Skip to content

Commit

Permalink
Merge pull request #452 from geneontology/dev
Browse files Browse the repository at this point in the history 
Bring master up to date with dev
  • Loading branch information
@balhoff
balhoff authored Jan 28, 2022
2 parents 33c0583 + 8893160 commit efd9247
Show file tree
Hide file tree
Showing 18 changed files with 1,749 additions and 189 deletions.
29 changes: 19 additions & 10 deletions minerva-cli/src/main/java/org/geneontology/minerva/cli/CommandLineInterface.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -315,14 +315,23 @@ public static void main(String[] args) {
gpad_options.addOption("cat", "catalog", true, "Catalog file for tbox ontology. " +
"Use this to specify local copies of the ontology and or its imports to " +
"speed and control the process. If not used, will download the tbox and all its imports.");
gpad_options.addOption("ontojournal", "ontojournal", true, "Specify a blazegraph journal file containing the merged, pre-reasoned tbox aka go-lego.owl");
cmd = parser.parse(gpad_options, args, false);
String inputDB = cmd.getOptionValue("input");
String gpadOutputFolder = cmd.getOptionValue("gpad-output");
String modelIdPrefix = cmd.getOptionValue("model-id-prefix");
String modelIdcurie = cmd.getOptionValue("model-id-curie");
String ontologyIRI = cmd.getOptionValue("ontology");
String catalog = cmd.getOptionValue("catalog");
legoToAnnotationsSPARQL(modelIdPrefix, modelIdcurie, inputDB, gpadOutputFolder, ontologyIRI, catalog);
String go_lego_journal_file = null;
if(cmd.hasOption("ontojournal")) {
go_lego_journal_file = cmd.getOptionValue("ontojournal");
}
if(go_lego_journal_file==null) {
System.err.println("Missing -- ontojournal . Need to specify blazegraph journal file containing the merged go-lego tbox (neo, GO-plus, etc..)");
System.exit(-1);
}
legoToAnnotationsSPARQL(modelIdPrefix, modelIdcurie, inputDB, gpadOutputFolder, ontologyIRI, catalog, go_lego_journal_file);
}else if(cmd.hasOption("version")) {
printVersion();
}else if(cmd.hasOption("validate-go-cams")) {
Expand Down Expand Up @@ -711,7 +720,7 @@ public static void owl2LegoJson(String input, String output, boolean usePretty)
* @param ontologyIRI
* @throws Exception
*/
public static void legoToAnnotationsSPARQL(String modelIdPrefix, String modelIdcurie, String inputDB, String gpadOutputFolder, String ontologyIRI, String catalog) throws Exception {
public static void legoToAnnotationsSPARQL(String modelIdPrefix, String modelIdcurie, String inputDB, String gpadOutputFolder, String ontologyIRI, String catalog, String go_lego_journal_file) throws Exception {
if(modelIdPrefix==null) {
modelIdPrefix = "http://model.geneontology.org/";
}
Expand All @@ -738,13 +747,13 @@ public static void legoToAnnotationsSPARQL(String modelIdPrefix, String modelIdc
OWLOntology ontology = ontman.loadOntology(IRI.create(ontologyIRI));
CurieMappings localMappings = new CurieMappings.SimpleCurieMappings(Collections.singletonMap(modelIdcurie, modelIdPrefix));
CurieHandler curieHandler = new MappedCurieHandler(DefaultCurieHandler.loadDefaultMappings(), localMappings);
BlazegraphMolecularModelManager<Void> m3 = new BlazegraphMolecularModelManager<>(ontology, curieHandler, modelIdPrefix, inputDB, null, null);
BlazegraphMolecularModelManager<Void> m3 = new BlazegraphMolecularModelManager<>(ontology, curieHandler, modelIdPrefix, inputDB, null, go_lego_journal_file);
final String immutableModelIdPrefix = modelIdPrefix;
final String immutableGpadOutputFolder = gpadOutputFolder;
m3.getAvailableModelIds().stream().parallel().forEach(modelIRI -> {
try {
//TODO investigate whether changing to a neo-lite model has an impact on this - may need to make use of ontology journal
String gpad = new GPADSPARQLExport(curieHandler, m3.getLegacyRelationShorthandIndex(), m3.getTboxShorthandIndex()).exportGPAD(m3.createInferredModel(modelIRI), modelIRI);
String gpad = new GPADSPARQLExport(curieHandler, m3.getLegacyRelationShorthandIndex(), m3.getTboxShorthandIndex(), m3.getGolego_repo().regulatorsToRegulated).exportGPAD(m3.createInferredModel(modelIRI), modelIRI);
String fileName = StringUtils.replaceOnce(modelIRI.toString(), immutableModelIdPrefix, "") + ".gpad";
Writer writer = new OutputStreamWriter(new FileOutputStream(Paths.get(immutableGpadOutputFolder, fileName).toFile()), StandardCharsets.UTF_8);
writer.write(gpad);
Expand Down Expand Up @@ -920,13 +929,13 @@ else if(!model_iris.add(modeluri)) {
// TODO Auto-generated catch block
e1.printStackTrace();
}

