s2

package
v1.15.12 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Oct 26, 2022 License: Apache-2.0, BSD-3-Clause, MIT, + 1 more Imports: 14 Imported by: 225

README

S2 Compression

S2 is an extension of Snappy.

S2 is aimed for high throughput, which is why it features concurrent compression for bigger payloads.

Decoding is compatible with Snappy compressed content, but content compressed with S2 cannot be decompressed by Snappy. This means that S2 can seamlessly replace Snappy without converting compressed content.

S2 can produce Snappy compatible output, faster and better than Snappy. If you want full benefit of the changes you should use s2 without Snappy compatibility.

S2 is designed to have high throughput on content that cannot be compressed. This is important, so you don't have to worry about spending CPU cycles on already compressed data.

Benefits over Snappy

  • Better compression
  • Adjustable compression (3 levels)
  • Concurrent stream compression
  • Faster decompression, even for Snappy compatible content
  • Concurrent Snappy/S2 stream decompression
  • Ability to quickly skip forward in compressed stream
  • Random seeking with indexes
  • Compatible with reading Snappy compressed content
  • Smaller block size overhead on incompressible blocks
  • Block concatenation
  • Uncompressed stream mode
  • Automatic stream size padding
  • Snappy compatible block compression

Drawbacks over Snappy

  • Not optimized for 32 bit systems
  • Streams use slightly more memory due to larger blocks and concurrency (configurable)

Usage

Installation: go get -u github.com/klauspost/compress/s2

Full package documentation:

godoc

Compression

func EncodeStream(src io.Reader, dst io.Writer) error {
    enc := s2.NewWriter(dst)
    _, err := io.Copy(enc, src)
    if err != nil {
        enc.Close()
        return err
    }
    // Blocks until compression is done.
    return enc.Close() 
}

You should always call enc.Close(), otherwise you will leak resources and your encode will be incomplete.

For the best throughput, you should attempt to reuse the Writer using the Reset() method.

The Writer in S2 is always buffered, therefore NewBufferedWriter in Snappy can be replaced with NewWriter in S2. It is possible to flush any buffered data using the Flush() method. This will block until all data sent to the encoder has been written to the output.

S2 also supports the io.ReaderFrom interface, which will consume all input from a reader.

As a final method to compress data, if you have a single block of data you would like to have encoded as a stream, a slightly more efficient method is to use the EncodeBuffer method. This will take ownership of the buffer until the stream is closed.

func EncodeStream(src []byte, dst io.Writer) error {
    enc := s2.NewWriter(dst)
    // The encoder owns the buffer until Flush or Close is called.
    err := enc.EncodeBuffer(buf)
    if err != nil {
        enc.Close()
        return err
    }
    // Blocks until compression is done.
    return enc.Close()
}

Each call to EncodeBuffer will result in discrete blocks being created without buffering, so it should only be used a single time per stream. If you need to write several blocks, you should use the regular io.Writer interface.

Decompression

func DecodeStream(src io.Reader, dst io.Writer) error {
    dec := s2.NewReader(src)
    _, err := io.Copy(dst, dec)
    return err
}

Similar to the Writer, a Reader can be reused using the Reset method.

For the best possible throughput, there is a EncodeBuffer(buf []byte) function available. However, it requires that the provided buffer isn't used after it is handed over to S2 and until the stream is flushed or closed.

For smaller data blocks, there is also a non-streaming interface: Encode(), EncodeBetter() and Decode(). Do however note that these functions (similar to Snappy) does not provide validation of data, so data corruption may be undetected. Stream encoding provides CRC checks of data.

It is possible to efficiently skip forward in a compressed stream using the Skip() method. For big skips the decompressor is able to skip blocks without decompressing them.

Single Blocks

Similar to Snappy S2 offers single block compression. Blocks do not offer the same flexibility and safety as streams, but may be preferable for very small payloads, less than 100K.

Using a simple dst := s2.Encode(nil, src) will compress src and return the compressed result. It is possible to provide a destination buffer. If the buffer has a capacity of s2.MaxEncodedLen(len(src)) it will be used. If not a new will be allocated.

Alternatively EncodeBetter/EncodeBest can also be used for better, but slightly slower compression.

Similarly to decompress a block you can use dst, err := s2.Decode(nil, src). Again an optional destination buffer can be supplied. The s2.DecodedLen(src) can be used to get the minimum capacity needed. If that is not satisfied a new buffer will be allocated.

Block function always operate on a single goroutine since it should only be used for small payloads.

Commandline tools

Some very simply commandline tools are provided; s2c for compression and s2d for decompression.

Binaries can be downloaded on the Releases Page.

Installing then requires Go to be installed. To install them, use:

go install github.com/klauspost/compress/s2/cmd/s2c@latest && go install github.com/klauspost/compress/s2/cmd/s2d@latest

To build binaries to the current folder use:

go build github.com/klauspost/compress/s2/cmd/s2c && go build github.com/klauspost/compress/s2/cmd/s2d

s2c

Usage: s2c [options] file1 file2

Compresses all files supplied as input separately.
Output files are written as 'filename.ext.s2' or 'filename.ext.snappy'.
By default output files will be overwritten.
Use - as the only file name to read from stdin and write to stdout.

Wildcards are accepted: testdir/*.txt will compress all files in testdir ending with .txt
Directories can be wildcards as well. testdir/*/*.txt will match testdir/subdir/b.txt

File names beginning with 'http://' and 'https://' will be downloaded and compressed.
Only http response code 200 is accepted.

Options:
  -bench int
    	Run benchmark n times. No output will be written
  -blocksize string
    	Max  block size. Examples: 64K, 256K, 1M, 4M. Must be power of two and <= 4MB (default "4M")
  -c	Write all output to stdout. Multiple input files will be concatenated
  -cpu int
    	Compress using this amount of threads (default 32)
  -faster
    	Compress faster, but with a minor compression loss
  -help
    	Display help
  -index
        Add seek index (default true)    	
  -o string
        Write output to another file. Single input file only
  -pad string
    	Pad size to a multiple of this value, Examples: 500, 64K, 256K, 1M, 4M, etc (default "1")
  -q	Don't write any output to terminal, except errors
  -rm
    	Delete source file(s) after successful compression
  -safe
    	Do not overwrite output files
  -slower
    	Compress more, but a lot slower
  -snappy
        Generate Snappy compatible output stream
  -verify
    	Verify written files

s2d

Usage: s2d [options] file1 file2

Decompresses all files supplied as input. Input files must end with '.s2' or '.snappy'.
Output file names have the extension removed. By default output files will be overwritten.
Use - as the only file name to read from stdin and write to stdout.

Wildcards are accepted: testdir/*.txt will compress all files in testdir ending with .txt
Directories can be wildcards as well. testdir/*/*.txt will match testdir/subdir/b.txt

File names beginning with 'http://' and 'https://' will be downloaded and decompressed.
Extensions on downloaded files are ignored. Only http response code 200 is accepted.

Options:
  -bench int
    	Run benchmark n times. No output will be written
  -c	Write all output to stdout. Multiple input files will be concatenated
  -help
    	Display help
  -o string
        Write output to another file. Single input file only
  -offset string
        Start at offset. Examples: 92, 64K, 256K, 1M, 4M. Requires Index
  -q    Don't write any output to terminal, except errors
  -rm
        Delete source file(s) after successful decompression
  -safe
        Do not overwrite output files
  -tail string
        Return last of compressed file. Examples: 92, 64K, 256K, 1M, 4M. Requires Index
  -verify
    	Verify files, but do not write output

s2sx: self-extracting archives

s2sx allows creating self-extracting archives with no dependencies.

By default, executables are created for the same platforms as the host os, but this can be overridden with -os and -arch parameters.

