Awesome

pzip

pzip, short for parallel-zip, is a blazing fast concurrent zip archiver and extractor.

Features

• Archives files and directories into a valid zip archive, using DEFLATE.
• Preserves modification times of files.
• Files are read and compressed concurrently

Installation

Command Line

For command-line usage, we provide two binaries which can be installed separately:

• pzip- concurrent zip archiving
• punzip- concurrent zip extraction

To install, run:

macOS

For zip archiving: brew install ybirader/pzip/pzip

For zip extraction: brew install ybirader/pzip/punzip

Debian, Ubuntu, Raspbian

For the latest stable release:

curl -1sLf 'https://dl.cloudsmith.io/public/pzip/stable/setup.deb.sh' | sudo -E bash
sudo apt update
sudo apt install pzip

curl -1sLf 'https://dl.cloudsmith.io/public/pzip/stable/setup.deb.sh' | sudo -E bash
sudo apt update
sudo apt install punzip

Go

Alternatively, if you have Go installed:

go install github.com/ybirader/pzip

Build from source

To build from source, we require Go 1.21 or newer.

1. Clone the repository by running git clone "https://github.com/ybirader/pzip.git"
2. Build both pzip and punzip by running make build or build separately via cd cmd/pzip && go build and cd cmd/punzip && go build

Usage

Archiving

pzip's API is similar to that of the standard zip utlity found on most *-nix systems.

pzip /path/to/compressed.zip path/to/file_or_directory1 path/to/file_or_directory2 ... path/to/file_or_directoryN

Alternatively, pzip can be imported as a package

archive, err := os.Create("archive.zip")
if err != nil {
  log.Fatal(err)
}

archiver, err := pzip.NewArchiver(archive)
if err != nil {
  log.Fatal(err)
}
defer archiver.Close()

files := []string{ "./hello", "./hello.txt", "./bye.md" }

err = archiver.Archive(context.Background(), files)
if err != nil {
  log.Fatal(err)
}

The concurrency of the archiver can be configured using the corresponding flag:

pzip --concurrency 2 /path/to/compressed.zip path/to/file_or_directory1 path/to/file_or_directory2 ... path/to/file_or_directoryN

or by passing the ArchiverConcurrency option:

archiver, err := pzip.NewArchiver(archive, ArchiverConcurrency(2))

Extraction

punzip's API is similar to that of the standard unzip utlity found on most *-nix systems.

punzip /path/to/compressed.zip

By default, punzip extracts into the current directory. We can extract to a particular path by:

punzip -d /path/to/output /path/to/compressed.zip

Using the Go package, we have:

outputDirPath := "./output"
archivePath := "./archive.zip"

extractor, err := pzip.NewExtractor(outputDirPath)
if err != nil {
  log.Fatal(err)
}
defer extractor.Close()

err = extractor.Extract(context.Background(), archivePath)
if err != nil {
  log.Fatal(err)
}

As with pzip, we can configure the concurrency of the extractor using:

punzip --concurrency 2 /path/to/compressed.zip

Similarly, with the Go package, we pass in the ExtractorConcurrency option:

extractor, err := pzip.NewExtractor(outputDirPath, ExtractorConcurrency(2))

Benchmarks

pzip was benchmarked using Matt Mahoney's sample directory.

Using the standard zip utlity, we get the following time to archive:

real    14m31.809s
user    13m12.833s
sys     0m24.193s

Running the same benchmark with pzip, we find that:

real    0m56.851s
user    3m32.619s
sys     1m25.040s

Contributing

To contribute to pzip, first submit or comment in an issue to discuss your contribution, then open a pull request (PR).

License

pzip is released under the Apache 2.0 license.

Acknowledgements

Many thanks to the folks at Cloudsmith for graciously providing Debian package hosting. Cloudsmith is the only fully hosted, cloud-native, universal package management solution, that enables your organization to create, store and share packages in any format, to any place, with total confidence.