowerCaseBiobankNames);
Map<String, Map<String, List<Integer>>> rankCollection =
new HashMap<String, Map<String, List<Integer>>>();
List<Object> allRanks = new ArrayList<Object>();
for (Entry<String, Map<String, List<String>>> entry : maunalMappings.entrySet()) {
String variableName = entry.getKey();
List<String> ranks = new ArrayList<String>();
ranks.add(variableName);
Map<String, List<String>> mappingDetail = entry.getValue();
List<ObservableFeature> features =
dataService.findAllAsList(
ObservableFeature.ENTITY_NAME,
new QueryImpl().eq(ObservableFeature.NAME, variableName));
String description = features.get(0).getDescription();
if (!rankCollection.containsKey(description))
rankCollection.put(description, new HashMap<String, List<Integer>>());
if (!features.isEmpty()) {
Entity row = new MapEntity();
row.set(firstColumn.toLowerCase(), description);
for (DataSet dataSet : dataSets) {
List<Integer> ranksBiobank = new ArrayList<Integer>();
if (mappingDetail.containsKey(dataSet.getName().toLowerCase())) {
Map<String, Hit> mappedFeatureIds =
findFeaturesFromIndex(
"name",
mappingDetail.get(dataSet.getName().toLowerCase()),
dataSet.getId());
String mappingDataSetIdentifier =
SecurityUtils.getCurrentUsername()
+ "-"
+ selectedDataSetId
+ "-"
+ dataSet.getId();
Query q =
new QueryImpl()
.eq("store_mapping_feature", features.get(0).getId())
.pageSize(50)
.sort(new Sort(Direction.DESC, "store_mapping_score"));
SearchRequest searchRequest =
new SearchRequest(mappingDataSetIdentifier, q, null);
SearchResult result = searchService.search(searchRequest);
if (mappedFeatureIds.size() == 0) {
row.set(dataSet.getName().toLowerCase(), "N/A2");
continue;
}
List<String> ids = new ArrayList<String>();
for (Hit hit : result.getSearchHits()) {
Map<String, Object> columnValueMap = hit.getColumnValueMap();
ids.add(columnValueMap.get("store_mapping_mapped_feature").toString());
}
Map<String, Hit> featureInfos = findFeaturesFromIndex("id", ids, dataSet.getId());
String previousDescription = null;
int rank = 0;
for (Hit hit : result.getSearchHits()) {
Map<String, Object> columnValueMap = hit.getColumnValueMap();
String mappedFeatureId =
columnValueMap.get("store_mapping_mapped_feature").toString();
String mappedFeatureDescription =
featureInfos
.get(mappedFeatureId)
.getColumnValueMap()
.get("description")
.toString()
.replaceAll("[^0-9a-zA-Z ]", " ");
rank++;
if (previousDescription != null
&& previousDescription.equalsIgnoreCase(mappedFeatureDescription)) rank--;
if (mappedFeatureIds.containsKey(mappedFeatureId)) {
ranksBiobank.add(rank);
allRanks.add(rank);
mappedFeatureIds.remove(mappedFeatureId);
}
previousDescription = mappedFeatureDescription;
}
if (mappedFeatureIds.size() == 0) {
String output = StringUtils.join(ranksBiobank, ',');
if (ranksBiobank.size() > 1) {
output += " (" + averageRank(ranksBiobank) + ")";
}
row.set(dataSet.getName().toLowerCase(), output);
} else {
for (int i = 0; i < mappedFeatureIds.size(); i++) allRanks.add("Not mapped");
row.set(dataSet.getName().toLowerCase(), "Not mapped");
ranksBiobank.clear();
}
} else row.set(dataSet.getName().toLowerCase(), "N/A1");
rankCollection.get(description).put(dataSet.getName().toLowerCase(), ranksBiobank);
}
sheetWriterRank.add(row);
}
}
Map<String, List<Integer>> rankCollectionPerBiobank =
new HashMap<String, List<Integer>>();
{
sheetWriterRankStatistics =
excelWriterRanks.createWritable(
"rank statistics",
Arrays.asList(
firstColumn.toLowerCase(),
"average rank",
"round-up rank",
"median rank",
"minium",
"maximum"));
for (Entry<String, Map<String, List<Integer>>> entry : rankCollection.entrySet()) {
String variableName = entry.getKey();
Entity row = new MapEntity();
row.set(firstColumn.toLowerCase(), variableName);
List<Integer> rankAllBiobanks = new ArrayList<Integer>();
for (Entry<String, List<Integer>> rankBiobanks : entry.getValue().entrySet()) {
if (!rankCollectionPerBiobank.containsKey(rankBiobanks.getKey()))
rankCollectionPerBiobank.put(rankBiobanks.getKey(), new ArrayList<Integer>());
rankCollectionPerBiobank.get(rankBiobanks.getKey()).addAll(rankBiobanks.getValue());
rankAllBiobanks.addAll(rankBiobanks.getValue());
}
row.set("average rank", averageRank(rankAllBiobanks));
row.set("round-up rank", Math.ceil(averageRank(rankAllBiobanks)));
Collections.sort(rankAllBiobanks);
if (!rankAllBiobanks.isEmpty()) {
row.set("minium", rankAllBiobanks.get(0));
row.set("maximum", rankAllBiobanks.get(rankAllBiobanks.size() - 1));
double medianRank = 0;
if (rankAllBiobanks.size() % 2 == 0) {
medianRank =
(double)
(rankAllBiobanks.get(rankAllBiobanks.size() / 2 - 1)
+ rankAllBiobanks.get(rankAllBiobanks.size() / 2))
