Examples with GeoDomainObjectGeneratorLocal ubic.gemma.core.loader.expression.geo.GeoDomainObjectGeneratorLocal used on opensource projects

Example 21 with GeoDomainObjectGeneratorLocal

use of ubic.gemma.core.loader.expression.geo.GeoDomainObjectGeneratorLocal in project Gemma by PavlidisLab.

the class ProcessedExpressionDataVectorServiceTest method getDataset.

private Collection<ExpressionExperiment> getDataset() throws Exception {
    // Dataset uses spotted arrays, 11 samples.
    ExpressionExperiment ee;
    try {
        geoService.setGeoDomainObjectGenerator(new GeoDomainObjectGeneratorLocal(this.getTestFileBasePath("gse432Short")));
        // noinspection unchecked
        Collection<ExpressionExperiment> results = (Collection<ExpressionExperiment>) geoService.fetchAndLoad("GSE432", false, true, false);
        ee = results.iterator().next();
        TwoChannelMissingValues tcmv = this.getBean(TwoChannelMissingValues.class);
        tcmv.computeMissingValues(ee, 1.5, null);
    // No masked preferred computation.
    } catch (AlreadyExistsInSystemException e) {
        if (e.getData() instanceof List) {
            ee = (ExpressionExperiment) ((List<?>) e.getData()).iterator().next();
        } else {
            ee = (ExpressionExperiment) e.getData();
        }
    } catch (Exception e) {
        if (e.getCause() instanceof IOException && e.getCause().getMessage().contains("502")) {
            return null;
        }
        throw e;
    }
    ee.setShortName(RandomStringUtils.randomAlphabetic(12));
    expressionExperimentService.update(ee);
    ee = expressionExperimentService.thawLite(ee);
    processedDataVectorService.createProcessedDataVectors(ee);
    Collection<ExpressionExperiment> e = new HashSet<>();
    e.add(ee);
    return e;
}

Example 22 with GeoDomainObjectGeneratorLocal

use of ubic.gemma.core.loader.expression.geo.GeoDomainObjectGeneratorLocal in project Gemma by PavlidisLab.

the class SubsettedAnalysis3Test method setup.

@Before
public void setup() throws Exception {
    geoService.setGeoDomainObjectGenerator(new GeoDomainObjectGeneratorLocal(FileTools.resourceToPath("/data/analysis/expression/gse26927short")));
    try {
        Collection<?> results = geoService.fetchAndLoad("GSE26927", false, true, false);
        ee = (ExpressionExperiment) results.iterator().next();
    } catch (AlreadyExistsInSystemException e) {
        ee = (ExpressionExperiment) ((Collection<?>) e.getData()).iterator().next();
    }
    ee = expressionExperimentService.thawLite(ee);
    Collection<ExperimentalFactor> toremove = new HashSet<>(ee.getExperimentalDesign().getExperimentalFactors());
    for (ExperimentalFactor ef : toremove) {
        experimentalFactorService.delete(ef);
        ee.getExperimentalDesign().getExperimentalFactors().remove(ef);
    }
    expressionExperimentService.update(ee);
    processedExpressionDataVectorService.computeProcessedExpressionData(ee);
    ee = expressionExperimentService.thaw(ee);
    designImporter.importDesign(ee, this.getClass().getResourceAsStream("/data/analysis/expression/gse26927short/2684_GSE26927_expdesign.data.txt"));
}

Example 23 with GeoDomainObjectGeneratorLocal

use of ubic.gemma.core.loader.expression.geo.GeoDomainObjectGeneratorLocal in project Gemma by PavlidisLab.

the class MeanVarianceServiceTest method setUp.

@Before
public void setUp() throws Exception {
    ee = eeService.findByShortName("GSE2982");
    if (ee != null) {
        // might work, but array designs might be in the way.
        eeService.remove(ee);
    }
    geoService.setGeoDomainObjectGenerator(new GeoDomainObjectGeneratorLocal(this.getTestFileBasePath("gse2982Short")));
    Collection<?> results = geoService.fetchAndLoad("GSE2982", false, false, false);
    ee = (ExpressionExperiment) results.iterator().next();
    qt = this.createOrUpdateQt(ScaleType.LINEAR);
    qt.setIsNormalized(true);
    quantitationTypeService.update(qt);
    // important bit, need to createProcessedVectors manually before using it
    ee = processedExpressionDataVectorService.createProcessedDataVectors(ee);
}

Example 24 with GeoDomainObjectGeneratorLocal

use of ubic.gemma.core.loader.expression.geo.GeoDomainObjectGeneratorLocal in project Gemma by PavlidisLab.

the class ProcessedExpressionDataCreateServiceTest method testReorder.