Extracted files have 0666 permissions, except when untar option used.

Usage: s2sx [options] file1 file2

Compresses all files supplied as input separately.
If files have '.s2' extension they are assumed to be compressed already.
Output files are written as 'filename.s2sx' and with '.exe' for windows targets.
If output is big, an additional file with ".more" is written. This must be included as well.
By default output files will be overwritten.

Wildcards are accepted: testdir/*.txt will compress all files in testdir ending with .txt
Directories can be wildcards as well. testdir/*/*.txt will match testdir/subdir/b.txt

Options:
  -arch string
        Destination architecture (default "amd64")
  -c    Write all output to stdout. Multiple input files will be concatenated
  -cpu int
        Compress using this amount of threads (default 32)
  -help
        Display help
  -max string
        Maximum executable size. Rest will be written to another file. (default "1G")
  -os string
        Destination operating system (default "windows")
  -q    Don't write any output to terminal, except errors
  -rm
        Delete source file(s) after successful compression
  -safe
        Do not overwrite output files
  -untar
        Untar on destination

Available platforms are:

  • darwin-amd64
  • darwin-arm64
  • linux-amd64
  • linux-arm
  • linux-arm64
  • linux-mips64
  • linux-ppc64le
  • windows-386
  • windows-amd64

By default, there is a size limit of 1GB for the output executable.

When this is exceeded the remaining file content is written to a file called output+.more. This file must be included for a successful extraction and placed alongside the executable for a successful extraction.

This file must have the same name as the executable, so if the executable is renamed, so must the .more file.

This functionality is disabled with stdin/stdout.

Self-extracting TAR files

If you wrap a TAR file you can specify -untar to make it untar on the destination host.

Files are extracted to the current folder with the path specified in the tar file.

Note that tar files are not validated before they are wrapped.

For security reasons files that move below the root folder are not allowed.

Performance

This section will focus on comparisons to Snappy. This package is solely aimed at replacing Snappy as a high speed compression package. If you are mainly looking for better compression zstandard gives better compression, but typically at speeds slightly below "better" mode in this package.

Compression is increased compared to Snappy, mostly around 5-20% and the throughput is typically 25-40% increased (single threaded) compared to the Snappy Go implementation.

Streams are concurrently compressed. The stream will be distributed among all available CPU cores for the best possible throughput.

A "better" compression mode is also available. This allows to trade a bit of speed for a minor compression gain. The content compressed in this mode is fully compatible with the standard decoder.

Snappy vs S2 compression speed on 16 core (32 thread) computer, using all threads and a single thread (1 CPU):

File S2 Speed S2 Throughput S2 % smaller S2 "better" "better" throughput "better" % smaller
rawstudio-mint14.tar 16.33x 10556 MB/s 8.0% 6.04x 5252 MB/s 14.7%
(1 CPU) 1.08x 940 MB/s - 0.46x 400 MB/s -
github-june-2days-2019.json 16.51x 15224 MB/s 31.70% 9.47x 8734 MB/s 37.71%
(1 CPU) 1.26x 1157 MB/s - 0.60x 556 MB/s -
github-ranks-backup.bin 15.14x 12598 MB/s -5.76% 6.23x 5675 MB/s 3.62%
(1 CPU) 1.02x 932 MB/s - 0.47x 432 MB/s -
consensus.db.10gb 11.21x 12116 MB/s 15.95% 3.24x 3500 MB/s 18.00%
(1 CPU) 1.05x 1135 MB/s - 0.27x 292 MB/s -
apache.log 8.55x 16673 MB/s 20.54% 5.85x 11420 MB/s 24.97%
(1 CPU) 1.91x 1771 MB/s - 0.53x 1041 MB/s -
gob-stream 15.76x 14357 MB/s 24.01% 8.67x 7891 MB/s 33.68%
(1 CPU) 1.17x 1064 MB/s - 0.65x 595 MB/s -
10gb.tar 13.33x 9835 MB/s 2.34% 6.85x 4863 MB/s 9.96%
(1 CPU) 0.97x 689 MB/s - 0.55x 387 MB/s -
sharnd.out.2gb 9.11x 13213 MB/s 0.01% 1.49x 9184 MB/s 0.01%
(1 CPU) 0.88x 5418 MB/s - 0.77x 5417 MB/s -
sofia-air-quality-dataset csv 22.00x 11477 MB/s 18.73% 11.15x 5817 MB/s 27.88%
(1 CPU) 1.23x 642 MB/s - 0.71x 642 MB/s -
silesia.tar 11.23x 6520 MB/s 5.9% 5.35x 3109 MB/s 15.88%
(1 CPU) 1.05x 607 MB/s - 0.52x 304 MB/s -
enwik9 19.28x 8440 MB/s 4.04% 9.31x 4076 MB/s 18.04%
(1 CPU) 1.12x 488 MB/s - 0.57x 250 MB/s -
Legend
  • S2 Speed: Speed of S2 compared to Snappy, using 16 cores and 1 core.
  • S2 Throughput: Throughput of S2 in MB/s.
  • S2 % smaller: How many percent of the Snappy output size is S2 better.
  • S2 "better": Speed when enabling "better" compression mode in S2 compared to Snappy.
  • "better" throughput: Speed when enabling "better" compression mode in S2 compared to Snappy.
  • "better" % smaller: How many percent of the Snappy output size is S2 better when using "better" compression.

There is a good speedup across the board when using a single thread and a significant speedup when using multiple threads.

Machine generated data gets by far the biggest compression boost, with size being reduced by up to 35% of Snappy size.

The "better" compression mode sees a good improvement in all cases, but usually at a performance cost.

Incompressible content (sharnd.out.2gb, 2GB random data) sees the smallest speedup. This is likely dominated by synchronization overhead, which is confirmed by the fact that single threaded performance is higher (see above).

Decompression

S2 attempts to create content that is also fast to decompress, except in "better" mode where the smallest representation is used.

S2 vs Snappy decompression speed. Both operating on single core:

File S2 Throughput vs. Snappy Better Throughput vs. Snappy
rawstudio-mint14.tar 2117 MB/s 1.14x 1738 MB/s 0.94x
github-june-2days-2019.json 2401 MB/s 1.25x 2307 MB/s 1.20x
github-ranks-backup.bin 2075 MB/s 0.98x 1764 MB/s 0.83x
consensus.db.10gb 2967 MB/s 1.05x 2885 MB/s 1.02x
adresser.json 4141 MB/s 1.07x 4184 MB/s 1.08x
gob-stream 2264 MB/s 1.12x 2185 MB/s 1.08x
10gb.tar 1525 MB/s 1.03x 1347 MB/s 0.91x
sharnd.out.2gb 3813 MB/s 0.79x 3900 MB/s 0.81x
enwik9 1246 MB/s 1.29x 967 MB/s 1.00x
silesia.tar 1433 MB/s 1.12x 1203 MB/s 0.94x
enwik10 1284 MB/s 1.32x 1010 MB/s 1.04x
Legend
  • S2 Throughput: Decompression speed of S2 encoded content.
  • Better Throughput: Decompression speed of S2 "better" encoded content.
  • vs Snappy: Decompression speed of S2 "better" mode compared to Snappy and absolute speed.

While the decompression code hasn't changed, there is a significant speedup in decompression speed. S2 prefers longer matches and will typically only find matches that are 6 bytes or longer. While this reduces compression a bit, it improves decompression speed.

The "better" compression mode will actively look for shorter matches, which is why it has a decompression speed quite similar to Snappy.

