Compress a data stream provided as a lazy ByteString.

There are no expected error conditions. All input data streams are valid. It is possible for unexpected errors to occur, such as running out of memory, or finding the wrong version of the zlib C library, these are thrown as exceptions.

Decompress a data stream provided as a lazy ByteString.

It will throw an exception if any error is encountered in the input data. If you need more control over error handling then use one the incremental versions, decompressST or decompressIO.

The unfolding of the compression process, where you provide a sequence of uncompressed data chunks as input and receive a sequence of compressed data chunks as output. The process is incremental, in that the demand for input and provision of output are interleaved.

Incremental compression in the ST monad. Using ST makes it possible to write pure lazy functions while making use of incremental compression.

Incremental compression in the IO monad.

A fold over the CompressStream in the given monad.

One way to look at this is that it runs the stream, using callback functions for the three stream events.

A variant on foldCompressStream that is pure rather than operating in a monad and where the input is provided by a lazy ByteString. So we only have to deal with the output and end parts, making it just like a foldr on a list of output chunks.

For example:

toChunks = foldCompressStreamWithInput (:) []

The unfolding of the decompression process, where you provide a sequence of compressed data chunks as input and receive a sequence of uncompressed data chunks as output. The process is incremental, in that the demand for input and provision of output are interleaved.

To indicate the end of the input supply an empty input chunk. Note that for gzipFormat with the default decompressAllMembers True you will have to do this, as the decompressor will look for any following members. With decompressAllMembers False the decompressor knows when the data ends and will produce DecompressStreamEnd without you having to supply an empty chunk to indicate the end of the input.

The possible error cases when decompressing a stream.

This can be shown to give a human readable error message.

Incremental decompression in the ST monad. Using ST makes it possible to write pure lazy functions while making use of incremental decompression.

Incremental decompression in the IO monad.

A fold over the DecompressStream in the given monad.

One way to look at this is that it runs the stream, using callback functions for the four stream events.

A variant on foldCompressStream that is pure rather than operating in a monad and where the input is provided by a lazy ByteString. So we only have to deal with the output, end and error parts, making it like a foldr on a list of output chunks.

For example:

toChunks = foldDecompressStreamWithInput (:) [] throw

The full set of parameters for compression. The defaults are defaultCompressParams.

The compressBufferSize is the size of the first output buffer containing the compressed data. If you know an approximate upper bound on the size of the compressed data then setting this parameter can save memory. The default compression output buffer size is 16k. If your estimate is wrong it does not matter too much, the default buffer size will be used for the remaining chunks.

The default set of parameters for compression. This is typically used with the compressWith function with specific parameters overridden.

The full set of parameters for decompression. The defaults are defaultDecompressParams.

The decompressBufferSize is the size of the first output buffer, containing the uncompressed data. If you know an exact or approximate upper bound on the size of the decompressed data then setting this parameter can save memory. The default decompression output buffer size is 32k. If your estimate is wrong it does not matter too much, the default buffer size will be used for the remaining chunks.

One particular use case for setting the decompressBufferSize is if you know the exact size of the decompressed data and want to produce a strict ByteString. The compression and decompression functions use lazy ByteStrings but if you set the decompressBufferSize correctly then you can generate a lazy ByteString with exactly one chunk, which can be converted to a strict ByteString in O(1) time using concat . toChunks.

The default set of parameters for decompression. This is typically used with the compressWith function with specific parameters overridden.

The format used for compression or decompression. There are three variations.

The gzip format uses a header with a checksum and some optional meta-data about the compressed file. It is intended primarily for compressing individual files but is also sometimes used for network protocols such as HTTP. The format is described in detail in RFC #1952 http://www.ietf.org/rfc/rfc1952.txt

The zlib format uses a minimal header with a checksum but no other meta-data. It is especially designed for use in network protocols. The format is described in detail in RFC #1950 http://www.ietf.org/rfc/rfc1950.txt

The 'raw' format is just the compressed data stream without any additional header, meta-data or data-integrity checksum. The format is described in detail in RFC #1951 http://www.ietf.org/rfc/rfc1951.txt

This is not a format as such. It enabled zlib or gzip decoding with automatic header detection. This only makes sense for decompression.

The compression level parameter controls the amount of compression. This is a trade-off between the amount of compression and the time required to do the compression.

The default compression level is 6 (that is, biased towards higher compression at expense of speed).

No compression, just a block copy.

The fastest compression method (less compression)

The slowest compression method (best compression).

A specific compression level between 0 and 9.

The compression method

'Deflate' is the only method supported in this version of zlib. Indeed it is likely to be the only method that ever will be supported.

This specifies the size of the compression window. Larger values of this parameter result in better compression at the expense of higher memory usage.

The compression window size is the value of the the window bits raised to the power 2. The window bits must be in the range 9..15 which corresponds to compression window sizes of 512b to 32Kb. The default is 15 which is also the maximum size.

The total amount of memory used depends on the window bits and the MemoryLevel. See the MemoryLevel for the details.

The default WindowBits is 15 which is also the maximum size.

A specific compression window size, specified in bits in the range 9..15

The MemoryLevel parameter specifies how much memory should be allocated for the internal compression state. It is a tradeoff between memory usage, compression ratio and compression speed. Using more memory allows faster compression and a better compression ratio.

The total amount of memory used for compression depends on the WindowBits and the MemoryLevel. For decompression it depends only on the WindowBits. The totals are given by the functions:

compressTotal windowBits memLevel = 4 * 2^windowBits + 512 * 2^memLevel
decompressTotal windowBits = 2^windowBits

For example, for compression with the default windowBits = 15 and memLevel = 8 uses 256Kb. So for example a network server with 100 concurrent compressed streams would use 25Mb. The memory per stream can be halved (at the cost of somewhat degraded and slower compression) by reducing the windowBits and memLevel by one.

Decompression takes less memory, the default windowBits = 15 corresponds to just 32Kb.

The default memory level. (Equivalent to memoryLevel 8)

Use minimum memory. This is slow and reduces the compression ratio. (Equivalent to memoryLevel 1)

Use maximum memory for optimal compression speed. (Equivalent to memoryLevel 9)

A specific level in the range 1..9

The strategy parameter is used to tune the compression algorithm.

The strategy parameter only affects the compression ratio but not the correctness of the compressed output even if it is not set appropriately.

Use this default compression strategy for normal data.

Use the filtered compression strategy for data produced by a filter (or predictor). Filtered data consists mostly of small values with a somewhat random distribution. In this case, the compression algorithm is tuned to compress them better. The effect of this strategy is to force more Huffman coding and less string matching; it is somewhat intermediate between defaultCompressionStrategy and huffmanOnlyCompressionStrategy.

Use the Huffman-only compression strategy to force Huffman encoding only (no string match).