chimp-encoding 0.2.0

Chimp time-series floating point compression for Python

chimp-encoding

Chimp time-series floating point compression for Python.

A Python port of the Chimp algorithm for lossless compression of floating-point time-series data. Supports both 64-bit (double) and 32-bit (float) variants, in base mode and history-based (Chimp128/ChimpN) mode.

Installation

pip install chimp-encoding

Usage

Encoding and decoding 64-bit doubles

from chimp_encoding import ChimpEncoder, ChimpDecoder

# Encode
encoder = ChimpEncoder()
for value in [23.5, 23.6, 23.4, 24.0, 22.9]:
    encoder.add_value(value)
encoder.close()

compressed = encoder.get_bytes()
print(f"Compressed to {len(compressed)} bytes")

# Decode
decoder = ChimpDecoder(compressed)
values = decoder.get_values()
print(values)  # [23.5, 23.6, 23.4, 24.0, 22.9]

Using Chimp128 (history-based, better compression)

from chimp_encoding import ChimpNEncoder, ChimpNDecoder

encoder = ChimpNEncoder(previous_values=128)
for value in data:
    encoder.add_value(value)
encoder.close()

decoder = ChimpNDecoder(encoder.get_bytes(), previous_values=128)
values = decoder.get_values()

32-bit float variants

from chimp_encoding import Chimp32Encoder, Chimp32Decoder
from chimp_encoding import ChimpN32Encoder, ChimpN32Decoder

CLI

# Encode values to hex
chimp encode 23.5 23.6 23.4 24.0 22.9

# Encode to file
chimp encode 23.5 23.6 23.4 -o compressed.bin

# Decode from hex
chimp decode --hex "4037800000000000..."

# Decode from file
chimp decode compressed.bin

# Use different variants
chimp encode --variant chimp128 23.5 23.6 23.4
chimp decode --variant chimp128 --hex "..."

Variants

Variant	Class	Description
Chimp	ChimpEncoder / ChimpDecoder	Base 64-bit, XOR with previous value
Chimp32	Chimp32Encoder / Chimp32Decoder	Base 32-bit
ChimpN	ChimpNEncoder / ChimpNDecoder	64-bit with N-value history (default N=128)
ChimpN32	ChimpN32Encoder / ChimpN32Decoder	32-bit with N-value history (default N=64)

License

MIT