Without assembly decompression is also very fast; single goroutine decompression speed. No assembly:

File S2 Throughput S2 throughput
consensus.db.10gb.s2 1.84x 2289.8 MB/s
10gb.tar.s2 1.30x 867.07 MB/s
rawstudio-mint14.tar.s2 1.66x 1329.65 MB/s
github-june-2days-2019.json.s2 2.36x 1831.59 MB/s
github-ranks-backup.bin.s2 1.73x 1390.7 MB/s
enwik9.s2 1.67x 681.53 MB/s
adresser.json.s2 3.41x 4230.53 MB/s
silesia.tar.s2 1.52x 811.58

Even though S2 typically compresses better than Snappy, decompression speed is always better.

Concurrent Stream Decompression

For full stream decompression S2 offers a DecodeConcurrent that will decode a full stream using multiple goroutines.

Example scaling, AMD Ryzen 3950X, 16 cores, decompression using s2d -bench=3 <input>, best of 3:

Input -cpu=1 -cpu=2 -cpu=4 -cpu=8 -cpu=16
enwik10.snappy 1098.6MB/s 1819.8MB/s 3625.6MB/s 6910.6MB/s 10818.2MB/s
enwik10.s2 1303.5MB/s 2606.1MB/s 4847.9MB/s 8878.4MB/s 9592.1MB/s
sofia-air-quality-dataset.tar.snappy 1302.0MB/s 2165.0MB/s 4244.5MB/s 8241.0MB/s 12920.5MB/s
sofia-air-quality-dataset.tar.s2 1399.2MB/s 2463.2MB/s 5196.5MB/s 9639.8MB/s 11439.5MB/s
sofia-air-quality-dataset.tar.s2 (no asm) 837.5MB/s 1652.6MB/s 3183.6MB/s 5945.0MB/s 9620.7MB/s

Scaling can be expected to be pretty linear until memory bandwidth is saturated.

For now the DecodeConcurrent can only be used for full streams without seeking or combining with regular reads.

Block compression

When compressing blocks no concurrent compression is performed just as Snappy. This is because blocks are for smaller payloads and generally will not benefit from concurrent compression.

An important change is that incompressible blocks will not be more than at most 10 bytes bigger than the input. In rare, worst case scenario Snappy blocks could be significantly bigger than the input.

Mixed content blocks

The most reliable is a wide dataset. For this we use webdevdata.org-2015-01-07-subset, 53927 files, total input size: 4,014,735,833 bytes. Single goroutine used.

* Input Output Reduction MB/s
S2 4014735833 1059723369 73.60% 936.73
S2 Better 4014735833 961580539 76.05% 451.10
S2 Best 4014735833 899182886 77.60% 46.84
Snappy 4014735833 1128706759 71.89% 790.15
S2, Snappy Output 4014735833 1093823291 72.75% 936.60
LZ4 4014735833 1063768713 73.50% 452.02

S2 delivers both the best single threaded throughput with regular mode and the best compression rate with "best". "Better" mode provides the same compression speed as LZ4 with better compression ratio.

When outputting Snappy compatible output it still delivers better throughput (150MB/s more) and better compression.

As can be seen from the other benchmarks decompression should also be easier on the S2 generated output.

Though they cannot be compared due to different decompression speeds here are the speed/size comparisons for other Go compressors:

* Input Output Reduction MB/s
Zstd Fastest (Go) 4014735833 794608518 80.21% 236.04
Zstd Best (Go) 4014735833 704603356 82.45% 35.63
Deflate (Go) l1 4014735833 871294239 78.30% 214.04
Deflate (Go) l9 4014735833 730389060 81.81% 41.17
Standard block compression

Benchmarking single block performance is subject to a lot more variation since it only tests a limited number of file patterns. So individual benchmarks should only be seen as a guideline and the overall picture is more important.

These micro-benchmarks are with data in cache and trained branch predictors. For a more realistic benchmark see the mixed content above.

Block compression. Parallel benchmark running on 16 cores, 16 goroutines.

AMD64 assembly is use for both S2 and Snappy.

Absolute Perf Snappy size S2 Size Snappy Speed S2 Speed Snappy dec S2 dec
html 22843 20868 16246 MB/s 18617 MB/s 40972 MB/s 49263 MB/s
urls.10K 335492 286541 7943 MB/s 10201 MB/s 22523 MB/s 26484 MB/s
fireworks.jpeg 123034 123100 349544 MB/s 303228 MB/s 718321 MB/s 827552 MB/s
fireworks.jpeg (200B) 146 155 8869 MB/s 20180 MB/s 33691 MB/s 52421 MB/s
paper-100k.pdf 85304 84202 167546 MB/s 112988 MB/s 326905 MB/s 291944 MB/s
html_x_4 92234 20870 15194 MB/s 54457 MB/s 30843 MB/s 32217 MB/s
alice29.txt 88034 85934 5936 MB/s 6540 MB/s 12882 MB/s 20044 MB/s
asyoulik.txt 77503 79575 5517 MB/s 6657 MB/s 12735 MB/s 22806 MB/s
lcet10.txt 234661 220383 6235 MB/s 6303 MB/s 14519 MB/s 18697 MB/s
plrabn12.txt 319267 318196 5159 MB/s 6074 MB/s 11923 MB/s 19901 MB/s
geo.protodata 23335 18606 21220 MB/s 25432 MB/s 56271 MB/s 62540 MB/s
kppkn.gtb 69526 65019 9732 MB/s 8905 MB/s 18491 MB/s 18969 MB/s
alice29.txt (128B) 80 82 6691 MB/s 17179 MB/s 31883 MB/s 38874 MB/s
alice29.txt (1000B) 774 774 12204 MB/s 13273 MB/s 48056 MB/s 52341 MB/s
alice29.txt (10000B) 6648 6933 10044 MB/s 12824 MB/s 32378 MB/s 46322 MB/s
alice29.txt (20000B) 12686 13516 7733 MB/s 12160 MB/s 30566 MB/s 58969 MB/s

Speed is generally at or above Snappy. Small blocks gets a significant speedup, although at the expense of size.

Decompression speed is better than Snappy, except in one case.

Since payloads are very small the variance in terms of size is rather big, so they should only be seen as a general guideline.

Size is on average around Snappy, but varies on content type. In cases where compression is worse, it usually is compensated by a speed boost.

Better compression

Benchmarking single block performance is subject to a lot more variation since it only tests a limited number of file patterns. So individual benchmarks should only be seen as a guideline and the overall picture is more important.

Absolute Perf Snappy size Better Size Snappy Speed Better Speed Snappy dec Better dec
html 22843 18972 16246 MB/s 8621 MB/s 40972 MB/s 40292 MB/s
urls.10K 335492 248079 7943 MB/s 5104 MB/s 22523 MB/s 20981 MB/s
fireworks.jpeg 123034 123100 349544 MB/s 84429 MB/s 718321 MB/s 823698 MB/s
fireworks.jpeg (200B) 146 149 8869 MB/s 7125 MB/s 33691 MB/s 30101 MB/s
paper-100k.pdf 85304 82887 167546 MB/s 11087 MB/s 326905 MB/s 198869 MB/s
html_x_4 92234 18982 15194 MB/s 29316 MB/s 30843 MB/s 30937 MB/s
alice29.txt 88034 71611 5936 MB/s 3709 MB/s 12882 MB/s 16611 MB/s
asyoulik.txt 77503 65941 5517 MB/s 3380 MB/s 12735 MB/s 14975 MB/s
lcet10.txt 234661 184939 6235 MB/s 3537 MB/s 14519 MB/s 16634 MB/s
plrabn12.txt 319267 264990 5159 MB/s 2960 MB/s 11923 MB/s 13382 MB/s
geo.protodata 23335 17689 21220 MB/s 10859 MB/s 56271 MB/s 57961 MB/s
kppkn.gtb 69526 55398 9732 MB/s 5206 MB/s 18491 MB/s 16524 MB/s
alice29.txt (128B) 80 78 6691 MB/s 7422 MB/s 31883 MB/s 34225 MB/s
alice29.txt (1000B) 774 746 12204 MB/s 5734 MB/s 48056 MB/s 42068 MB/s
alice29.txt (10000B) 6648 6218 10044 MB/s 6055 MB/s 32378 MB/s 28813 MB/s
alice29.txt (20000B) 12686 11492 7733 MB/s 3143 MB/s 30566 MB/s 27315 MB/s

Except for the mostly incompressible JPEG image compression is better and usually in the double digits in terms of percentage reduction over Snappy.

