public void DoTestOlderProteomeDb(TestContext testContext, bool doActualWebAccess)
{
using (var testFilesDir = new TestFilesDir(testContext, ZIP_FILE))
{
string fastaPath = testFilesDir.GetTestPath("tiny.fasta");
string protDbPath = testFilesDir.GetTestPath("celegans_mini.protdb"); // a version 0 protdb file
string blibPath = testFilesDir.GetTestPath("random.blib"); // a bibliospec file
// What happens when you try to open a random file as a protdb file?
AssertEx.ThrowsException<DbException>(() => ProteomeDb.OpenProteomeDb(fastaPath));
// What happens when you try to open a non-protdb database file as a protdb file?
AssertEx.ThrowsException<FileLoadException>(() => ProteomeDb.OpenProteomeDb(blibPath));
using (ProteomeDb proteomeDb = ProteomeDb.OpenProteomeDb(protDbPath))
{
Assert.IsTrue(proteomeDb.GetSchemaVersionMajor() == 0); // the initial db from our zipfile should be ancient
Assert.IsTrue(proteomeDb.GetSchemaVersionMinor() == 0); // the initial db from our zipfile should be ancient
Assert.AreEqual(9, proteomeDb.GetProteinCount());
var protein = proteomeDb.GetProteinByName("Y18D10A.20");
Assert.IsNotNull(protein);
Assert.IsTrue(String.IsNullOrEmpty(protein.Accession)); // old db won't have this populated
WebEnabledFastaImporter searcher = new WebEnabledFastaImporter(doActualWebAccess ? null :new WebEnabledFastaImporter.FakeWebSearchProvider());
Assert.IsTrue(proteomeDb.LookupProteinMetadata(Progress, searcher)); // add any missing protein metadata
protein = proteomeDb.GetProteinByName("Y18D10A.20");
Assert.IsNotNull(protein);
if (doActualWebAccess) // We can actually go to the web for metadata
Assert.AreEqual( "Q9XW16", protein.Accession);
using (var reader = new StreamReader(fastaPath))
{
proteomeDb.AddFastaFile(reader, (msg, progress) => true);
}
// the act of writing should update to the current version
Assert.AreEqual(ProteomeDb.SCHEMA_VERSION_MAJOR_CURRENT, proteomeDb.GetSchemaVersionMajor());
Assert.AreEqual(ProteomeDb.SCHEMA_VERSION_MINOR_CURRENT, proteomeDb.GetSchemaVersionMinor());
Assert.AreEqual(19, proteomeDb.GetProteinCount());
// check for propery processed protein metadata
Assert.IsTrue(proteomeDb.LookupProteinMetadata(Progress,searcher));
protein = proteomeDb.GetProteinByName("IPI00000044");
Assert.IsNotNull(protein);
Assert.AreEqual("P01127", protein.Accession); // We get this offline with our ipi->uniprot mapper
if (doActualWebAccess)
Assert.AreEqual("PDGFB_HUMAN", protein.PreferredName); // But this we get only with web access
// TODO: bspratt fix this - GetDigestion has no notion of a Db that has been added to, doesn't digest the new proteins and returns immediately (issue #304)
// Enzyme trypsin = EnzymeList.GetDefault();
//proteomeDb.Digest(new ProteaseImpl(trypsin), (msg, progress) => true);
//Digestion digestion = proteomeDb.GetDigestion(trypsin.Name);
//var digestedProteins0 = digestion.GetProteinsWithSequencePrefix("EDGWVK", 100);
//Assert.IsTrue(digestedProteins0.Count >= 1);
}
}
}
public void AddFastaFile(StreamReader reader, ProgressMonitor progressMonitor)
{
Dictionary<string, ProtIdNames> proteinIds = new Dictionary<string, ProtIdNames>();
using (ISession session = OpenWriteSession())
{
foreach (DbProtein protein in session.CreateCriteria(typeof(DbProtein)).List())
{
if (protein.Id.HasValue)
proteinIds.Add(protein.Sequence, new ProtIdNames(protein.Id.Value, protein.Names));
}
int proteinCount = 0;
using (var transaction = session.BeginTransaction())
