/* Encodes sequencing starting at the given index (i) until buffer
* is full or the sequence ends and writes the result in the given WriteBitSheperd.
* Updates i according to its progression.
*/
void EncodeSeq(WriteBitShepherd bits, ref long i, FastqReader reader)
{
int writtenBits = 0;
while (reader.HasSeqLeft(i, 4) && writtenBits < BIT_BUFFER)
{
//Check for run-length run.
if (reader.GetSeq(i) == reader.GetSeq(i + 3) && reader.GetSeq(i) == reader.GetSeq(i + 2) &&
reader.GetSeq(i) == reader.GetSeq(i + 1))
{
long j = i + 4;
long l = i + 8199;
while (j < l && reader.HasSeqLeft(j, 1) && reader.GetSeq(j) == reader.GetSeq(j - 1))
{
j++;
}
int length = (int)(j - i);
if (length > 35)
{
bits.Write(127, 7); //flag for long run length
bits.Write(GetRandomBit(), 1);
bits.Write(Array.IndexOf(Bases, reader.GetSeq(j - 1)), 3);
bits.Write(length - 4, 13);
writtenBits += 24;
}
else
{
bits.Write(0, 7); //flag for short run length
bits.Write(GetRandomBit(), 1);
bits.Write(Array.IndexOf(Bases, reader.GetSeq(j - 1)), 3);
bits.Write(length - 4, 5);
writtenBits += 16;
}
i = j;
}
else
{
bits.Write(Encode(reader.GetSeq(i), reader.GetSeq(i + 1), reader.GetSeq(i + 2)), 7);
bits.Write(GetRandomBit(), 1);
i += 3;
writtenBits += 8;
}
}
bool end = false;
string last = "";
if (!reader.HasSeqLeft(i, 4))
{
while (reader.HasSeqLeft(i, 1)) //could still have 1, 2 or 3 bases
{
last += reader.GetSeq(i++);
}
end = true;
}
if (last != "")
{
last = last.PadRight(3, 'N');
bits.Write(Encode(last[0], last[1], last[2]), 7);
bits.Write(GetRandomBit(), 1);
writtenBits += 8;
}
if (end)
{
bits.Write(126, 7); // mark end of sequences blocks
bits.Write(GetRandomBit(), 1);
writtenBits += 8;
}
}
/* Encodes sequencing starting at the given index (i) until buffer
* is full or the sequence ends and writes the result in the given BinaryWriter.
* Updates i according to its progression.
*/
void EncodeSeq(byte[] buffer, ref long i, BinaryWriter writer, FastqReader reader)
{
//the first byte starts with 1 if we are encoding a seq
byte first = (byte)128;
int b = 0;
while (reader.HasSeqLeft(i, 4) && b < BUFFER)
{
//Check for run-length run.
if (reader.GetSeq(i) == reader.GetSeq(i + 3) && reader.GetSeq(i) == reader.GetSeq(i + 2) && reader.GetSeq(i) == reader.GetSeq(i + 1))
{
long j = i + 4;
long l = i + 19;
while (j < l && reader.HasSeqLeft(j, 1) && reader.GetSeq(j) == reader.GetSeq(i))
{
j++;
}
buffer[b++] = (byte)(128 + ((j - i - 4) << 3) + Array.IndexOf(Bases, reader.GetSeq(i)));
i = j;
}
else
{
buffer[b++] = Encode(reader.GetSeq(i++), reader.GetSeq(i++), reader.GetSeq(i++));
}
}
string last = "";
byte lastSeqByte = 0;
bool lastByte = false;
if (!reader.HasSeqLeft(i, 4))
{
while (reader.HasSeqLeft(i, 1)) //could still have 1, 2 or 3 bases
{
last += reader.GetSeq(i++);
}
}
if (last != "")
{
lastByte = true;
last = last.PadRight(3, 'N');
lastSeqByte = Encode(last[0], last[1], last[2]);
}
if (b == BUFFER && !lastByte)
{
writer.Write(first);
writer.Write(buffer);
}
else
{
first += (byte)32; //we have to tell the decoder that we have a block with a length
//different than BUFFER
writer.Write(first);
writer.Write(b + (lastByte ? 1 : 0));
writer.Write(buffer, 0, b);
if (lastByte)
{
writer.Write(lastSeqByte);
}
}
}