if(checkShex) {
if(checkShex) {
shex.setActive(true);
}else {
shex.setActive(false);
}
shex.setActive(true);
}else {
shex.setActive(false);
}

//shex validator is ready, now build the inference provider (which provides access to the shex validator and provides inferences useful for shex)
String reasonerOpt = "arachne";
LOGGER.info("Building OWL inference provider: "+reasonerOpt);
Expand Down Expand Up @@ -1038,7 +1047,7 @@ else if(!model_iris.add(modeluri)) {
int n_rows_gpad = 0;
if(isConsistent) {
try {
Set<GPADData> gpad = new GPADSPARQLExport(curieHandler, m3.getLegacyRelationShorthandIndex(), m3.getTboxShorthandIndex()).getGPAD(m3.createInferredModel(modelIRI), modelIRI);
Set<GPADData> gpad = new GPADSPARQLExport(curieHandler, m3.getLegacyRelationShorthandIndex(), m3.getTboxShorthandIndex(), m3.getGolego_repo().regulatorsToRegulated).getGPAD(m3.createInferredModel(modelIRI), modelIRI);
if(gpad!=null) {
n_rows_gpad = gpad.size();
}
Expand Down
26 changes: 19 additions & 7 deletions minerva-converter/src/main/java/org/geneontology/minerva/legacy/sparql/GPADSPARQLExport.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,9 @@

import scala.collection.JavaConverters;

import static java.util.stream.Collectors.mapping;
import static java.util.stream.Collectors.toSet;

/* Note: the example GPAD files are available at this link: http://www.informatics.jax.org/downloads/reports/mgi.gpa.gz */
public class GPADSPARQLExport {
private static final Logger LOG = Logger.getLogger(GPADSPARQLExport.class);
Expand All @@ -54,6 +57,9 @@ public class GPADSPARQLExport {
private static final String BP = "http://purl.obolibrary.org/obo/GO_0008150";
private static final String CC = "http://purl.obolibrary.org/obo/GO_0005575";
private static final Set<String> rootTerms = new HashSet<>(Arrays.asList(MF, BP, CC));
private static final String ENABLES = "http://purl.obolibrary.org/obo/RO_0002327";
private static final String CONTRIBUTES_TO = "http://purl.obolibrary.org/obo/RO_0002326";
private static final Set<String> functionRelations = new HashSet<>(Arrays.asList(ENABLES, CONTRIBUTES_TO));
private static final String EMAPA_NAMESPACE = "http://purl.obolibrary.org/obo/EMAPA_";
private static final String UBERON_NAMESPACE = "http://purl.obolibrary.org/obo/UBERON_";
private static final String inconsistentQuery =
Expand Down Expand Up @@ -96,11 +102,13 @@ public class GPADSPARQLExport {
private final CurieHandler curieHandler;
private final Map<IRI, String> relationShorthandIndex;
private final Map<IRI, String> tboxShorthandIndex;
private final Map<IRI, Set<IRI>> regulators;

public GPADSPARQLExport(CurieHandler handler, Map<IRI, String> shorthandIndex, Map<IRI, String> tboxShorthandIndex) {
public GPADSPARQLExport(CurieHandler handler, Map<IRI, String> shorthandIndex, Map<IRI, String> tboxShorthandIndex, Map<IRI, Set<IRI>> regulators) {
this.curieHandler = handler;
this.relationShorthandIndex = shorthandIndex;
this.tboxShorthandIndex = tboxShorthandIndex;
this.regulators = regulators;
}

public String exportGPAD(WorkingMemory wm, IRI modelIRI) throws InconsistentOntologyException {
Expand Down Expand Up @@ -147,11 +155,15 @@ public Set<GPADData> getGPAD(WorkingMemory wm, IRI modelIRI) throws Inconsistent
possibleExtensions.forEach(ae -> statementsToExplain.add(ae.getTriple()));
Map<Triple, Set<Explanation>> allExplanations = statementsToExplain.stream().collect(Collectors.toMap(Function.identity(), s -> toJava(wm.explain(Bridge.tripleFromJena(s)))));