The PDF sample shows a significant slowdown compared to Snappy, as this mode tries harder to compress the data. Very small blocks are also not favorable for better compression, so throughput is way down.

This mode aims to provide better compression at the expense of performance and achieves that without a huge performance penalty, except on very small blocks.

Decompression speed suffers a little compared to the regular S2 mode, but still manages to be close to Snappy in spite of increased compression.

Best compression mode

S2 offers a "best" compression mode.

This will compress as much as possible with little regard to CPU usage.

Mainly for offline compression, but where decompression speed should still be high and compatible with other S2 compressed data.

Some examples compared on 16 core CPU, amd64 assembly used:

* enwik10
Default... 10000000000 -> 4759950115 [47.60%]; 1.03s, 9263.0MB/s
Better...  10000000000 -> 4084706676 [40.85%]; 2.16s, 4415.4MB/s
Best...    10000000000 -> 3615520079 [36.16%]; 42.259s, 225.7MB/s

* github-june-2days-2019.json
Default... 6273951764 -> 1041700255 [16.60%]; 431ms, 13882.3MB/s
Better...  6273951764 -> 945841238 [15.08%]; 547ms, 10938.4MB/s
Best...    6273951764 -> 826392576 [13.17%]; 9.455s, 632.8MB/s

* nyc-taxi-data-10M.csv
Default... 3325605752 -> 1093516949 [32.88%]; 324ms, 9788.7MB/s
Better...  3325605752 -> 885394158 [26.62%]; 491ms, 6459.4MB/s
Best...    3325605752 -> 773681257 [23.26%]; 8.29s, 412.0MB/s

* 10gb.tar
Default... 10065157632 -> 5915541066 [58.77%]; 1.028s, 9337.4MB/s
Better...  10065157632 -> 5453844650 [54.19%]; 1.597s, 4862.7MB/s
Best...    10065157632 -> 5192495021 [51.59%]; 32.78s, 308.2MB/

* consensus.db.10gb
Default... 10737418240 -> 4549762344 [42.37%]; 882ms, 12118.4MB/s
Better...  10737418240 -> 4438535064 [41.34%]; 1.533s, 3500.9MB/s
Best...    10737418240 -> 4210602774 [39.21%]; 42.96s, 254.4MB/s

Decompression speed should be around the same as using the 'better' compression mode.

Snappy Compatibility

S2 now offers full compatibility with Snappy.

This means that the efficient encoders of S2 can be used to generate fully Snappy compatible output.

There is a snappy package that can be used by simply changing imports from github.com/golang/snappy to github.com/klauspost/compress/snappy. This uses "better" mode for all operations. If you would like more control, you can use the s2 package as described below:

Blocks

Snappy compatible blocks can be generated with the S2 encoder. Compression and speed is typically a bit better MaxEncodedLen is also smaller for smaller memory usage. Replace

Snappy S2 replacement
snappy.Encode(...) s2.EncodeSnappy(...)
snappy.MaxEncodedLen(...) s2.MaxEncodedLen(...)

s2.EncodeSnappy can be replaced with s2.EncodeSnappyBetter or s2.EncodeSnappyBest to get more efficiently compressed snappy compatible output.

s2.ConcatBlocks is compatible with snappy blocks.

Comparison of webdevdata.org-2015-01-07-subset, 53927 files, total input size: 4,014,735,833 bytes. amd64, single goroutine used:

Encoder Size MB/s Reduction
snappy.Encode 1128706759 725.59 71.89%
s2.EncodeSnappy 1093823291 899.16 72.75%
s2.EncodeSnappyBetter 1001158548 578.49 75.06%
s2.EncodeSnappyBest 944507998 66.00 76.47%

Streams

For streams, replace enc = snappy.NewBufferedWriter(w) with enc = s2.NewWriter(w, s2.WriterSnappyCompat()). All other options are available, but note that block size limit is different for snappy.

Comparison of different streams, AMD Ryzen 3950x, 16 cores. Size and throughput:

File snappy.NewWriter S2 Snappy S2 Snappy, Better S2 Snappy, Best
nyc-taxi-data-10M.csv 1316042016 - 539.47MB/s 1307003093 - 10132.73MB/s 1174534014 - 5002.44MB/s 1115904679 - 177.97MB/s
enwik10 (xml) 5088294643 - 451.13MB/s 5175840939 - 9440.69MB/s 4560784526 - 4487.21MB/s 4340299103 - 158.92MB/s
10gb.tar (mixed) 6056946612 - 729.73MB/s 6208571995 - 9978.05MB/s 5741646126 - 4919.98MB/s 5548973895 - 180.44MB/s
github-june-2days-2019.json 1525176492 - 933.00MB/s 1476519054 - 13150.12MB/s 1400547532 - 5803.40MB/s 1321887137 - 204.29MB/s
consensus.db.10gb (db) 5412897703 - 1102.14MB/s 5354073487 - 13562.91MB/s 5335069899 - 5294.73MB/s 5201000954 - 175.72MB/s

Decompression

All decompression functions map directly to equivalent s2 functions.

Snappy S2 replacement
snappy.Decode(...) s2.Decode(...)
snappy.DecodedLen(...) s2.DecodedLen(...)
snappy.NewReader(...) s2.NewReader(...)

Features like quick forward skipping without decompression are also available for Snappy streams.

If you know you are only decompressing snappy streams, setting ReaderMaxBlockSize(64<<10) on your Reader will reduce memory consumption.

Concatenating blocks and streams.

Concatenating streams will concatenate the output of both without recompressing them. While this is inefficient in terms of compression it might be usable in certain scenarios. The 10 byte 'stream identifier' of the second stream can optionally be stripped, but it is not a requirement.

Blocks can be concatenated using the ConcatBlocks function.

Snappy blocks/streams can safely be concatenated with S2 blocks and streams. Streams with indexes (see below) will currently not work on concatenated streams.

Stream Seek Index

S2 and Snappy streams can have indexes. These indexes will allow random seeking within the compressed data.

The index can either be appended to the stream as a skippable block or returned for separate storage.

When the index is appended to a stream it will be skipped by regular decoders, so the output remains compatible with other decoders.

Creating an Index

To automatically add an index to a stream, add WriterAddIndex() option to your writer. Then the index will be added to the stream when Close() is called.

	// Add Index to stream...
	enc := s2.NewWriter(w, s2.WriterAddIndex())
	io.Copy(enc, r)
	enc.Close()

If you want to store the index separately, you can use CloseIndex() instead of the regular Close(). This will return the index. Note that CloseIndex() should only be called once, and you shouldn't call Close().

	// Get index for separate storage... 
	enc := s2.NewWriter(w)
	io.Copy(enc, r)
	index, err := enc.CloseIndex()

The index can then be used needing to read from the stream. This means the index can be used without needing to seek to the end of the stream or for manually forwarding streams. See below.

Finally, an existing S2/Snappy stream can be indexed using the s2.IndexStream(r io.Reader) function.

