pcodec
FiniteStateEntropy
pcodec | FiniteStateEntropy | |
---|---|---|
19 | 4 | |
248 | 1,263 | |
- | - | |
8.8 | 0.0 | |
2 days ago | over 1 year ago | |
Rust | C | |
Apache License 2.0 | BSD 2-clause "Simplified" License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
pcodec
-
Learnings from making things fast
Context: I've been iterating on my side project pcodec (a codec for columns of numerical data) and have gradually improved decompression speed from ~150MB/s to ~1GB/s. Not everything here is novel or Rust-specific, but here's what I've learned in the process:
- Compressing bytes?
-
Worries about tANS?
For context: I'm creating an experimental successor to my library Quantile Compression, which does good compression for numerical sequences and has several users. I have a variable number of symbols which may be as high as 212 in some cases, but is ~26 in most cases. The data is typically 216 to 224 tokens long.
-
Quantile Compression, a compression format for numerical data that improves compression ratio by ~30% over alternatives
I'm not a member, but you can use the CLI to try it out pretty easily: https://github.com/mwlon/quantile-compression/tree/main/q_compress_cli . Let me know how it does
-
I built Quantile Compression, which could make all our numerical columnar data 25% smaller.
You can try it out very easily with the CLI which works on CSV and Parquet columns now, e.g. cargo run --release compress --csv my.csv --col-name my_column out.qco
-
Quantile Compression: 35% higher compression ratio for numeric sequences than any other compressor
Right, please don't try to use it for general files. It looks like zpaq is kinda hard to set up except on windows, so I'm probably not going to, but I encourage you to try it out! There's an example you can use to generate a bunch of random numerical distributions, outputting binary files, .qco, and other formats.
- Q_compress: Lossless compressor and decompressor for numerical data
-
q_compress 0.7: still has 35% higher compression ratio than .zstd.parquet for numerical sequences, now with delta encoding and 2x faster than before
Here's how you can generate benchmark data, including binary files: https://github.com/mwlon/quantile-compression/blob/main/q_compress/examples/primary.md
-
Quantile Compression, a format and algorithm for numerical sequences offering 35% higher compression ratio than .zstd.parquet.
I made a simple CLI for compressing and inspecting .qco files. Not available on package managers yet, but it's still pretty easy to try out: https://github.com/mwlon/quantile-compression/blob/main/CLI.md
- Quantile Compression (q-compress), a new compression format and rust library that shrinks real-world columns of numerical data 10-40% smaller than other methods
FiniteStateEntropy
-
Intel QuickAssist Technology Zstandard Plugin for Zstandard
It's obsolete. It's limited to 32KB LZ window with huffman coding. Zstd can use a much larger window (8MB recommended) and a much better entropy coder: https://github.com/Cyan4973/FiniteStateEntropy
-
Worries about tANS?
tANS block based : FSE
-
Silly Lossy Text Compression Idea
Sounds similar to: https://github.com/Cyan4973/FiniteStateEntropy
https://arxiv.org/abs/1311.2540
> The modern data compression is mainly based on two approaches to entropy coding: Huffman (HC) and arithmetic/range coding (AC). The former is much faster, but approximates probabilities with powers of 2, usually leading to relatively low compression rates. The latter uses nearly exact probabilities - easily approaching theoretical compression rate limit (Shannon entropy), but at cost of much larger computational cost.
-
C Deep
FiniteStateEntropy - Two highly efficient compression codecs optimized for modern CPUs. BSD-2-Clause
What are some alternatives?
ans-large-alphabet - Large-Alphabet Semi-Static Entropy Coding Via Asymmetric Numeral Systems
Snappy - A fast compressor/decompressor
encoding - Integer Compression Libraries for Go
zstd - Zstandard - Fast real-time compression algorithm
x3-rust - X3 Lossless Audio Compression for Rust
zlib-ng - zlib replacement with optimizations for "next generation" systems.
ryg_rans - Simple rANS encoder/decoder (arithmetic coding-ish entropy coder).
brotli - Brotli compression format
gdal - GDAL is an open source MIT licensed translator library for raster and vector geospatial data formats.
LZFSE - LZFSE compression library and command line tool
spark-pancake-connector - support for the "pancake" format in Spark
LZMA - (Unofficial) Git mirror of LZMA SDK releases