Map<Triple, Set<GPADEvidence>> allEvidences = evidencesForFacts(allExplanations.values().stream().flatMap(es -> es.stream()).flatMap(e -> toJava(e.facts()).stream().map(t -> Bridge.jenaFromTriple(t))).collect(Collectors.toSet()), model, modelID, modelLevelAnnotations);
Set<Node> gpNodesWithOtherThanRootMF = basicAnnotations.stream().filter(a -> !a.getOntologyClass().toString().equals(MF)).map(a -> a.getObjectNode()).collect(Collectors.toSet());
Map<Triple, Set<GPADEvidence>> allEvidences = evidencesForFacts(allExplanations.values().stream().flatMap(es -> es.stream()).flatMap(e -> toJava(e.facts()).stream().map(t -> Bridge.jenaFromTriple(t))).collect(toSet()), model, modelID, modelLevelAnnotations);
Set<IRI> gpsWithAnyMFNotRootMF = basicAnnotations.stream().filter(a -> functionRelations.contains(a.getQualifier().toString())).filter(a -> !a.getOntologyClass().toString().equals(MF)).map(a -> a.getObject()).collect(toSet());
Map<Node, Set<IRI>> nodesToOntologyClasses = basicAnnotations.stream().collect(Collectors.groupingBy(BasicGPADData::getObjectNode, mapping(BasicGPADData::getOntologyClass, toSet())));
for (BasicGPADData annotation : basicAnnotations) {
Set<IRI> termsRegulatedByAnnotationsForThisGPNode = nodesToOntologyClasses.get(annotation.getObjectNode()).stream().flatMap(term -> regulators.getOrDefault(term, Collections.emptySet()).stream()).collect(toSet());
boolean regulationViolation = termsRegulatedByAnnotationsForThisGPNode.contains(annotation.getOntologyClass());
if (regulationViolation) continue;
for (Explanation explanation : allExplanations.get(Triple.create(annotation.getObjectNode(), NodeFactory.createURI(annotation.getQualifier().toString()), annotation.getOntologyClassNode()))) {
Set<Triple> requiredFacts = toJava(explanation.facts()).stream().map(t -> Bridge.jenaFromTriple(t)).collect(Collectors.toSet());
Set<Triple> requiredFacts = toJava(explanation.facts()).stream().map(t -> Bridge.jenaFromTriple(t)).collect(toSet());
// Every statement in the explanation must have at least one evidence, unless the statement is a class assertion
if (requiredFacts.stream().filter(t -> !t.getPredicate().getURI().equals(RDF.type.getURI())).allMatch(f -> !(allEvidences.get(f).isEmpty()))) {
// The evidence used for the annotation must be on an edge to or from the target node
Expand Down Expand Up @@ -179,8 +191,8 @@ public Set<GPADData> getGPAD(WorkingMemory wm, IRI modelIRI) throws Inconsistent
if (rootTerms.contains(annotation.getOntologyClass().toString())) {
rootViolation = !ND.equals(currentEvidence.getEvidence().toString());
} else { rootViolation = false; }
final boolean rootMFWithBP = annotation.getOntologyClass().toString().equals(MF) && gpNodesWithOtherThanRootMF.contains(annotation.getObjectNode());
if (!rootViolation && !rootMFWithBP) {
final boolean rootMFWithOtherMF = annotation.getOntologyClass().toString().equals(MF) && gpsWithAnyMFNotRootMF.contains(annotation.getObject());
if (!rootViolation && !rootMFWithOtherMF) {
DefaultGPADData defaultGPADData = new DefaultGPADData(annotation.getObject(), annotation.getQualifier(), annotation.getOntologyClass(), goodExtensions,
reference, currentEvidence.getEvidence(), currentEvidence.getWithOrFrom(), Optional.empty(), currentEvidence.getModificationDate(),
currentEvidence.getAssignedBy(), currentEvidence.getAnnotations());
Expand Down Expand Up @@ -250,7 +262,7 @@ private Map<Triple, Set<GPADEvidence>> evidencesForFacts(Set<Triple> facts, Mode
Optional<String> with = Optional.ofNullable(eqs.getLiteral("with")).map(Literal::getLexicalForm);
Set<Pair<String, String>> annotationAnnotations = new HashSet<>();
annotationAnnotations.add(Pair.of("noctua-model-id", modelID));
annotationAnnotations.addAll(getContributors(eqs).stream().map(c -> Pair.of("contributor", c)).collect(Collectors.toSet()));
annotationAnnotations.addAll(getContributors(eqs).stream().map(c -> Pair.of("contributor", c)).collect(toSet()));
String modificationDate = eqs.getLiteral("modification_date").getLexicalForm();
Optional<String> creationDate = Optional.ofNullable(eqs.getLiteral("creation_date")).map(Literal::getLexicalForm);
creationDate.ifPresent(date -> annotationAnnotations.add(Pair.of("creation-date", date)));
Expand Down
1 change: 0 additions & 1 deletion ...va-converter/src/main/resources/org/geneontology/minerva/legacy/sparql/gpad-extensions.rq
View file
Original file line number Diff line number Diff line change
Expand Up @@ -37,7 +37,6 @@ VALUES ?extension_rel {
<http://purl.obolibrary.org/obo/RO_0002490>
<http://purl.obolibrary.org/obo/RO_0002491>
<http://purl.obolibrary.org/obo/RO_0000053>
<http://purl.obolibrary.org/obo/RO_0000057>
<http://purl.obolibrary.org/obo/RO_0000087>
<http://purl.obolibrary.org/obo/RO_0001025>
<http://purl.obolibrary.org/obo/RO_0002007>
Expand Down
2 changes: 1 addition & 1 deletion .../main/resources/org/geneontology/minerva/legacy/sparql/gpad-relation-evidence-multiple.rq
View file
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@ FILTER(?evidence_type NOT IN (rdfs:Resource, owl:Thing, owl:NamedIndividual, BFO
?evidence dc:contributor ?contributor .
?evidence dc:date ?modification_date .
?evidence dc:source ?source_pub .
OPTIONAL { ?axiom dct:created ?creation_date }
OPTIONAL { ?evidence dct:created ?creation_date }
OPTIONAL { ?axiom dct:dateAccepted ?import_date }
OPTIONAL { ?evidence pav:providedBy ?provided_by . }
OPTIONAL { ?evidence <http://geneontology.org/lego/evidence-with> ?with }
Expand Down
46 changes: 36 additions & 10 deletions minerva-converter/src/test/java/org/geneontology/minerva/legacy/sparql/GPADSPARQLTest.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -24,10 +24,7 @@
import scala.collection.JavaConverters;