Using Indexes

To use indexes there is a ReadSeeker(random bool, index []byte) (*ReadSeeker, error) function available.

Calling ReadSeeker will return an io.ReadSeeker compatible version of the reader.

If 'random' is specified the returned io.Seeker can be used for random seeking, otherwise only forward seeking is supported. Enabling random seeking requires the original input to support the io.Seeker interface.

	dec := s2.NewReader(r)
	rs, err := dec.ReadSeeker(false, nil)
	rs.Seek(wantOffset, io.SeekStart)

Get a seeker to seek forward. Since no index is provided, the index is read from the stream. This requires that an index was added and that r supports the io.Seeker interface.

A custom index can be specified which will be used if supplied. When using a custom index, it will not be read from the input stream.

	dec := s2.NewReader(r)
	rs, err := dec.ReadSeeker(false, index)
	rs.Seek(wantOffset, io.SeekStart)

This will read the index from index. Since we specify non-random (forward only) seeking r does not have to be an io.Seeker

	dec := s2.NewReader(r)
	rs, err := dec.ReadSeeker(true, index)
	rs.Seek(wantOffset, io.SeekStart)

Finally, since we specify that we want to do random seeking r must be an io.Seeker.

The returned ReadSeeker contains a shallow reference to the existing Reader, meaning changes performed to one is reflected in the other.

To check if a stream contains an index at the end, the (*Index).LoadStream(rs io.ReadSeeker) error can be used.

Manually Forwarding Streams

Indexes can also be read outside the decoder using the Index type. This can be used for parsing indexes, either separate or in streams.

In some cases it may not be possible to serve a seekable stream. This can for instance be an HTTP stream, where the Range request is sent at the start of the stream.

With a little bit of extra code it is still possible to use indexes to forward to specific offset with a single forward skip.

It is possible to load the index manually like this:

	var index s2.Index
	_, err = index.Load(idxBytes)

This can be used to figure out how much to offset the compressed stream:

	compressedOffset, uncompressedOffset, err := index.Find(wantOffset)

The compressedOffset is the number of bytes that should be skipped from the beginning of the compressed file.

The uncompressedOffset will then be offset of the uncompressed bytes returned when decoding from that position. This will always be <= wantOffset.

When creating a decoder it must be specified that it should not expect a stream identifier at the beginning of the stream. Assuming the io.Reader r has been forwarded to compressedOffset we create the decoder like this:

	dec := s2.NewReader(r, s2.ReaderIgnoreStreamIdentifier())

We are not completely done. We still need to forward the stream the uncompressed bytes we didn't want. This is done using the regular "Skip" function:

	err = dec.Skip(wantOffset - uncompressedOffset)

This will ensure that we are at exactly the offset we want, and reading from dec will start at the requested offset.

Compact storage

For compact storage RemoveIndexHeaders can be used to remove any redundant info from a serialized index. If you remove the header it must be restored before Loading.

This is expected to save 20 bytes. These can be restored using RestoreIndexHeaders. This removes a layer of security, but is the most compact representation. Returns nil if headers contains errors.

Index Format:

Each block is structured as a snappy skippable block, with the chunk ID 0x99.

The block can be read from the front, but contains information so it can be read from the back as well.

Numbers are stored as fixed size little endian values or zigzag encoded base 128 varints, with un-encoded value length of 64 bits, unless other limits are specified.

Content Format
ID, [1]byte Always 0x99.
Data Length, [3]byte 3 byte little-endian length of the chunk in bytes, following this.
Header [6]byte Header, must be [115, 50, 105, 100, 120, 0] or in text: "s2idx\x00".
UncompressedSize, Varint Total Uncompressed size.
CompressedSize, Varint Total Compressed size if known. Should be -1 if unknown.
EstBlockSize, Varint Block Size, used for guessing uncompressed offsets. Must be >= 0.
Entries, Varint Number of Entries in index, must be < 65536 and >=0.
HasUncompressedOffsets byte 0 if no uncompressed offsets are present, 1 if present. Other values are invalid.
UncompressedOffsets, [Entries]VarInt Uncompressed offsets. See below how to decode.
CompressedOffsets, [Entries]VarInt Compressed offsets. See below how to decode.
Block Size, [4]byte Little Endian total encoded size (including header and trailer). Can be used for searching backwards to start of block.
Trailer [6]byte Trailer, must be [0, 120, 100, 105, 50, 115] or in text: "\x00xdi2s". Can be used for identifying block from end of stream.

For regular streams the uncompressed offsets are fully predictable, so HasUncompressedOffsets allows to specify that compressed blocks all have exactly EstBlockSize bytes of uncompressed content.

Entries must be in order, starting with the lowest offset, and there must be no uncompressed offset duplicates.
Entries may point to the start of a skippable block, but it is then not allowed to also have an entry for the next block since that would give an uncompressed offset duplicate.

There is no requirement for all blocks to be represented in the index. In fact there is a maximum of 65536 block entries in an index.

The writer can use any method to reduce the number of entries. An implicit block start at 0,0 can be assumed.

Decoding entries:
// Read Uncompressed entries.
// Each assumes EstBlockSize delta from previous.
for each entry {
    uOff = 0
    if HasUncompressedOffsets == 1 {
        uOff = ReadVarInt // Read value from stream
    }
   
    // Except for the first entry, use previous values.
    if entryNum == 0 {
        entry[entryNum].UncompressedOffset = uOff
        continue
    }
    
    // Uncompressed uses previous offset and adds EstBlockSize
    entry[entryNum].UncompressedOffset = entry[entryNum-1].UncompressedOffset + EstBlockSize + uOff
}


// Guess that the first block will be 50% of uncompressed size.
// Integer truncating division must be used.
CompressGuess := EstBlockSize / 2

// Read Compressed entries.
// Each assumes CompressGuess delta from previous.
// CompressGuess is adjusted for each value.
for each entry {
    cOff = ReadVarInt // Read value from stream
    
    // Except for the first entry, use previous values.
    if entryNum == 0 {
        entry[entryNum].CompressedOffset = cOff
        continue
    }
    
    // Compressed uses previous and our estimate.
    entry[entryNum].CompressedOffset = entry[entryNum-1].CompressedOffset + CompressGuess + cOff
        
     // Adjust compressed offset for next loop, integer truncating division must be used. 
     CompressGuess += cOff/2               
}

To decode from any given uncompressed offset (wantOffset):

  • Iterate entries until entry[n].UncompressedOffset > wantOffset.
  • Start decoding from entry[n-1].CompressedOffset.
  • Discard entry[n-1].UncompressedOffset - wantOffset bytes from the decoded stream.

See using indexes for functions that perform the operations with a simpler interface.

Format Extensions

  • Frame Stream identifier changed from sNaPpY to S2sTwO.
  • Framed compressed blocks can be up to 4MB (up from 64KB).
  • Compressed blocks can have an offset of 0, which indicates to repeat the last seen offset.

Repeat offsets must be encoded as a 2.2.1. Copy with 1-byte offset (01), where the offset is 0.

The length is specified by reading the 3-bit length specified in the tag and decode using this table:

Length Actual Length
0 4
1 5
2 6
3 7
4 8
5 8 + read 1 byte
6 260 + read 2 bytes
7 65540 + read 3 bytes

This allows any repeat offset + length to be represented by 2 to 5 bytes. It also allows to emit matches longer than 64 bytes with one copy + one repeat instead of several 64 byte copies.

Lengths are stored as little endian values.

The first copy of a block cannot be a repeat offset and the offset is reset on every block in streams.

Default streaming block size is 1MB.

LICENSE

This code is based on the Snappy-Go implementation.

Use of this source code is governed by a BSD-style license that can be found in the LICENSE file.

Documentation

Overview

Package s2 implements the S2 compression format.

