use crate::CompressionType;
use arrow_buffer::Buffer;
use arrow_schema::ArrowError;
const LENGTH_NO_COMPRESSED_DATA: i64 = -1;
const LENGTH_OF_PREFIX_DATA: i64 = 8;
#[derive(Default)]
pub struct CompressionContext {
#[cfg(feature = "zstd")]
compressor: Option<zstd::bulk::Compressor<'static>>,
}
impl CompressionContext {
#[cfg(feature = "zstd")]
fn zstd_compressor(&mut self) -> &mut zstd::bulk::Compressor<'static> {
self.compressor.get_or_insert_with(|| {
zstd::bulk::Compressor::new(zstd::DEFAULT_COMPRESSION_LEVEL)
.expect("can use default compression level")
})
}
}
impl std::fmt::Debug for CompressionContext {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
let mut ds = f.debug_struct("CompressionContext");
#[cfg(feature = "zstd")]
ds.field(
"compressor",
&self.compressor.as_ref().map(|_| "zstd::bulk::Compressor"),
);
ds.finish()
}
}
pub struct DecompressionContext {
#[cfg(feature = "zstd")]
decompressor: Option<zstd::bulk::Decompressor<'static>>,
}
impl DecompressionContext {
pub(crate) fn new() -> Self {
Default::default()
}
#[cfg(feature = "zstd")]
fn zstd_decompressor(&mut self) -> &mut zstd::bulk::Decompressor<'static> {
self.decompressor.get_or_insert_with(|| {
zstd::bulk::Decompressor::new().expect("can create zstd decompressor")
})
}
}
#[allow(clippy::derivable_impls)]
impl Default for DecompressionContext {
fn default() -> Self {
DecompressionContext {
#[cfg(feature = "zstd")]
decompressor: None,
}
}
}
impl std::fmt::Debug for DecompressionContext {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
let mut ds = f.debug_struct("DecompressionContext");
#[cfg(feature = "zstd")]
ds.field(
"decompressor",
&self
.decompressor
.as_ref()
.map(|_| "zstd::bulk::Decompressor"),
);
ds.finish()
}
}
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum CompressionCodec {
Lz4Frame,
Zstd,
}
impl TryFrom<CompressionType> for CompressionCodec {
type Error = ArrowError;
fn try_from(compression_type: CompressionType) -> Result<Self, ArrowError> {
match compression_type {
CompressionType::ZSTD => Ok(CompressionCodec::Zstd),
CompressionType::LZ4_FRAME => Ok(CompressionCodec::Lz4Frame),
other_type => Err(ArrowError::NotYetImplemented(format!(
"compression type {other_type:?} not supported "
))),
}
}
}
impl CompressionCodec {
pub(crate) fn compress_to_vec(
&self,
input: &[u8],
output: &mut Vec<u8>,
context: &mut CompressionContext,
) -> Result<usize, ArrowError> {
let uncompressed_data_len = input.len();
let original_output_len = output.len();
if input.is_empty() {
} else {
output.extend_from_slice(&uncompressed_data_len.to_le_bytes());
self.compress(input, output, context)?;
let compression_len = output.len() - original_output_len;
if compression_len > uncompressed_data_len {
output.truncate(original_output_len);
output.extend_from_slice(&LENGTH_NO_COMPRESSED_DATA.to_le_bytes());
output.extend_from_slice(input);
}
}
Ok(output.len() - original_output_len)
}
pub(crate) fn decompress_to_buffer(
&self,
input: &Buffer,
context: &mut DecompressionContext,
) -> Result<Buffer, ArrowError> {
let decompressed_length = read_uncompressed_size(input)?;
let buffer = if decompressed_length == 0 {
Buffer::from([])
} else if decompressed_length == LENGTH_NO_COMPRESSED_DATA {
input.slice(LENGTH_OF_PREFIX_DATA as usize)
} else if let Ok(decompressed_length) = usize::try_from(decompressed_length) {
let input_data = &input[(LENGTH_OF_PREFIX_DATA as usize)..];
let v = self.decompress(input_data, decompressed_length as _, context)?;
Buffer::from_vec(v)
} else {
return Err(ArrowError::IpcError(format!(
"Invalid uncompressed length: {decompressed_length}"
)));
};
Ok(buffer)
}
fn compress(
&self,
input: &[u8],
output: &mut Vec<u8>,
context: &mut CompressionContext,
) -> Result<(), ArrowError> {
match self {
CompressionCodec::Lz4Frame => compress_lz4(input, output),
CompressionCodec::Zstd => compress_zstd(input, output, context),
}
}
fn decompress(
&self,
input: &[u8],
decompressed_size: usize,
context: &mut DecompressionContext,
) -> Result<Vec<u8>, ArrowError> {