import java.io.File;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Set;
import java.util.*;
import java.util.stream.Collectors;

public class GPADSPARQLTest {
Expand All @@ -47,7 +44,7 @@ public static void setupRules() throws OWLOntologyCreationException {
@BeforeClass
public static void setupExporter() {
JenaSystem.init();
exporter = new GPADSPARQLExport(DefaultCurieHandler.getDefaultHandler(), new HashMap<IRI, String>(), new HashMap<IRI, String>());
exporter = new GPADSPARQLExport(DefaultCurieHandler.getDefaultHandler(), new HashMap<IRI, String>(), new HashMap<IRI, String>(), new HashMap<>());
}

@Test
Expand Down Expand Up @@ -124,21 +121,23 @@ public void testGPADContainsAcceptedAndCreatedDates() throws Exception {
Set<Triple> triples = model.listStatements().toList().stream().map(s -> Bridge.tripleFromJena(s.asTriple())).collect(Collectors.toSet());
WorkingMemory mem = arachne.processTriples(JavaConverters.asScalaSetConverter(triples).asScala());
Set<GPADData> annotations = exporter.getGPAD(mem, IRI.create("http://test.org"));
IRI gene = IRI.create("http://identifiers.org/mgi/MGI:1922815");
Pair<String, String> creationDate = Pair.of("creation-date", "2012-09-17");
IRI gene = IRI.create("http://identifiers.org/wormbase/WBGene00001326");
Pair<String, String> creationDate = Pair.of("creation-date", "2021-05-13");
Assert.assertTrue(annotations.stream().anyMatch(a -> a.getObject().equals(gene) && a.getAnnotations().contains(creationDate)));
}

@Test
public void testFilterRootMFWhenRootBP() throws Exception {
public void testFilterRootMFWhenOtherMF() throws Exception {
IRI rootMF = IRI.create("http://purl.obolibrary.org/obo/GO_0003674");
IRI rootBP = IRI.create("http://purl.obolibrary.org/obo/GO_0008150");
IRI rootCC = IRI.create("http://purl.obolibrary.org/obo/GO_0005575");