S2 is an extension of Snappy. Similar to Snappy S2 is aimed for high throughput, which is why it features concurrent compression for bigger payloads.

Decoding is compatible with Snappy compressed content, but content compressed with S2 cannot be decompressed by Snappy.

For more information on Snappy/S2 differences see README in: https://github.com/klauspost/compress/tree/master/s2

There are actually two S2 formats: block and stream. They are related, but different: trying to decompress block-compressed data as a S2 stream will fail, and vice versa. The block format is the Decode and Encode functions and the stream format is the Reader and Writer types.

A "better" compression option is available. This will trade some compression speed

The block format, the more common case, is used when the complete size (the number of bytes) of the original data is known upfront, at the time compression starts. The stream format, also known as the framing format, is for when that isn't always true.

Blocks to not offer much data protection, so it is up to you to add data validation of decompressed blocks.

Streams perform CRC validation of the decompressed data. Stream compression will also be performed on multiple CPU cores concurrently significantly improving throughput.

Index

Examples

Constants

View Source 
const (
	S2IndexHeader  = "s2idx\x00"
	S2IndexTrailer = "\x00xdi2s"
)
View Source 
const (
	ChunkTypeIndex = 0x99
)
View Source 
const MaxBlockSize = math.MaxUint32 - binary.MaxVarintLen32 - 5

MaxBlockSize is the maximum value where MaxEncodedLen will return a valid block size. Blocks this big are highly discouraged, though.

Variables

View Source 
var (
	// ErrCorrupt reports that the input is invalid.
	ErrCorrupt = errors.New("s2: corrupt input")
	// ErrCRC reports that the input failed CRC validation (streams only)
	ErrCRC = errors.New("s2: corrupt input, crc mismatch")
	// ErrTooLarge reports that the uncompressed length is too large.
	ErrTooLarge = errors.New("s2: decoded block is too large")
	// ErrUnsupported reports that the input isn't supported.
	ErrUnsupported = errors.New("s2: unsupported input")
)

Functions

func ConcatBlocks 

func ConcatBlocks(dst []byte, blocks ...[]byte) ([]byte, error)

ConcatBlocks will concatenate the supplied blocks and append them to the supplied destination. If the destination is nil or too small, a new will be allocated. The blocks are not validated, so garbage in = garbage out. dst may not overlap block data. Any data in dst is preserved as is, so it will not be considered a block.

func Decode 

func Decode(dst, src []byte) ([]byte, error)

Decode returns the decoded form of src. The returned slice may be a sub- slice of dst if dst was large enough to hold the entire decoded block. Otherwise, a newly allocated slice will be returned.

The dst and src must not overlap. It is valid to pass a nil dst.

func DecodedLen 

func DecodedLen(src []byte) (int, error)

DecodedLen returns the length of the decoded block.

func Encode 

func Encode(dst, src []byte) []byte

Encode returns the encoded form of src. The returned slice may be a sub- slice of dst if dst was large enough to hold the entire encoded block. Otherwise, a newly allocated slice will be returned.

The dst and src must not overlap. It is valid to pass a nil dst.

The blocks will require the same amount of memory to decode as encoding, and does not make for concurrent decoding. Also note that blocks do not contain CRC information, so corruption may be undetected.

If you need to encode larger amounts of data, consider using the streaming interface which gives all of these features.

func EncodeBest added in v1.11.7 

func EncodeBest(dst, src []byte) []byte

EncodeBest returns the encoded form of src. The returned slice may be a sub- slice of dst if dst was large enough to hold the entire encoded block. Otherwise, a newly allocated slice will be returned.

EncodeBest compresses as good as reasonably possible but with a big speed decrease.

The dst and src must not overlap. It is valid to pass a nil dst.

The blocks will require the same amount of memory to decode as encoding, and does not make for concurrent decoding. Also note that blocks do not contain CRC information, so corruption may be undetected.

If you need to encode larger amounts of data, consider using the streaming interface which gives all of these features.

func EncodeBetter 

func EncodeBetter(dst, src []byte) []byte

EncodeBetter returns the encoded form of src. The returned slice may be a sub- slice of dst if dst was large enough to hold the entire encoded block. Otherwise, a newly allocated slice will be returned.

EncodeBetter compresses better than Encode but typically with a 10-40% speed decrease on both compression and decompression.

The dst and src must not overlap. It is valid to pass a nil dst.

The blocks will require the same amount of memory to decode as encoding, and does not make for concurrent decoding. Also note that blocks do not contain CRC information, so corruption may be undetected.

If you need to encode larger amounts of data, consider using the streaming interface which gives all of these features.

func EncodeSnappy added in v1.10.0 

func EncodeSnappy(dst, src []byte) []byte

EncodeSnappy returns the encoded form of src. The returned slice may be a sub- slice of dst if dst was large enough to hold the entire encoded block. Otherwise, a newly allocated slice will be returned.

The output is Snappy compatible and will likely decompress faster.

The dst and src must not overlap. It is valid to pass a nil dst.

The blocks will require the same amount of memory to decode as encoding, and does not make for concurrent decoding. Also note that blocks do not contain CRC information, so corruption may be undetected.

If you need to encode larger amounts of data, consider using the streaming interface which gives all of these features.

func EncodeSnappyBest added in v1.13.1 

func EncodeSnappyBest(dst, src []byte) []byte

EncodeSnappyBest returns the encoded form of src. The returned slice may be a sub- slice of dst if dst was large enough to hold the entire encoded block. Otherwise, a newly allocated slice will be returned.

The output is Snappy compatible and will likely decompress faster.

The dst and src must not overlap. It is valid to pass a nil dst.

The blocks will require the same amount of memory to decode as encoding, and does not make for concurrent decoding. Also note that blocks do not contain CRC information, so corruption may be undetected.

If you need to encode larger amounts of data, consider using the streaming interface which gives all of these features.

func EncodeSnappyBetter added in v1.13.1 

func EncodeSnappyBetter(dst, src []byte) []byte

EncodeSnappyBetter returns the encoded form of src. The returned slice may be a sub- slice of dst if dst was large enough to hold the entire encoded block. Otherwise, a newly allocated slice will be returned.

The output is Snappy compatible and will likely decompress faster.

The dst and src must not overlap. It is valid to pass a nil dst.

The blocks will require the same amount of memory to decode as encoding, and does not make for concurrent decoding. Also note that blocks do not contain CRC information, so corruption may be undetected.

If you need to encode larger amounts of data, consider using the streaming interface which gives all of these features.

func IndexStream added in v1.14.0 

func IndexStream(r io.Reader) ([]byte, error)

IndexStream will return an index for a stream. The stream structure will be checked, but data within blocks is not verified. The returned index can either be appended to the end of the stream or stored separately.

func MaxEncodedLen 

func MaxEncodedLen(srcLen int) int

MaxEncodedLen returns the maximum length of a snappy block, given its uncompressed length.

It will return a negative value if srcLen is too large to encode. 32 bit platforms will have lower thresholds for rejecting big content.

func RemoveIndexHeaders added in v1.15.8 

func RemoveIndexHeaders(b []byte) []byte

RemoveIndexHeaders will trim all headers and trailers from a given index. This is expected to save 20 bytes. These can be restored using RestoreIndexHeaders. This removes a layer of security, but is the most compact representation. Returns nil if headers contains errors. The returned slice references the provided slice.

func RestoreIndexHeaders added in v1.15.8 

func RestoreIndexHeaders(in []byte) []byte

RestoreIndexHeaders will index restore headers removed by RemoveIndexHeaders. No error checking is performed on the input. If a 0 length slice is sent, it is returned without modification.

