Struct q_compress::Compressor

pub struct Compressor<T> where
    T: NumberLike,  { /* private fields */ }

Converts vectors of numbers into compressed bytes.

You can use the compressor very easily:

use q_compress::Compressor;

let my_nums = vec![1, 2, 3];
let compressor = Compressor::<i32>::default();
let bytes = compressor.simple_compress(&my_nums);

You can also get full control over the compression process:

use q_compress::{BitWriter, Compressor, CompressorConfig};

let compressor = Compressor::<i32>::from_config(CompressorConfig {
  compression_level: 5,
  ..Default::default()
});
let mut writer = BitWriter::default();

compressor.header(&mut writer).expect("header failure");
let chunk_0 = vec![1, 2, 3];
compressor.chunk(&chunk_0, &mut writer).expect("chunk failure");
let chunk_1 = vec![4, 5];
compressor.chunk(&chunk_1, &mut writer).expect("chunk failure");
compressor.footer(&mut writer).expect("footer failure");

let bytes = writer.pop();

Note that in practice we would need larger chunks than this to achieve good compression, preferably containing 10k-10M numbers.

Implementations

impl<T> Compressor<T> where
    T: NumberLike, 

pub fn from_config(config: CompressorConfig) -> Self

Creates a new compressor, given a CompressorConfig. Internally, the compressor builds Flags as well as an internal configuration that doesn’t show up in the output file. You can inspect the flags it chooses with .flags().

pub fn flags(&self) -> &Flags

Returns a reference to the compressor’s flags.

pub fn header(&self, writer: &mut BitWriter) -> QCompressResult<()>

Writes out a header using the compressor’s data type and flags. Will return an error if the writer is not at a byte-aligned position.

Each .qco file must start with such a header, which contains:

• a 4-byte magic header for “qco!” in ascii,
• a byte for the data type (e.g. i64 has byte 1 and f64 has byte 5), and
• bytes for the flags used to compress.

pub fn chunk(
    &self,
    nums: &[T],
    writer: &mut BitWriter
) -> QCompressResult<ChunkMetadata<T>>

Writes out a chunk of data representing the provided numbers. Will return an error if the writer is not at a byte-aligned position or the slice of numbers is empty.

Each chunk contains a ChunkMetadata section followed by the chunk body. The chunk body encodes the numbers passed in here.

pub fn footer(&self, writer: &mut BitWriter) -> QCompressResult<()>

Writes out a single footer byte indicating that the .qco file has ended. Will return an error if the writer is not byte-aligned.

pub fn simple_compress(&self, nums: &[T]) -> Vec<u8>

Takes in a slice of numbers and returns compressed bytes.

Trait Implementations

impl<T: Clone> Clone for Compressor<T> where
    T: NumberLike, 

fn clone(&self) -> Compressor<T>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<T: Debug> Debug for Compressor<T> where
    T: NumberLike, 

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<T: NumberLike> Default for Compressor<T>

fn default() -> Self

Returns the “default value” for a type.