/// <summary>
/// Create a DensePairAnsi object from a sequence of RowKeyColKeyValue triples.
/// </summary>
/// <param name="tripleEnumerable">a sequence of RowKeyColKeyValue</param>
/// <returns>A DensePairAnsi object</returns>
public static DensePairAnsi GetInstanceFromSparse(IEnumerable <RowKeyColKeyValue <string, string, UOPair <char> > > tripleEnumerable)
{
DensePairAnsi densePairAnsi = new DensePairAnsi();
densePairAnsi.GetInstanceFromSparseInternal(tripleEnumerable);
return(densePairAnsi);
}
//Similar code elsewhere, but hard to move static code to common location
//Would be nice to convert UOPair<double>, etc to UOPair<char>
/// <summary>
/// Write a matrix to DensePairAnsi file format
/// </summary>
/// <param name="matrix">The matrix to write</param>
/// <param name="textWriter">The TextWriter to write to</param>
/// <param name="parallelOptions">A ParallelOptions instance that configures the multithreaded behavior of this operation.</param>
public static void WriteDensePairAnsi(this Matrix <string, string, UOPair <char> > matrix, TextWriter textWriter, ParallelOptions parallelOptions)
{
textWriter.WriteLine("var\t{0}", matrix.ColKeys.StringJoin("\t"));
foreach (string rowKey in matrix.RowKeys)
{
textWriter.Write(rowKey);
textWriter.Write("\t");
int rowIndex = matrix.IndexOfRowKey[rowKey];
List <UOPair <byte> > storeList = new List <UOPair <byte> >(matrix.ColCount);
for (int colIndex = 0; colIndex < matrix.ColCount; ++colIndex)
{
UOPair <char> value;
UOPair <byte> store;
if (matrix.TryGetValue(rowIndex, colIndex, out value))
{
store = DensePairAnsi.StaticValueToStore(value);
}
else
{
store = DensePairAnsi.StaticStoreMissingValue;
}
storeList.Add(store);
}
Helper.CheckCondition(storeList.Count == matrix.ColCount, Properties.Resource.ExpectedStoreListCountToEqualColCount, storeList.Count, matrix.ColCount);
string s = DensePairAnsi.StoreListToString(storeList, matrix.ColCount);
textWriter.WriteLine(s);
}
}
/// <summary>
/// Creates a DensePairAnsi object from a file in dense pair ansi format.
/// </summary>
/// <param name="densePairAnsiFileName">a file in dense pair ansi format</param>
/// <param name="parallelOptions">A ParallelOptions instance that configures the multithreaded behavior of this operation.</param>
/// <returns>the DensePairAnsi object</returns>
public static DensePairAnsi GetInstance(string densePairAnsiFileName, ParallelOptions parallelOptions)
{
DensePairAnsi densePairAnsi = new DensePairAnsi();
densePairAnsi.GetInstanceInternal(densePairAnsiFileName, parallelOptions);
return(densePairAnsi);
}
/// <summary>
/// Parses directly from the SparseFile format. Useful if memory must be concerved.
/// </summary>
/// <param name="inputSparseFileName">Name of a file in Sparse format.</param>
/// <returns>An instance of DensePairAnsi.</returns>
public static DensePairAnsi GetInstanceFromSparse(string inputSparseFileName)
{
DensePairAnsi densePairAnsi = new DensePairAnsi();
densePairAnsi.GetInstanceFromSparseInternal(inputSparseFileName);
return(densePairAnsi);
}
/// <summary>
/// Creates an DensePairAnsi with values missing.
/// </summary>
/// <param name="rowKeySequence">A sequence of row keys. The items will become the RowKeys of the Matrix.</param>
/// <param name="colKeySequence">A sequence of colKeys. The items will come the ColKeys of the Matrix.</param>
/// <param name="missingValue">The special Missing value, which must be (UO '?','?')</param>
/// <returns>A new, empty, DensePairAnsi</returns>
static public DensePairAnsi CreateEmptyInstance(IEnumerable <string> rowKeySequence, IEnumerable <string> colKeySequence, UOPair <char> missingValue)
{
Helper.CheckCondition(missingValue.Equals(StaticMissingValue), "For DensePairAnsi the missingValue must be (UO '?', '?')"); //OK to use Equals because UOPair<char> can't be null.
DensePairAnsi densePairAnsi = new DensePairAnsi();
densePairAnsi.InternalCreateEmptyInstance(rowKeySequence, colKeySequence);
return(densePairAnsi);
}
/// <summary>
/// Creates an DensePairAnsi with values missing.
/// </summary>
/// <param name="rowKeySequence">A sequence of row keys. The items will become the RowKeys of the Matrix.</param>
/// <param name="colKeySequence">A sequence of colKeys. The items will come the ColKeys of the Matrix.</param>
/// <param name="missingValue">The special Missing value, which must be (UO '?','?')</param>
/// <returns>A new, empty, DensePairAnsi</returns>
static public DensePairAnsi CreateEmptyInstance(IEnumerable <string> rowKeySequence, IEnumerable <string> colKeySequence, UOPair <char> missingValue)
{
//OK to use Equals because UOPair<char> can't be null.
Helper.CheckCondition(missingValue.Equals(StaticMissingValue), Properties.Resource.DensePairAnsiMissingValueSignatureMustBe);
DensePairAnsi densePairAnsi = new DensePairAnsi();
densePairAnsi.InternalCreateEmptyInstance(rowKeySequence, colKeySequence);
return(densePairAnsi);
}
/// <summary>
/// Converts matrix to a DensePairAnsi. Even if the matrix is already an densePairAnsi, a new one is created..