/ 2;
} else {
medianRank = rankAllBiobanks.get(rankAllBiobanks.size() / 2);
}
row.set("median rank", medianRank);
}
sheetWriterRankStatistics.add(row);
}
}
{
lowerCaseBiobankNames.remove(0);
sheetWriteBiobankRanks =
excelWriterRanks.createWritable("biobank average ranks", lowerCaseBiobankNames);
Entity entity = new MapEntity();
for (Entry<String, List<Integer>> entry : rankCollectionPerBiobank.entrySet()) {
entity.set(entry.getKey(), averageRank(entry.getValue()));
}
sheetWriteBiobankRanks.add(entity);
}
{
sheetWriteSpssInput =
excelWriterRanks.createWritable("spss ranks", Arrays.asList("rank"));
for (Object rank : allRanks) {
Entity entity = new MapEntity("rank", rank);
sheetWriteSpssInput.add(entity);
}
}
}
}
} finally {
if (reader != null) reader.close();
if (excelWriterRanks != null) IOUtils.closeQuietly(excelWriterRanks);
}
}
@Override
@Transactional(rollbackFor = IOException.class)
public EntityImportReport doImport(
RepositoryCollection repositories, DatabaseAction databaseAction) throws IOException {
// All new repository identifiers
List<String> newRepoIdentifiers = new ArrayList<String>();
// First import entities, the data sheets are ignored in the entitiesimporter
EntityImportReport importReport = entitiesImporter.importEntities(repositories, databaseAction);
// RULE: Feature can only belong to one Protocol in a DataSet. Check it (see issue #1136)
checkFeatureCanOnlyBelongToOneProtocolForOneDataSet();
// Import data sheets
for (String name : repositories.getEntityNames()) {
Repository repository = repositories.getRepositoryByEntityName(name);
if (repository.getName().startsWith(DATASET_SHEET_PREFIX)) {
// Import DataSet sheet, create new OmxRepository
String identifier = repository.getName().substring(DATASET_SHEET_PREFIX.length());
if (!dataService.hasRepository(identifier)) {
dataService.addRepository(
new AggregateableCrudRepositorySecurityDecorator(
new OmxRepository(dataService, searchService, identifier, entityValidator)));
newRepoIdentifiers.add(identifier);
DataSet dataSet =
dataService.findOne(
DataSet.ENTITY_NAME,
new QueryImpl().eq(DataSet.IDENTIFIER, identifier),
DataSet.class);
List<Protocol> protocols =
ProtocolUtils.getProtocolDescendants(dataSet.getProtocolUsed());
List<ObservableFeature> categoricalFeatures = new ArrayList<ObservableFeature>();
for (Protocol protocol : protocols) {
List<ObservableFeature> observableFeatures = protocol.getFeatures();
if (observableFeatures != null) {
for (ObservableFeature observableFeature : observableFeatures) {
String dataType = observableFeature.getDataType();
FieldType type = MolgenisFieldTypes.getType(dataType);
if (type.getEnumType() == FieldTypeEnum.CATEGORICAL) {
categoricalFeatures.add(observableFeature);
}
}
}
}
for (ObservableFeature categoricalFeature : categoricalFeatures) {
if (!dataService.hasRepository(
OmxLookupTableEntityMetaData.createOmxLookupTableEntityMetaDataName(
categoricalFeature.getIdentifier()))) {
dataService.addRepository(
new OmxLookupTableRepository(
dataService, categoricalFeature.getIdentifier(), queryResolver));
newRepoIdentifiers.add(
OmxLookupTableEntityMetaData.createOmxLookupTableEntityMetaDataName(
categoricalFeature.getIdentifier()));
}
}
}
// Check if all column names in the excel sheet exist as attributes of the entity
Set<ConstraintViolation> violations = Sets.newLinkedHashSet();
EntityMetaData meta = dataService.getEntityMetaData(identifier);
for (AttributeMetaData attr : repository.getEntityMetaData().getAttributes()) {
if (meta.getAttribute(attr.getName()) == null) {
String message =
String.format(
"Unknown attributename '%s' for entity '%s'. Sheet: '%s'",
attr.getName(), meta.getName(), repository.getName());
violations.add(new ConstraintViolation(message, attr.getName(), null, null, meta, 0));
}
}
if (!violations.isEmpty()) {
throw new MolgenisValidationException(violations);
}
// Import data into new OmxRepository
try {
dataService.add(identifier, repository);
} catch (MolgenisValidationException e) {
// Add sheet info
for (ConstraintViolation violation : e.getViolations()) {
if (violation.getRownr() > 0) {
// Rownr +1 for header
violation.setImportInfo(
String.format(
"Sheet: '%s', row: %d", repository.getName(), violation.getRownr() + 1));
} else {
violation.setImportInfo(String.format("Sheet: '%s'", repository.getName()));
}
}
for (String newRepoIdentifier : newRepoIdentifiers) {
dataService.removeRepository(newRepoIdentifier);
}
throw e;
}
int count = (int) RepositoryUtils.count(repository);
importReport.addEntityCount(identifier, count);
importReport.addNrImported(count);
}
}
return importReport;
}