Types

type ErrCantSeek added in v1.14.0 

type ErrCantSeek struct {
	Reason string
}

ErrCantSeek is returned if the stream cannot be seeked.

func (ErrCantSeek) Error added in v1.14.0 

func (e ErrCantSeek) Error() string

Error returns the error as string.

type Index added in v1.14.0 

type Index struct {
	TotalUncompressed int64 // Total Uncompressed size if known. Will be -1 if unknown.
	TotalCompressed   int64 // Total Compressed size if known. Will be -1 if unknown.
	// contains filtered or unexported fields
}

Index represents an S2/Snappy index.

func (*Index) Find added in v1.14.0 

func (i *Index) Find(offset int64) (compressedOff, uncompressedOff int64, err error)

Find the offset at or before the wanted (uncompressed) offset. If offset is 0 or positive it is the offset from the beginning of the file. If the uncompressed size is known, the offset must be within the file. If an offset outside the file is requested io.ErrUnexpectedEOF is returned. If the offset is negative, it is interpreted as the distance from the end of the file, where -1 represents the last byte. If offset from the end of the file is requested, but size is unknown, ErrUnsupported will be returned.

func (*Index) JSON added in v1.14.0 

func (i *Index) JSON() []byte

JSON returns the index as JSON text.

func (*Index) Load added in v1.14.0 

func (i *Index) Load(b []byte) ([]byte, error)

Load a binary index. A zero value Index can be used or a previous one can be reused.

Example 
package main

import (
	"bytes"
	"fmt"
	"io"
	"math/rand"
	"sync"

	"github.com/klauspost/compress/s2"
)

func main() {
	fatalErr := func(err error) {
		if err != nil {
			panic(err)
		}
	}

	// Create a test corpus
	tmp := make([]byte, 5<<20)
	rng := rand.New(rand.NewSource(0xbeefcafe))
	rng.Read(tmp)
	// Make it compressible...
	for i, v := range tmp {
		tmp[i] = '0' + v&3
	}
	// Compress it...
	var buf bytes.Buffer
	// We use smaller blocks just for the example...
	enc := s2.NewWriter(&buf, s2.WriterBlockSize(100<<10))
	err := enc.EncodeBuffer(tmp)
	fatalErr(err)

	// Close and get index...
	idxBytes, err := enc.CloseIndex()
	fatalErr(err)

	// This is our compressed stream...
	compressed := buf.Bytes()

	var once sync.Once
	for wantOffset := int64(0); wantOffset < int64(len(tmp)); wantOffset += 555555 {
		// Let's assume we want to read from uncompressed offset 'i'
		// and we cannot seek in input, but we have the index.
		want := tmp[wantOffset:]

		// Load the index.
		var index s2.Index
		_, err = index.Load(idxBytes)
		fatalErr(err)

		// Find offset in file:
		compressedOffset, uncompressedOffset, err := index.Find(wantOffset)
		fatalErr(err)

		// Offset the input to the compressed offset.
		// Notice how we do not provide any bytes before the offset.
		input := io.Reader(bytes.NewBuffer(compressed[compressedOffset:]))
		if _, ok := input.(io.Seeker); !ok {
			// Notice how the input cannot be seeked...
			once.Do(func() {
				fmt.Println("Input does not support seeking...")
			})
		} else {
			panic("did you implement seeking on bytes.Buffer?")
		}

		// When creating the decoder we must specify that it should not
		// expect a stream identifier at the beginning og the frame.
		dec := s2.NewReader(input, s2.ReaderIgnoreStreamIdentifier())

		// We now have a reader, but it will start outputting at uncompressedOffset,
		// and not the actual offset we want, so skip forward to that.
		toSkip := wantOffset - uncompressedOffset
		err = dec.Skip(toSkip)
		fatalErr(err)

		// Read the rest of the stream...
		got, err := io.ReadAll(dec)
		fatalErr(err)
		if bytes.Equal(got, want) {
			fmt.Println("Successfully skipped forward to", wantOffset)
		} else {
			fmt.Println("Failed to skip forward to", wantOffset)
		}
	}
}

Output:

Input does not support seeking...
Successfully skipped forward to 0
Successfully skipped forward to 555555
Successfully skipped forward to 1111110
Successfully skipped forward to 1666665
Successfully skipped forward to 2222220
Successfully skipped forward to 2777775
Successfully skipped forward to 3333330
Successfully skipped forward to 3888885
Successfully skipped forward to 4444440
Successfully skipped forward to 4999995

func (*Index) LoadStream added in v1.14.0 

func (i *Index) LoadStream(rs io.ReadSeeker) error

LoadStream will load an index from the end of the supplied stream. ErrUnsupported will be returned if the signature cannot be found. ErrCorrupt will be returned if unexpected values are found. io.ErrUnexpectedEOF is returned if there are too few bytes. IO errors are returned as-is.

type ReadSeeker added in v1.14.0 

type ReadSeeker struct {
	*Reader
}

ReadSeeker provides random or forward seeking in compressed content. See Reader.ReadSeeker

func (*ReadSeeker) Seek added in v1.14.0 

func (r *ReadSeeker) Seek(offset int64, whence int) (int64, error)

Seek allows seeking in compressed data.

type Reader 

type Reader struct {
	// contains filtered or unexported fields
}

Reader is an io.Reader that can read Snappy-compressed bytes.

func NewReader 

func NewReader(r io.Reader, opts ...ReaderOption) *Reader

NewReader returns a new Reader that decompresses from r, using the framing format described at https://github.com/google/snappy/blob/master/framing_format.txt with S2 changes.

func (*Reader) DecodeConcurrent added in v1.15.5 

func (r *Reader) DecodeConcurrent(w io.Writer, concurrent int) (written int64, err error)

DecodeConcurrent will decode the full stream to w. This function should not be combined with reading, seeking or other operations. Up to 'concurrent' goroutines will be used. If <= 0, runtime.NumCPU will be used. On success the number of bytes decompressed nil and is returned. This is mainly intended for bigger streams.

func (*Reader) Read 

func (r *Reader) Read(p []byte) (int, error)

Read satisfies the io.Reader interface.

func (*Reader) ReadByte added in v1.13.4 

func (r *Reader) ReadByte() (byte, error)

ReadByte satisfies the io.ByteReader interface.

func (*Reader) ReadSeeker added in v1.14.0 

func (r *Reader) ReadSeeker(random bool, index []byte) (*ReadSeeker, error)

ReadSeeker will return an io.ReadSeeker compatible version of the reader. If 'random' is specified the returned io.Seeker can be used for random seeking, otherwise only forward seeking is supported. Enabling random seeking requires the original input to support the io.Seeker interface. A custom index can be specified which will be used if supplied. When using a custom index, it will not be read from the input stream. The returned ReadSeeker contains a shallow reference to the existing Reader, meaning changes performed to one is reflected in the other.

func (*Reader) Reset 

func (r *Reader) Reset(reader io.Reader)

Reset discards any buffered data, resets all state, and switches the Snappy reader to read from r. This permits reusing a Reader rather than allocating a new one.

func (*Reader) Skip 

func (r *Reader) Skip(n int64) error

Skip will skip n bytes forward in the decompressed output. For larger skips this consumes less CPU and is faster than reading output and discarding it. CRC is not checked on skipped blocks. io.ErrUnexpectedEOF is returned if the stream ends before all bytes have been skipped. If a decoding error is encountered subsequent calls to Read will also fail.

func (*Reader) SkippableCB added in v1.14.0 

func (r *Reader) SkippableCB(id uint8, fn func(r io.Reader) error) error

SkippableCB will register a callback for chunks with the specified ID. ID must be a Reserved skippable chunks ID, 0x80-0xfe (inclusive). For each chunk with the ID, the callback is called with the content. Any returned non-nil error will abort decompression. Only one callback per ID is supported, latest sent will be used. Sending a nil function will disable previous callbacks.

type ReaderOption added in v1.11.7 

type ReaderOption func(*Reader) error

ReaderOption is an option for creating a decoder.

func ReaderAllocBlock added in v1.11.8 

func ReaderAllocBlock(blockSize int) ReaderOption

ReaderAllocBlock allows to control upfront stream allocations and not allocate for frames bigger than this initially. If frames bigger than this is seen a bigger buffer will be allocated.

