private IEnumerable <AnnotationAndOffset> GetAnnotationsAndOffsets(FastaDatabase database, out long estimatedProteins, CancellationToken?cancellationToken = null)
{
var indexedDb = new IndexedDatabase(database);
indexedDb.Read();
estimatedProteins = indexedDb.EstimateTotalPeptides(SearchMode, MinSequenceLength, MaxSequenceLength, MaxNumNTermCleavages, MaxNumCTermCleavages);
IEnumerable <AnnotationAndOffset> annotationsAndOffsets;
if (SearchMode == InternalCleavageType.MultipleInternalCleavages)
{
//annotationsAndOffsets = indexedDb.AnnotationsAndOffsetsNoEnzyme(MinSequenceLength, MaxSequenceLength);
annotationsAndOffsets = indexedDb.AnnotationsAndOffsetsNoEnzymeParallel(MinSequenceLength, MaxSequenceLength, MaxNumThreads, cancellationToken);
}
else if (SearchMode == InternalCleavageType.NoInternalCleavage)
{
annotationsAndOffsets = indexedDb.IntactSequenceAnnotationsAndOffsets(MinSequenceLength, MaxSequenceLength, MaxNumCTermCleavages);
}
else
{
annotationsAndOffsets = indexedDb
.SequenceAnnotationsAndOffsetsWithNtermOrCtermCleavageNoLargerThan(
MinSequenceLength, MaxSequenceLength, MaxNumNTermCleavages, MaxNumCTermCleavages);
}
return(annotationsAndOffsets);
}
[TestCase(3, @"TEST_FOLDER\MSPathFinderT\ID_005133_8491EFA2.fasta", 323719193)] // 3MB
//[TestCase(6, @"TEST_FOLDER\MSPathFinderT\ID_004530_B63BD900.fasta", 595227563)] // 6MB
//[TestCase(15, @"TEST_FOLDER\MSPathFinderT\ID_004208_295531A4.fasta", 1882434687)] // 15MB
public void TestSequenceEnumeration(double size, string dbFile, int expected)
{
var methodName = MethodBase.GetCurrentMethod().Name;
Utils.ShowStarting(methodName, dbFile);
var fastaFile = Utils.GetTestFile(methodName, dbFile.Replace("TEST_FOLDER", Utils.DEFAULT_TEST_FILE_FOLDER));
var sw = new System.Diagnostics.Stopwatch();
sw.Start();
var db = new FastaDatabase(fastaFile.FullName);
var indexedDb = new IndexedDatabase(db);
var numSequences = 0L;
var timeDB = sw.Elapsed;
Console.WriteLine("Read DB in " + timeDB.TotalSeconds + " Seconds");
var estimatedAnnOff = indexedDb.EstimateTotalPeptides(0, 30, 250);
var timeEstimate = sw.Elapsed;
Console.WriteLine("Read Estimate in " + (timeEstimate - timeDB).TotalSeconds + " Seconds");
//int coreCount = 0;
//foreach (var item in new System.Management.ManagementObjectSearcher("Select NumberOfCores from Win32_Processor").Get())
//{
// coreCount += int.Parse(item["NumberOfCores"].ToString());
//}
//Console.WriteLine("Number Of Cores: {0}", coreCount);
//Console.WriteLine("Processors: " + System.Environment.ProcessorCount);
Console.WriteLine("Estimated results: " + estimatedAnnOff);
var annotationsAndOffsets = indexedDb.AnnotationsAndOffsetsNoEnzymeParallel(30, 250);
var timeGetAnn = sw.Elapsed;
Console.WriteLine("Read Annotations in " + (timeGetAnn - timeEstimate).TotalSeconds + " Seconds");
/*/Parallel.ForEach(
* annotationsAndOffsets,
* // new ParallelOptions { MaxDegreeOfParallelism = 2},
* annotationAndOffset =>
* {
* Interlocked.Increment(ref numSequences);
* //++numSequences;
* }
* );/**/
//annotationsAndOffsets.Select(annotationsAndOffset => annotationsAndOffset.)
// Below, original: 110, 109(total) seconds
// Parallelizing AnnotationsAndOffsetsNoEnzyme: 86 seconds
// Parallelizing AnnotationsAndOffsetsNoEnzyme, yield returns: 79.6, 94, 60, 60 seconds
//
// 3MB
// serial:
// Parallel2: 107,
//
// 6MB
// serial:
// Parallel2:
//
// 15MB
// serial:
// Parallel2:
//using (var ofstream = new FileStream(Path.Combine(@"F:\InformedProteomicsTestFiles", Path.GetFileNameWithoutExtension(fastaFile) + "_par.txt"), FileMode.Create))
//using (var fout = new StreamWriter(ofstream))
//{
// foreach (var annOff in annotationsAndOffsets)
// {
// numSequences++;
// fout.WriteLine(annOff.Annotation);
// }
//}
numSequences = annotationsAndOffsets.Count();
var timeParForEach = sw.Elapsed;
Console.WriteLine("Parallel ForEach in " + (timeParForEach - timeGetAnn).TotalSeconds + " Seconds");
Console.WriteLine("NumPeptides: {0}", numSequences);
sw.Stop();
Console.WriteLine(@"{0:f4} sec", sw.Elapsed.TotalSeconds);
//Assert.AreEqual(188961836, numSequences);
Assert.AreEqual(expected, numSequences);
}
