Class ExpressionExperimentBibRefFinderNetworkTest

java.lang.Object
ubic.gemma.core.loader.entrez.pubmed.ExpressionExperimentBibRefFinderNetworkTest
@Tag("pubmed") @Tag("geo") @Tag("slow") @NetworkAvailable(url="https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi") @ExtendWith(NetworkAvailableExtension.class) public class ExpressionExperimentBibRefFinderNetworkTest extends Object
The over-the-wire half of ExpressionExperimentBibRefFinderTest, which stubs GeoBrowser and never leaves the JVM.

Split into its own class so the unit tests stay in the default suite: a class-level @Tag("slow") filters every method in the class, so the two cannot live together.

These reach Entrez twice per method — esearch/esummary on db=gds to resolve the accession, then efetch on db=pubmed to build the reference. They are the only remaining check that the accession-to-UID mapping is done the supported way rather than by assuming GEO's internal 2 + zero-padded UID encoding.

Author:
pavlidis

  • Constructor Details

    • ExpressionExperimentBibRefFinderNetworkTest

      public ExpressionExperimentBibRefFinderNetworkTest()

  • Method Details

    • testLocatePrimaryReference

      @Test public void testLocatePrimaryReference() throws IOException
      Throws:
      IOException

    • testLocatePubMedId

      @Test public void testLocatePubMedId() throws IOException
      The id behind the assertion above, without the second PubMed round trip. Kept separate because this is the hop that moved off acc.cgi: if Entrez resolution regresses, this fails first and names the reason.
      Throws:
      IOException