Model model = ModelFactory.createDefaultModel();
model.read(this.getClass().getResourceAsStream("/test_root_mf_filter.ttl"), "", "ttl");
Set<Triple> triples = model.listStatements().toList().stream().map(s -> Bridge.tripleFromJena(s.asTriple())).collect(Collectors.toSet());
WorkingMemory mem = arachne.processTriples(JavaConverters.asScalaSetConverter(triples).asScala());
Set<GPADData> annotations = exporter.getGPAD(mem, IRI.create("http://test.org"));
IRI gene = IRI.create("http://identifiers.org/mgi/MGI:2153470");
IRI rootMF = IRI.create("http://purl.obolibrary.org/obo/GO_0003674");
IRI rootBP = IRI.create("http://purl.obolibrary.org/obo/GO_0008150");
Assert.assertTrue(annotations.stream().noneMatch(a -> a.getObject().equals(gene) && a.getOntologyClass().equals(rootMF)));

Model model2 = ModelFactory.createDefaultModel();
Expand All @@ -149,6 +148,33 @@ public void testFilterRootMFWhenRootBP() throws Exception {
IRI gene2 = IRI.create("http://identifiers.org/mgi/MGI:98392");
Assert.assertTrue(annotations2.stream().anyMatch(a -> a.getObject().equals(gene2) && a.getOntologyClass().equals(rootMF)));
Assert.assertTrue(annotations2.stream().anyMatch(a -> a.getObject().equals(gene2) && a.getOntologyClass().equals(rootBP)));

Model model3 = ModelFactory.createDefaultModel();
model3.read(this.getClass().getResourceAsStream("/test_root_mf_filter3.ttl"), "", "ttl");
Set<Triple> triples3 = model3.listStatements().toList().stream().map(s -> Bridge.tripleFromJena(s.asTriple())).collect(Collectors.toSet());
WorkingMemory mem3 = arachne.processTriples(JavaConverters.asScalaSetConverter(triples3).asScala());
Set<GPADData> annotations3 = exporter.getGPAD(mem3, IRI.create("http://test.org"));
IRI gene3 = IRI.create("http://identifiers.org/sgd/S000002650");
Assert.assertTrue(annotations3.stream().anyMatch(a -> a.getObject().equals(gene3) && a.getOntologyClass().equals(rootMF)));
Assert.assertTrue(annotations3.stream().anyMatch(a -> a.getObject().equals(gene3) && a.getOntologyClass().equals(rootBP)));
Assert.assertTrue(annotations3.stream().anyMatch(a -> a.getObject().equals(gene3) && a.getOntologyClass().equals(rootCC)));
}

@Test
public void testFilterAnnotationsToRegulatedProcess() throws Exception {
HashMap<IRI, Set<IRI>> regulators = new HashMap<>();
regulators.put(IRI.create("http://purl.obolibrary.org/obo/GO_0030511"), Collections.singleton(IRI.create("http://purl.obolibrary.org/obo/GO_0007179")));
GPADSPARQLExport exporter = new GPADSPARQLExport(DefaultCurieHandler.getDefaultHandler(), new HashMap<IRI, String>(), new HashMap<IRI, String>(), regulators);
Model model = ModelFactory.createDefaultModel();
model.read(this.getClass().getResourceAsStream("/test_filter_regulated_process.ttl"), "", "ttl");
Set<Triple> triples = model.listStatements().toList().stream().map(s -> Bridge.tripleFromJena(s.asTriple())).collect(Collectors.toSet());
WorkingMemory mem = arachne.processTriples(JavaConverters.asScalaSetConverter(triples).asScala());
Set<GPADData> annotations = exporter.getGPAD(mem, IRI.create("http://test.org"));
IRI gene = IRI.create("http://identifiers.org/mgi/MGI:2148811");
IRI regulator = IRI.create("http://purl.obolibrary.org/obo/GO_0030511");
IRI regulated = IRI.create("http://purl.obolibrary.org/obo/GO_0007179");
Assert.assertTrue(annotations.stream().anyMatch(a -> a.getObject().equals(gene) && a.getOntologyClass().equals(regulator)));
Assert.assertTrue(annotations.stream().noneMatch(a -> a.getObject().equals(gene) && a.getOntologyClass().equals(regulated)));
}

}
Loading

0 comments on commit efd9247

Please sign in to comment.