/// </summary>
/// <param name="matrix">The matrix to convert from</param>
/// <param name="parallelOptions">Options for controlling any parallelism.</param>
/// <returns>A densePairAnsi version of the matrix</returns>
public static DensePairAnsi ToDensePairAnsi(this Matrix <string, string, UOPair <char> > matrix, ParallelOptions parallelOptions)
{
var densePairAnsi = DensePairAnsi.CreateEmptyInstance(matrix.RowKeys, matrix.ColKeys, DensePairAnsi.StaticMissingValue);
//Console.WriteLine("Convert no more than {0} values", matrix.RowCount * matrix.ColCount);
//CounterWithMessages counterWithMessages = new CounterWithMessages("adding value #{0}", 100000, null);
Parallel.ForEach(matrix.RowKeyColKeyValues, parallelOptions, triple =>
{
//counterWithMessages.Increment();
densePairAnsi[triple.RowKey, triple.ColKey] = triple.Value;
});
return(densePairAnsi);
}
public static DensePairAnsi GetInstance(
IEnumerable <IGrouping <string, KeyValuePair <string, UOPair <char> > > > snpGroupsCidToNucPairEnumerable)
{
if (snpGroupsCidToNucPairEnumerable == null)
{
throw new ArgumentNullException("snpGroupsCidToNucPairEnumerable");
}
DensePairAnsi densePairAnsi = new DensePairAnsi();
densePairAnsi.ColSerialNumbers = new SerialNumbers <string>();
densePairAnsi.RowKeyToStoreList = new Dictionary <string, List <UOPair <byte> > >();
foreach (var snpGroupsCidToNucPair in snpGroupsCidToNucPairEnumerable)
{
List <UOPair <byte> > storeArray = Enumerable.Repeat(StaticStoreMissingValue, densePairAnsi.ColSerialNumbers.Count).ToList();
string var = snpGroupsCidToNucPair.Key;
densePairAnsi.RowKeyToStoreList.Add(var, storeArray);
foreach (var cidAndNucPair in snpGroupsCidToNucPair)
{
UOPair <char> valPair = cidAndNucPair.Value;
int colIndex = densePairAnsi.ColSerialNumbers.GetNewOrOld(cidAndNucPair.Key);
if (colIndex < storeArray.Count)
{
Helper.CheckCondition(storeArray[colIndex] == StaticStoreMissingValue, () => string.Format(CultureInfo.InvariantCulture, "Each pair of keys, i,e,.<{0},{1}>, should only be seen once", var, cidAndNucPair.Key));
storeArray[colIndex] = densePairAnsi.ValueToStore(valPair);
}
else
{
Debug.Assert(colIndex == storeArray.Count); // real assert
storeArray.Add(densePairAnsi.ValueToStore(valPair));
}
}
storeArray.TrimExcess();
}
densePairAnsi._rowKeys = densePairAnsi.RowKeyToStoreList.Keys.ToList();
densePairAnsi._indexOfRowKey = densePairAnsi.RowKeys.Select((key, index) => new { key, index }).ToDictionary(keyAndIndex => keyAndIndex.key, keyAndIndex => keyAndIndex.index);
return(densePairAnsi);
}
//!!! very, very similar to other code
/// <summary>
/// Creates a DensePairAnsi object from a sequence of groupings. A grouping here is a roew
/// </summary>
/// <param name="snpGroupsCidToNucPairEnumerable"></param>
/// <param name="inputsMissingChar"></param>
/// <returns></returns>
public static DensePairAnsi GetInstance(IEnumerable <IGrouping <string, KeyValuePair <string, UOPair <char> > > >
snpGroupsCidToNucPairEnumerable, char inputsMissingChar)
{
DensePairAnsi densePairAnsi = new DensePairAnsi();
densePairAnsi.ColSerialNumbers = new SerialNumbers <string>();
densePairAnsi.RowKeyToStoreList = new Dictionary <string, List <UOPair <byte> > >();
foreach (var snpGroupsCidToNucPair in snpGroupsCidToNucPairEnumerable)
{
//!!!const
List <UOPair <byte> > storeArray = Enumerable.Repeat(StaticStoreMissingValue, densePairAnsi.ColSerialNumbers.Count).ToList();
string var = snpGroupsCidToNucPair.Key;
densePairAnsi.RowKeyToStoreList.Add(var, storeArray);
foreach (var cidAndNucPair in snpGroupsCidToNucPair)
{
UOPair <char> valPair = cidAndNucPair.Value;
int colIndex = densePairAnsi.ColSerialNumbers.GetNewOrOld(cidAndNucPair.Key);
if (colIndex < storeArray.Count)
{
storeArray[colIndex] = densePairAnsi.ValueToStore(valPair);
}
else
{
Debug.Assert(colIndex == storeArray.Count); // real assert
storeArray.Add(densePairAnsi.ValueToStore(valPair));
}
}
storeArray.TrimExcess();
}
densePairAnsi._rowKeys = densePairAnsi.RowKeyToStoreList.Keys.ToList();
densePairAnsi._indexOfRowKey = densePairAnsi.RowKeys.Select((key, index) => new { key, index }).ToDictionary(keyAndIndex => keyAndIndex.key, keyAndIndex => keyAndIndex.index);
return(densePairAnsi);
}