using (IDbCommand insertProtein = session.Connection.CreateCommand())
using (IDbCommand insertName = session.Connection.CreateCommand())
{
WebEnabledFastaImporter fastaImporter = new WebEnabledFastaImporter(new WebEnabledFastaImporter.DelayedWebSearchProvider()); // just parse, no search for now
insertProtein.CommandText =
"INSERT INTO ProteomeDbProtein (Version, Sequence) Values (1,?);select last_insert_rowid();"; // Not L10N
insertProtein.Parameters.Add(new SQLiteParameter());
insertName.CommandText =
"INSERT INTO ProteomeDbProteinName (Version, Protein, IsPrimary, Name, Description, PreferredName, Accession, Gene, Species, WebSearchStatus) Values(1,?,?,?,?,?,?,?,?,?)"; // Not L10N
insertName.Parameters.Add(new SQLiteParameter()); // Id
insertName.Parameters.Add(new SQLiteParameter()); // IsPrimary
insertName.Parameters.Add(new SQLiteParameter()); // Name
insertName.Parameters.Add(new SQLiteParameter()); // Description
insertName.Parameters.Add(new SQLiteParameter()); // PreferredName
insertName.Parameters.Add(new SQLiteParameter()); // Accession
insertName.Parameters.Add(new SQLiteParameter()); // Gene
insertName.Parameters.Add(new SQLiteParameter()); // Species
insertName.Parameters.Add(new SQLiteParameter()); // WebSearchInfo
foreach (DbProtein protein in fastaImporter.Import(reader))
{
int iProgress = (int)(reader.BaseStream.Position * 100 / (reader.BaseStream.Length + 1));
if (!progressMonitor.Invoke(string.Format(Resources.ProteomeDb_AddFastaFile_Added__0__proteins,proteinCount), iProgress))
{
return;
}
bool existingProtein = false;
ProtIdNames proteinIdNames;
if (proteinIds.TryGetValue(protein.Sequence, out proteinIdNames))
{
existingProtein = true;
}
else
{
((SQLiteParameter)insertProtein.Parameters[0]).Value = protein.Sequence;
proteinIdNames = new ProtIdNames(Convert.ToInt64(insertProtein.ExecuteScalar()), new DbProteinName[0]);
proteinIds.Add(protein.Sequence, proteinIdNames);
proteinCount++;
}
foreach (var proteinName in protein.Names)
{
// Skip any names that already exist
if (proteinIdNames.Names.Any(dbProteinName => Equals(dbProteinName.Name, proteinName.Name)))
continue;
try
{
((SQLiteParameter)insertName.Parameters[0]).Value = proteinIdNames.Id;
((SQLiteParameter)insertName.Parameters[1]).Value = proteinName.IsPrimary && !existingProtein;
((SQLiteParameter)insertName.Parameters[2]).Value = proteinName.Name;
((SQLiteParameter)insertName.Parameters[3]).Value = proteinName.Description;
((SQLiteParameter)insertName.Parameters[4]).Value = proteinName.PreferredName;
((SQLiteParameter)insertName.Parameters[5]).Value = proteinName.Accession;
((SQLiteParameter)insertName.Parameters[6]).Value = proteinName.Gene;
((SQLiteParameter)insertName.Parameters[7]).Value = proteinName.Species;
((SQLiteParameter)insertName.Parameters[8]).Value = proteinName.WebSearchStatus; // represent as a string for ease of serialization
insertName.ExecuteNonQuery();
}
catch (Exception exception)
{
Console.Out.WriteLine(exception);
}
}
}
if (!progressMonitor.Invoke(Resources.ProteomeDb_AddFastaFile_Saving_changes, 99))
{
return;
}
transaction.Commit();
}
AnalyzeDb(session);
progressMonitor.Invoke(
string.Format(Resources.ProteomeDb_AddFastaFile_Finished_importing__0__proteins, proteinCount), 100);
}
}
/// <summary>
/// Access the web to resolve protein metadata not directly found in fasta file.
/// The fasta text importer will have left search hints in ProteinMetadata.
