ox is a modular command-line toolkit written in C for compressing, encoding, analyzing, and profiling symbolic sequences. It includes bit-packing, entropy metrics, histogramming, finite-context modeling, and CRC-based hashing.

git clone https://github.com/cobilab/ox
cd ox/src/
make

./ox <command> [options]

Calculate basic statistics from a file.

./ox stats [-b 8|16] <filename>

• -b: bits data type (8 or 16 bits)

Generate random symbolic sequences.

./ox generate [-s <size>] [-c <cardinality>] [-e <seed>] <filename>

• -s: sequence size
• -c: alphabet cardinality (0–255)
• -e: random seed

Bit-pack sequences using 2-bit (ABCD) or 4-bit (A–P) encodings.

./ox pack2 pack <input> <output>
./ox pack2 unpack <input> <output>

./ox pack4 pack <input> <output>
./ox pack4 unpack <input> <output>

Encode/decode sequences with a custom XRC-256 codec: order-0 followed by a Range Coder.

./ox xrc-256 encode <input> <output>
./ox xrc-256 decode <input> <output>

Compute Shannon entropy of binary input.

./ox entropy [-v] <filename>

• -v: verbose output (byte frequencies and count)

Analyze distribution of values in a file (supports 8 and 16 bits).

./ox histogram [-h] [-t 8|16] [-w <width>] [-p] <filename>

• -t: data type (8 or 16 bits)
• -w: histogram width
• -p: plot instead of raw values
• -h: hide zero-count bins

Measure pattern distances in a sequence.

./ox distance -t <pattern> <filename>

• -t: pattern (e.g., RRR, EXFGGHH)

Compute CRC32 checksum.

./ox crc32-hash <filename>

Estimate local complexity using a finite-context model.

./ox profile [-k <ctx>] [-a <alphaDen>] [-w <window>] <filename>

• -k: model context order
• -a: smoothing parameter (1/a)
• -w: sliding window size

Print predefined analysis pipelines.

./ox pipelines

Example pipeline for DNA compression and decompression:

#!/bin/bash
grep -v '>' DNA.fa | tr -d -c 'ACGT' | tr 'ACGT' 'ABCD' > A.seq
./ox pack2 pack A.seq A.packed
./ox xrc-256 encode A.packed A.encoded

./ox xrc-256 decode A.encoded A.decoded
./ox pack2 unpack A.decoded A.unpacked
cmp A.unpacked A.seq

Print program version.

./ox version

• pack2: expects sequence with only 'A', 'B', 'C', 'D'
• pack4: expects symbols from 'A' to 'P'

# Prepare sequence
grep -v '>' input.fa | tr -d -c 'ACGT' | tr 'ACGT' 'ABCD' > seq.txt

# Pack using 2-bit
./ox pack2 pack seq.txt packed.bin

# Encode with custom codec
./ox xrc-256 encode packed.bin encoded.bin

# Decode
./ox xrc-256 decode encoded.bin decoded.bin

# Unpack to original
./ox pack2 unpack decoded.bin unpacked.txt

# Validate
cmp seq.txt unpacked.txt

GPLv3 License