@Test
public void testReorder() throws Exception {
    ExpressionExperiment old = eeService.findByShortName("GSE404");
    if (old != null) {
        eeService.remove(old);
    }
    try {
        geoService.setGeoDomainObjectGenerator(new GeoDomainObjectGeneratorLocal(this.getTestFileBasePath("gse404Short")));
        @SuppressWarnings("unchecked") Collection<ExpressionExperiment> results = (Collection<ExpressionExperiment>) geoService.fetchAndLoad("GSE404", false, true, false);
        this.ee = results.iterator().next();
    } catch (AlreadyExistsInSystemException e) {
        this.ee = (ExpressionExperiment) e.getData();
    }
    ee = this.eeService.thawLite(ee);
    processedExpressionDataVectorService.computeProcessedExpressionData(ee);
    ExperimentalFactor factor = ExperimentalFactor.Factory.newInstance();
    factor.setType(FactorType.CATEGORICAL);
    factor.setName(ee.getShortName() + " design");
    factor.setExperimentalDesign(ee.getExperimentalDesign());
    factor = eeService.addFactor(ee, factor);
    FactorValue fv1 = FactorValue.Factory.newInstance();
    FactorValue fv2 = FactorValue.Factory.newInstance();
    fv1.setValue("foo");
    fv1.setExperimentalFactor(factor);
    fv2.setValue("bar");
    fv2.setIsBaseline(true);
    fv2.setExperimentalFactor(factor);
    eeService.addFactorValue(ee, fv1);
    eeService.addFactorValue(ee, fv2);
    List<BioAssay> basInOrder = new ArrayList<>(ee.getBioAssays());
    Collections.sort(basInOrder, new Comparator<BioAssay>() {

        @Override
        public int compare(BioAssay o1, BioAssay o2) {
            return o1.getId().compareTo(o2.getId());
        }
    });
    int i = 0;
    for (BioAssay ba : basInOrder) {
        // bioAssayService.thawRawAndProcessed( ba );
        BioMaterial bm = ba.getSampleUsed();
        assert fv1.getId() != null;
        if (!bm.getFactorValues().isEmpty()) {
            continue;
        }
        if (i % 2 == 0) {
            bm.getFactorValues().add(fv1);
        // log.info( bm + " " + bm.getId() + " => " + fv1 );
        } else {
            bm.getFactorValues().add(fv2);
        // log.info( bm + " " + bm.getId() + " => " + fv2 );
        }
        bioMaterialService.update(bm);
        i++;
    }
    factor = this.experimentalFactorService.load(factor.getId());
    assertEquals(2, factor.getFactorValues().size());
    /*
         * All that was setup. Now do the interesting bit
         */
    processedExpressionDataVectorService.reorderByDesign(ee.getId());
    /*
         * Now check the vectors...
         */
    Collection<ProcessedExpressionDataVector> resortedVectors = processedExpressionDataVectorService.getProcessedDataVectors(ee);
    // ExpressionDataDoubleMatrix newMat = new ExpressionDataDoubleMatrix( resortedVectors );
    // log.info( newMat );
    boolean foundVector = false;
    assertTrue(resortedVectors.size() > 0);
    for (ProcessedExpressionDataVector vector : resortedVectors) {
        i = 0;
        log.debug(vector.getDesignElement().getName() + " .........................");
        // thawingto avoid lazy error because we are outside of transaction in this test. All references in code run
        // inside a transaction
        BioAssayDimension bioAssayDimension = vector.getBioAssayDimension();
        bioAssayDimensionService.thawLite(bioAssayDimension);
        Collection<BioAssay> bioAssays = bioAssayDimension.getBioAssays();
        for (BioAssay ba : bioAssays) {
            BioMaterial bm = ba.getSampleUsed();
            assertEquals(1, bm.getFactorValues().size());
            FactorValue fv = bm.getFactorValues().iterator().next();
            assertNotNull(fv.getId());
            log.debug(ba.getId() + " " + fv.getId() + " " + fv);
            if (i < 10) {
                // first because it is baseline;
                assertEquals(fv2, fv);
            }
            i++;
        }
        /*
             * spot check the data, same place as before.
             */
        if (vector.getDesignElement().getName().equals("40")) {
            foundVector = true;
            ByteArrayConverter conv = new ByteArrayConverter();
            Double[] d = ArrayUtils.toObject(conv.byteArrayToDoubles(vector.getData()));
            assertEquals(20, d.length);
            assertEquals(-0.08, d[1], 0.001);
            assertEquals(0.45, d[10], 0.001);
            assertEquals(Double.NaN, d[19], 0.001);
        }
    }
    assertTrue("test vector not found", foundVector);
}

Example 25 with GeoDomainObjectGeneratorLocal

use of ubic.gemma.core.loader.expression.geo.GeoDomainObjectGeneratorLocal in project Gemma by PavlidisLab.

the class ProcessedExpressionDataCreateServiceTest method testComputeDevRankForExpressionExperimentMultiArrayWithGaps.