/// </summary>
/// <param name="progressMonitor"></param>
/// <param name="fastaImporter">object that accesses the web, or pretends to if in a test</param>
/// <param name="polite">if true, don't try to resolve everything in one go, assume we can come back later</param>
/// <returns>true on success</returns>
public bool LookupProteinMetadata(ProgressMonitor progressMonitor, WebEnabledFastaImporter fastaImporter, bool polite = false)
{
var unsearchedProteins = new List<ProteinSearchInfo>();
List<DbProteinName> untaggedProteins;
using (ISession session = OpenSession())
{
if (!progressMonitor.Invoke(Resources.ProteomeDb_LookupProteinMetadata_looking_for_unresolved_protein_details, 0))
{
return false;
}
// get a list of proteins with unresolved metadata websearches
var proteinNames = session.CreateCriteria(typeof (DbProteinName)).List<DbProteinName>();
var proteinsToSearch =
proteinNames.Where(proteinName => (proteinName.GetProteinMetadata().GetPendingSearchTerm().Length > 0))
.ToList();
// and a list of proteins which have never been considered for metadata search
untaggedProteins =
proteinNames.Where(proteinName => proteinName.WebSearchInfo.IsEmpty()).ToList();
foreach (var untaggedProtein in untaggedProteins)
{
untaggedProtein.SetWebSearchCompleted(); // by default take this out of consideration for next time
var metadata = untaggedProtein.GetProteinMetadata();
if (metadata.HasMissingMetadata())
{
var search = fastaImporter.ParseProteinMetaData(metadata);
if (search!=null)
{
metadata = untaggedProtein.ChangeProteinMetadata(metadata.Merge(search)); // don't stomp name by accident
metadata = untaggedProtein.ChangeProteinMetadata(metadata.ChangeWebSearchInfo(search.WebSearchInfo));
}
}
if (metadata.NeedsSearch())
proteinsToSearch.Add(untaggedProtein); // add to the list of things to commit back to the db
}
// Get the lengths of the sequences without getting the sequences themselves, for best speed
var proteinIds = proteinsToSearch.Select(name => name.Protein.Id.Value).Distinct().ToArray();
var proteinLengths = new Dictionary<long, int>();
using (var cmd = session.Connection.CreateCommand())
{
string sql = "SELECT Id, LENGTH(Sequence) AS SequenceLength FROM ProteomeDbProtein P"; // Not L10N
if (proteinIds.Length < 1000)
{
sql += " WHERE P.Id IN (" + // Not L10N
string.Join(",", proteinIds) + ")"; // Not L10N
}
cmd.CommandText = sql;
using (var reader = cmd.ExecuteReader())
{
while (reader.Read())
{
var id = reader.GetValue(0);
var len = reader.GetValue(1);
proteinLengths.Add(Convert.ToInt64(id), Convert.ToInt32(len));
}
}
}
foreach (var p in proteinsToSearch)
{
int length;
proteinLengths.TryGetValue(p.Protein.Id.GetValueOrDefault(), out length);
unsearchedProteins.Add(new ProteinSearchInfo(p, length));
}
}
if (untaggedProteins.Any(untagged => !untagged.GetProteinMetadata().NeedsSearch())) // did any get set as unsearchable?
{
// Write back the ones that were formerly without search terms, but which now indicate no search is possible
using (ISession session = OpenWriteSession())
{
using (var transaction = session.BeginTransaction())
{
foreach (var untagged in untaggedProteins.Where(untagged => !untagged.GetProteinMetadata().NeedsSearch()))
session.SaveOrUpdate(untagged); // update the metadata
transaction.Commit();
}
}
}
if (unsearchedProteins.Any())
{
int resultsCount = 0;
int unsearchedCount = unsearchedProteins.Count;
for (bool success = true; success;)
{
success = false; // Until we see at least one succeed this round
var results = new List<DbProteinName>();
// The "true" arg means "do just one batch then return"
foreach (var result in fastaImporter.DoWebserviceLookup(unsearchedProteins, null, true))
{
if (result != null)
{
if (
!progressMonitor.Invoke(
string.Format(
Resources.ProteomeDb_LookupProteinMetadata_Retrieving_details_for__0__proteins,
unsearchedProteins.Count), 100 * resultsCount++ / unsearchedCount))
{
return false;
}
success = true;
results.Add(result.ProteinDbInfo);
}
}
if (results.Any()) // save this batch
{
using (var session = OpenWriteSession())
{
using (var transaction = session.BeginTransaction())
{
foreach (var result in results)
session.SaveOrUpdate(result);
transaction.Commit();
session.Close();
}
}
}
// Edit this list rather than rederive with database access
var hits = unsearchedProteins.Where(p => !p.GetProteinMetadata().NeedsSearch()).ToList();
foreach (var hit in hits)
{
unsearchedProteins.Remove(hit);
}
}
}
return true;
}
private void DoWork()
{
try
{
// just do the basic name+description parsing, no regex or web access - we don't use extended metadata here
var fastaImporter = new WebEnabledFastaImporter(new WebEnabledFastaImporter.FakeWebSearchProvider());
var proteins = new Dictionary<string, ProteinData>();
foreach (var path in FastaFilePaths)
{
var statusText = "Reading FASTA file " + Path.GetFileName(path);
var fileInfo = new FileInfo(path);
var reader = File.OpenText(path);
foreach (var protein in fastaImporter.Import(File.OpenText(path)))
{
if (!UpdateProgress(statusText, (int) (reader.BaseStream.Position * 100 / fileInfo.Length)))
{
return;
}
ProteinData proteinData;
if (!proteins.TryGetValue(protein.Sequence, out proteinData))
{
proteinData = new ProteinData(protein.Sequence);
proteins.Add(protein.Sequence, proteinData);
}
foreach (var name in protein.Names)
{
proteinData.AddName(name.Name, name.Description);
}
}
}
UpdatePeptides(proteins);
if (!IsDisposed)
{
BeginInvoke(new Action(Close));
}
}
finally
{
lock(this)
{
_running = false;
Monitor.PulseAll(this);
}
}
}