/**
* Insert a List of new Protein instances.
*
* @param newInstances being a List of instances to persist.
* @return the Collection of persisted instances. This MAY NOT contain the same objects as have
* been passed in, for sub-classes that check for the pre-existence of the object in the
* database.
*/
@Override
@Transactional
public Collection<Protein> insert(Collection<Protein> newInstances) {
final PersistedProteins persistedProteins = insertNewProteins(newInstances);
final Collection<Protein> allProteins =
new ArrayList<Protein>(persistedProteins.getNewProteins());
allProteins.addAll(persistedProteins.getPreExistingProteins());
entityManager.flush();
return allProteins;
}
/**
* Inserts new Proteins. If there are Protein objects with the same MD5 / sequence in the
* database, this method updates these proteins, rather than inserting the new ones.
*
* <p>Note that this method inserts the new Protein objects AND and new Xrefs (possibly updating
* an existing Protein object if necessary with the new Xref.)
*
* @param newProteins being a List of new Protein objects to insert
* @return a new List<Protein> containing all of the inserted / updated Protein objects. (Allows
* the caller to retrieve the primary keys for the proteins).
*/
@Transactional
@SuppressWarnings("unchecked")
public PersistedProteins insertNewProteins(Collection<Protein> newProteins) {
PersistedProteins persistentProteins = new PersistedProteins();
if (newProteins.size() > 0) {
// Create a List of MD5s (just as Strings) to query the database with
final List<String> newMd5s = new ArrayList<String>(newProteins.size());
for (Protein newProtein : newProteins) {
newMd5s.add(newProtein.getMd5());
if (LOGGER.isDebugEnabled()) {
LOGGER.debug("MD5 of new protein: " + newProtein.getMd5());
}
}
// Retrieve any proteins AND associated xrefs that have the same MD5 as one of the 'new'
// proteins
// being inserted and place in a Map of MD5 to Protein object.
final Map<String, Protein> md5ToExistingProtein = new HashMap<String, Protein>();
final Query query =
entityManager.createQuery(
"select p from Protein p left outer join fetch P.crossReferences where p.md5 in (:md5)");
query.setParameter("md5", newMd5s);
for (Protein existingProtein : (List<Protein>) query.getResultList()) {
if (LOGGER.isDebugEnabled()) {
LOGGER.debug("Found 1 existing protein with MD5: " + existingProtein.getMd5());
}
md5ToExistingProtein.put(existingProtein.getMd5(), existingProtein);
}
// Now have the List of 'new' proteins, and a list of existing proteins that match
// them. Insert / update proteins as appropriate.
for (Protein candidate : newProteins) {
// PROTEIN ALREADY EXISTS in the DB. - update cross references and save.
if (md5ToExistingProtein.keySet().contains(candidate.getMd5())) {
// This protein is already in the database - add any new Xrefs and update.
Protein existingProtein = md5ToExistingProtein.get(candidate.getMd5());
boolean updateRequired = false;
if (candidate.getCrossReferences() != null) {
if (LOGGER.isTraceEnabled()) {
LOGGER.trace("Protein TO BE STORED has xrefs:");
}
for (ProteinXref xref : candidate.getCrossReferences()) {
if (LOGGER.isTraceEnabled()) {
LOGGER.trace(xref.getIdentifier());
}
// Add any NEW cross references.
if (!existingProtein.getCrossReferences().contains(xref)) {
if (LOGGER.isTraceEnabled()) {
LOGGER.trace(
"Adding " + xref.getIdentifier() + " and setting updateRequired = true");
}
existingProtein.addCrossReference(xref);
updateRequired = true;
}
}
}
if (updateRequired) {
// PROTEIN is NOT new, but CHANGED (new Xrefs)
if (LOGGER.isTraceEnabled()) {
LOGGER.trace("Merging protein with new Xrefs: " + existingProtein.getMd5());
}
entityManager.merge(existingProtein);
}
persistentProteins.addPreExistingProtein(existingProtein);
}
// PROTEIN IS NEW - save it.
else {
if (LOGGER.isTraceEnabled()) {
LOGGER.trace("Saving new protein: " + candidate.getMd5());
}
entityManager.persist(candidate);
persistentProteins.addNewProtein(candidate);
// Check for this new protein next time through the loop, just in case the new source of
// proteins is redundant (e.g. a FASTA file with sequences repeated).
md5ToExistingProtein.put(candidate.getMd5(), candidate);
}
}
}
// Finally return all the persisted Protein objects (new or existing)
entityManager.flush();
return persistentProteins;
}