Default is 1MB, which is default output size.

func ReaderIgnoreCRC added in v1.15.6 

func ReaderIgnoreCRC() ReaderOption

ReaderIgnoreCRC will make the reader skip CRC calculation and checks.

func ReaderIgnoreStreamIdentifier added in v1.14.0 

func ReaderIgnoreStreamIdentifier() ReaderOption

ReaderIgnoreStreamIdentifier will make the reader skip the expected stream identifier at the beginning of the stream. This can be used when serving a stream that has been forwarded to a specific point.

func ReaderMaxBlockSize added in v1.11.7 

func ReaderMaxBlockSize(blockSize int) ReaderOption

ReaderMaxBlockSize allows to control allocations if the stream has been compressed with a smaller WriterBlockSize, or with the default 1MB. Blocks must be this size or smaller to decompress, otherwise the decoder will return ErrUnsupported.

For streams compressed with Snappy this can safely be set to 64KB (64 << 10).

Default is the maximum limit of 4MB.

func ReaderSkippableCB added in v1.14.0 

func ReaderSkippableCB(id uint8, fn func(r io.Reader) error) ReaderOption

ReaderSkippableCB will register a callback for chuncks with the specified ID. ID must be a Reserved skippable chunks ID, 0x80-0xfd (inclusive). For each chunk with the ID, the callback is called with the content. Any returned non-nil error will abort decompression. Only one callback per ID is supported, latest sent will be used.

type Writer 

type Writer struct {
	// contains filtered or unexported fields
}

Writer is an io.Writer that can write Snappy-compressed bytes.

func NewWriter 

func NewWriter(w io.Writer, opts ...WriterOption) *Writer

NewWriter returns a new Writer that compresses to w, using the framing format described at https://github.com/google/snappy/blob/master/framing_format.txt

Users must call Close to guarantee all data has been forwarded to the underlying io.Writer and that resources are released. They may also call Flush zero or more times before calling Close.

func (*Writer) AddSkippableBlock added in v1.14.0 

func (w *Writer) AddSkippableBlock(id uint8, data []byte) (err error)

AddSkippableBlock will add a skippable block to the stream. The ID must be 0x80-0xfe (inclusive). Length of the skippable block must be <= 16777215 bytes.

func (*Writer) Close 

func (w *Writer) Close() error

Close calls Flush and then closes the Writer. Calling Close multiple times is ok, but calling CloseIndex after this will make it not return the index.

func (*Writer) CloseIndex added in v1.14.0 

func (w *Writer) CloseIndex() ([]byte, error)

CloseIndex calls Close and returns an index on first call. This is not required if you are only adding index to a stream.

func (*Writer) EncodeBuffer added in v1.10.0 

func (w *Writer) EncodeBuffer(buf []byte) (err error)

EncodeBuffer will add a buffer to the stream. This is the fastest way to encode a stream, but the input buffer cannot be written to by the caller until Flush or Close has been called when concurrency != 1.

If you cannot control that, use the regular Write function.

Note that input is not buffered. This means that each write will result in discrete blocks being created. For buffered writes, use the regular Write function.

func (*Writer) Flush 

func (w *Writer) Flush() error

Flush flushes the Writer to its underlying io.Writer. This does not apply padding.

func (*Writer) ReadFrom 

func (w *Writer) ReadFrom(r io.Reader) (n int64, err error)

ReadFrom implements the io.ReaderFrom interface. Using this is typically more efficient since it avoids a memory copy. ReadFrom reads data from r until EOF or error. The return value n is the number of bytes read. Any error except io.EOF encountered during the read is also returned.

func (*Writer) Reset 

func (w *Writer) Reset(writer io.Writer)

Reset discards the writer's state and switches the Snappy writer to write to w. This permits reusing a Writer rather than allocating a new one.

func (*Writer) Write 

func (w *Writer) Write(p []byte) (nRet int, errRet error)

Write satisfies the io.Writer interface.

type WriterOption 

type WriterOption func(*Writer) error

WriterOption is an option for creating a encoder.

func WriterAddIndex added in v1.14.0 

func WriterAddIndex() WriterOption

WriterAddIndex will append an index to the end of a stream when it is closed.

func WriterBestCompression added in v1.11.7 

func WriterBestCompression() WriterOption

WriterBestCompression will enable better compression. EncodeBetter compresses better than Encode but typically with a big speed decrease on compression.

func WriterBetterCompression 

func WriterBetterCompression() WriterOption

WriterBetterCompression will enable better compression. EncodeBetter compresses better than Encode but typically with a 10-40% speed decrease on both compression and decompression.

func WriterBlockSize 

func WriterBlockSize(n int) WriterOption

WriterBlockSize allows to override the default block size. Blocks will be this size or smaller. Minimum size is 4KB and and maximum size is 4MB.

Bigger blocks may give bigger throughput on systems with many cores, and will increase compression slightly, but it will limit the possible concurrency for smaller payloads for both encoding and decoding. Default block size is 1MB.

When writing Snappy compatible output using WriterSnappyCompat, the maximum block size is 64KB.

func WriterConcurrency 

func WriterConcurrency(n int) WriterOption

WriterConcurrency will set the concurrency, meaning the maximum number of decoders to run concurrently. The value supplied must be at least 1. By default this will be set to GOMAXPROCS.

func WriterFlushOnWrite added in v1.13.4 

func WriterFlushOnWrite() WriterOption

WriterFlushOnWrite will compress blocks on each call to the Write function.

This is quite inefficient as blocks size will depend on the write size.

Use WriterConcurrency(1) to also make sure that output is flushed. When Write calls return, otherwise they will be written when compression is done.

func WriterPadding 

func WriterPadding(n int) WriterOption

WriterPadding will add padding to all output so the size will be a multiple of n. This can be used to obfuscate the exact output size or make blocks of a certain size. The contents will be a skippable frame, so it will be invisible by the decoder. n must be > 0 and <= 4MB. The padded area will be filled with data from crypto/rand.Reader. The padding will be applied whenever Close is called on the writer.

func WriterPaddingSrc added in v1.11.4 

func WriterPaddingSrc(reader io.Reader) WriterOption

WriterPaddingSrc will get random data for padding from the supplied source. By default crypto/rand is used.

func WriterSnappyCompat added in v1.13.1 

func WriterSnappyCompat() WriterOption

WriterSnappyCompat will write snappy compatible output. The output can be decompressed using either snappy or s2. If block size is more than 64KB it is set to that.

func WriterUncompressed added in v1.11.4 

func WriterUncompressed() WriterOption

WriterUncompressed will bypass compression. The stream will be written as uncompressed blocks only. If concurrency is > 1 CRC and output will still be done async.

Source Files

View all Source files

Directories

Path Synopsis
_generate module
cmd
internal/filepathx
Package filepathx adds double-star globbing support to the Glob function from the core path/filepath package.
 Package filepathx adds double-star globbing support to the Glob function from the core path/filepath package.
internal/readahead
Package readahead will do asynchronous read-ahead from an input io.Reader and make the data available as an io.Reader.
 Package readahead will do asynchronous read-ahead from an input io.Reader and make the data available as an io.Reader.
s2c
s2d

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.