let ret = match self {
CompressionCodec::Lz4Frame => decompress_lz4(input, decompressed_size)?,
CompressionCodec::Zstd => decompress_zstd(input, decompressed_size, context)?,
};
if ret.len() != decompressed_size {
return Err(ArrowError::IpcError(format!(
"Expected compressed length of {decompressed_size} got {}",
ret.len()
)));
}
Ok(ret)
}
}
#[cfg(feature = "lz4")]
fn compress_lz4(input: &[u8], output: &mut Vec<u8>) -> Result<(), ArrowError> {
use std::io::Write;
let mut encoder = lz4_flex::frame::FrameEncoder::new(output);
encoder.write_all(input)?;
encoder
.finish()
.map_err(|e| ArrowError::ExternalError(Box::new(e)))?;
Ok(())
}
#[cfg(not(feature = "lz4"))]
#[allow(clippy::ptr_arg)]
fn compress_lz4(_input: &[u8], _output: &mut Vec<u8>) -> Result<(), ArrowError> {
Err(ArrowError::InvalidArgumentError(
"lz4 IPC compression requires the lz4 feature".to_string(),
))
}
#[cfg(feature = "lz4")]
fn decompress_lz4(input: &[u8], decompressed_size: usize) -> Result<Vec<u8>, ArrowError> {
use std::io::Read;
let mut output = Vec::with_capacity(decompressed_size);
lz4_flex::frame::FrameDecoder::new(input).read_to_end(&mut output)?;
Ok(output)
}
#[cfg(not(feature = "lz4"))]
#[allow(clippy::ptr_arg)]
fn decompress_lz4(_input: &[u8], _decompressed_size: usize) -> Result<Vec<u8>, ArrowError> {
Err(ArrowError::InvalidArgumentError(
"lz4 IPC decompression requires the lz4 feature".to_string(),
))
}
#[cfg(feature = "zstd")]
fn compress_zstd(
input: &[u8],
output: &mut Vec<u8>,
context: &mut CompressionContext,
) -> Result<(), ArrowError> {
let result = context.zstd_compressor().compress(input)?;
output.extend_from_slice(&result);
Ok(())
}
#[cfg(not(feature = "zstd"))]
#[allow(clippy::ptr_arg)]
fn compress_zstd(
_input: &[u8],
_output: &mut Vec<u8>,
_context: &mut CompressionContext,
) -> Result<(), ArrowError> {
Err(ArrowError::InvalidArgumentError(
"zstd IPC compression requires the zstd feature".to_string(),
))
}
#[cfg(feature = "zstd")]
fn decompress_zstd(
input: &[u8],
decompressed_size: usize,
context: &mut DecompressionContext,
) -> Result<Vec<u8>, ArrowError> {
let output = context
.zstd_decompressor()
.decompress(input, decompressed_size)?;
Ok(output)
}
#[cfg(not(feature = "zstd"))]
#[allow(clippy::ptr_arg)]
fn decompress_zstd(
_input: &[u8],
_decompressed_size: usize,
_context: &mut DecompressionContext,
) -> Result<Vec<u8>, ArrowError> {
Err(ArrowError::InvalidArgumentError(
"zstd IPC decompression requires the zstd feature".to_string(),
))
}
#[inline]
fn read_uncompressed_size(buffer: &[u8]) -> Result<i64, ArrowError> {
let len_buffer = buffer.get(..LENGTH_OF_PREFIX_DATA as usize).ok_or_else(|| {
ArrowError::IpcError(format!(
"Compressed IPC buffer is too short: expected at least {LENGTH_OF_PREFIX_DATA} bytes, got {}",
buffer.len()
))
})?;
Ok(i64::from_le_bytes(len_buffer.try_into().unwrap()))
}
#[cfg(test)]
mod tests {
#[test]
#[cfg(feature = "lz4")]
fn test_lz4_compression() {
let input_bytes = b"hello lz4";
let codec = super::CompressionCodec::Lz4Frame;
let mut output_bytes: Vec<u8> = Vec::new();
codec
.compress(input_bytes, &mut output_bytes, &mut Default::default())
.unwrap();
let result = codec
.decompress(
output_bytes.as_slice(),
input_bytes.len(),
&mut Default::default(),
)
.unwrap();
assert_eq!(input_bytes, result.as_slice());
}
#[test]
#[cfg(feature = "zstd")]
fn test_zstd_compression() {
let input_bytes = b"hello zstd";
let codec = super::CompressionCodec::Zstd;
let mut output_bytes: Vec<u8> = Vec::new();
codec
.compress(input_bytes, &mut output_bytes, &mut Default::default())
.unwrap();
let result = codec
.decompress(
output_bytes.as_slice(),
input_bytes.len(),
&mut Default::default(),
)
.unwrap();
assert_eq!(input_bytes, result.as_slice());
}
#[test]
fn test_read_uncompressed_size_rejects_short_prefix() {
let err = super::read_uncompressed_size(&[1, 2, 3, 4, 5, 6, 7])
.expect_err("short compressed IPC prefix should return an error");
assert!(
err.to_string()
.contains("Compressed IPC buffer is too short"),
"unexpected error: {err}"
);
}
}