Bio.Iteratee.Bgzf

Handling of BGZF files. Right now, we have an Enumeratee each for input and output. The input iteratee can optionally supply virtual file offsets, so that seeking is possible.

Documentation

data Block

One BGZF block: virtual offset and contents. Could also be a block of an uncompressed file, if we want to support indexing of uncompressed BAM or some silliness like that.

Constructors

Block

Fields

• block_offset :: !FileOffset
• block_contents :: !Bytes

Instances

Semigroup Block

Methods

(<>) :: Block -> Block -> Block

sconcat :: NonEmpty Block -> Block

stimes :: Integral b => b -> Block -> Block

Monoid Block

Methods

mempty :: Block

mappend :: Block -> Block -> Block

mconcat :: [Block] -> Block

Nullable Block

Methods

nullC :: Block -> Bool

NullPoint Block

Methods

emptyP :: Block

decompressBgzfBlocks' :: MonadIO m => Int -> Enumeratee Bytes Block m a

Decompress a BGZF stream into a stream of Blocks, np fold parallel.

decompressBgzfBlocks :: MonadIO m => Enumeratee Bytes Block m a

decompressBgzf :: MonadIO m => Enumeratee Bytes Bytes m a

Decompress a BGZF stream into a stream of Bytess.

decompressPlain :: MonadIO m => Enumeratee Bytes Block m a

Decompresses a plain file. What's actually happening is that the offset in the input stream is tracked and added to the Bytess giving Blocks. This results in the same interface as decompressing actual Bgzf.

maxBlockSize :: Int

Maximum block size for Bgzf: 64k with some room for headers and uncompressible stuff

bgzfEofMarker :: Bytes

The EOF marker for BGZF files. This is just an empty string compressed as BGZF. Appended to BAM files to indicate their end.

liftBlock :: Monad m => Iteratee Bytes m a -> Iteratee Block m a

Runs an Iteratee for Bytess when decompressing BGZF. Adds internal bookkeeping.

getOffset :: Iteratee Block m FileOffset

Get the current virtual offset. The virtual address in a BGZF stream contains the offset of the current block in the upper 48 bits and the current offset into that block in the lower 16 bits. This scheme is compatible with the way BAM files are indexed.

data BgzfChunk

Constructors

SpecialChunk !Bytes BgzfChunk

RecordChunk !Bytes BgzfChunk

LeftoverChunk !Bytes BgzfChunk

NoChunk

Instances

Nullable BgzfChunk

Methods

nullC :: BgzfChunk -> Bool

NullPoint BgzfChunk

Methods

emptyP :: BgzfChunk

isBgzf :: Monad m => Iteratee Bytes m Bool

Tests whether a stream is in BGZF format. Does not consume any input.

isGzip :: Monad m => Iteratee Bytes m Bool

Tests whether a stream is in GZip format. Also returns True on a Bgzf stream, which is technically a special case of GZip.

parMapChunksIO :: (MonadIO m, Nullable s) => Int -> (s -> IO t) -> Enumeratee s t m a

Parallel map of an IO action over the elements of a stream

This Enumeratee applies an IO action to every chunk of the input stream. These IO actions are run asynchronously in a limited parallel way. Don't forget to evaluate

compressBgzf :: MonadIO m => Enumeratee BgzfChunk Bytes m a

Like compressBgzf', with sensible defaults.

compressBgzfLv :: MonadIO m => Int -> Enumeratee BgzfChunk Bytes m a

compressBgzf' :: MonadIO m => CompressParams -> Enumeratee BgzfChunk Bytes m a

Compresses a stream of Bytess into a stream of BGZF blocks, in parallel

data CompressParams

Constructors

CompressParams

Fields

• compression_level :: Int
• queue_depth :: Int

Instances

Show CompressParams

Methods

showsPrec :: Int -> CompressParams -> ShowS

show :: CompressParams -> String

showList :: [CompressParams] -> ShowS

compressChunk :: Int -> Ptr Word8 -> CUInt -> IO Bytes