/**
 * Three platforms, one sample was not run on GPL81. It's 'Norm-1a', but the name we use for the sample is random.
 */
@SuppressWarnings("unchecked")
@Test
public void testComputeDevRankForExpressionExperimentMultiArrayWithGaps() throws Exception {
    try {
        geoService.setGeoDomainObjectGenerator(new GeoDomainObjectGeneratorLocal(this.getTestFileBasePath("gse482short")));
        Collection<ExpressionExperiment> results = (Collection<ExpressionExperiment>) geoService.fetchAndLoad("GSE482", false, true, false);
        this.ee = results.iterator().next();
    } catch (AlreadyExistsInSystemException e) {
        this.ee = ((Collection<ExpressionExperiment>) e.getData()).iterator().next();
    }
    ee = this.eeService.thawLite(ee);
    processedExpressionDataVectorService.computeProcessedExpressionData(ee);
    Collection<ProcessedExpressionDataVector> preferredVectors = this.processedExpressionDataVectorService.getProcessedDataVectors(ee);
    ee = eeService.load(ee.getId());
    ee = this.eeService.thawLite(ee);
    processedExpressionDataVectorService.thaw(preferredVectors);
    ExpressionDataDoubleMatrix mat = new ExpressionDataDoubleMatrix(preferredVectors);
    assertEquals(10, mat.columns());
    boolean found = false;
    for (int i = 0; i < mat.rows(); i++) {
        Double[] row = mat.getRow(i);
        // debugging
        if (i == 0) {
            for (int j = 0; j < row.length; j++) {
                BioAssay ba = mat.getBioAssaysForColumn(j).iterator().next();
                System.err.println(ba.getName());
            }
        }
        System.err.print(mat.getRowElement(i).getDesignElement().getName() + "\t");
        for (double d : row) {
            System.err.print(String.format("%4.2f\t", d));
        }
        System.err.print("\n");
        CompositeSequence el = mat.getDesignElementForRow(i);
        for (int j = 0; j < row.length; j++) {
            BioAssay ba = mat.getBioAssaysForColumn(j).iterator().next();
            if (ba.getName().matches("PGA-MurLungHyper-Norm-1a[ABC]v2-s2") && (el.getName().equals("100001_at") || el.getName().equals("100002_at") || el.getName().equals("100003_at") || el.getName().equals("100004_at") || el.getName().equals("100005_at") || el.getName().equals("100006_at") || el.getName().equals("100007_at") || el.getName().equals("100009_r_at") || el.getName().equals("100010_at") || el.getName().equals("100011_at"))) {
                assertEquals(Double.NaN, row[j], 0.0001);
                found = true;
            } else {
                assertTrue("Got unexpected NA value for " + ba.getName() + " for " + el.getName(), !Double.isNaN(row[j]));
            }
        }
    }
    assertTrue(found);
    /*
         * Now do this through the processedExpressionDataVectorService
         */
    Collection<DoubleVectorValueObject> da = this.processedExpressionDataVectorService.getProcessedDataArrays(ee);
    assertEquals(30, da.size());
    found = false;
    boolean first = true;
    for (DoubleVectorValueObject v : da) {
        CompositeSequenceValueObject el = v.getDesignElement();
        double[] row = v.getData();
        // debugging
        if (first) {
            for (int j = 0; j < row.length; j++) {
                BioAssayValueObject ba = v.getBioAssays().get(j);
                System.err.println(ba.getName());
            }
            first = false;
        }
        System.err.print(el.getName() + "\t");
        for (double d : row) {
            System.err.print(String.format("%4.2f\t", d));
        }
        System.err.print("\n");
        assertEquals(10, row.length);
        for (int j = 0; j < row.length; j++) {
            assertNotNull(v.getBioAssays());
            BioAssayValueObject ba = v.getBioAssays().get(j);
            if (ba.getName().startsWith("Missing bioassay for biomaterial") && (el.getName().equals("100001_at") || el.getName().equals("100002_at") || el.getName().equals("100003_at") || el.getName().equals("100004_at") || el.getName().equals("100005_at") || el.getName().equals("100006_at") || el.getName().equals("100007_at") || el.getName().equals("100009_r_at") || el.getName().equals("100010_at") || el.getName().equals("100011_at"))) {
                assertEquals(Double.NaN, row[j], 0.0001);
                found = true;
            } else {
                assertTrue("Got unexpected NA value for " + ba.getName() + " for " + el.getName(), !Double.isNaN(row[j]));
            }
        }
    }
    assertTrue(found);
}