Skip to main content

Consistent Overhead Byte Stuffing (COBS)

Project description

Author:

Craig McQueen

Contact:
http://craig.mcqueen.id.au/

Python functions for encoding and decoding COBS.

Intro

The cobs package is provided, which contains modules containing functions for encoding and decoding according to COBS methods.

What Is COBS?

COBS is a method of encoding a packet of bytes into a form that contains no bytes with value zero (0x00). The input packet of bytes can contain bytes in the full range of 0x00 to 0xFF. The COBS encoded packet is guaranteed to generate packets with bytes only in the range 0x01 to 0xFF. Thus, in a communication protocol, packet boundaries can be reliably delimited with 0x00 bytes.

The COBS encoding does have to increase the packet size to achieve this encoding. However, compared to other byte-stuffing methods, the packet size increase is reasonable and predictable. COBS always adds 1 byte to the message length. Additionally, for longer packets of length n, it may add n/254 (rounded down) additional bytes to the encoded packet size.

For example, compare to the PPP protocol, which uses 0x7E bytes to delimit PPP packets. The PPP protocol uses an “escape” style of byte stuffing, replacing all occurences of 0x7E bytes in the packet with 0x7D 0x5E. But that byte-stuffing method can potentially double the size of the packet in the worst case. COBS uses a different method for byte-stuffing, which has a much more reasonable worst-case overhead.

For more details about COBS, see the references [1] [2].

I have included a variant on COBS, COBS/R, which slightly modifies COBS to often avoid the +1 byte overhead of COBS. So in many cases, especially for smaller packets, the size of a COBS/R encoded packet is the same size as the original packet. See below for more details about COBS/R.

References

Modules Provided

Module

Short Name

Long Name

cobs.cobs

COBS

Consistent Overhead Byte Stuffing (basic method) [1]

cobs.cobsr

COBS/R

Consistent Overhead Byte Stuffing–Reduced

Consistent Overhead Byte Stuffing–Reduced” (COBS/R) is my own invention, a modification of basic COBS encoding, and is described in more detail below.

The following are not implemented:

Short Name

Long Name

COBS/ZPE

Consistent Overhead Byte Stuffing–Zero Pair Elimination [1]

COBS/ZRE

Consistent Overhead Byte Stuffing–Zero Run Elimination [2]

A pure Python implementation and a C extension implementation are provided. If the C extension is not available for some reason, the pure Python version will be used.

Usage

The modules provide an encode and a decode function.

The input should be a byte string, not a Unicode string. Basic usage:

>>> from cobs import cobs
>>> encoded = cobs.encode(b'Hello world\x00This is a test')
>>> encoded
b'\x0cHello world\x0fThis is a test'
>>> cobs.decode(encoded)
b'Hello world\x00This is a test'

COBS/R usage is almost identical:

>>> from cobs import cobsr
>>> encoded = cobsr.encode(b'Hello world\x00This is a test')
>>> encoded
b'\x0cHello worldtThis is a tes'
>>> cobsr.decode(encoded)
b'Hello world\x00This is a test'

Any type that implements the buffer protocol, providing a single block of bytes, is also acceptable as input:

>>> from cobs import cobs
>>> encoded = cobs.encode(bytearray(b'Hello world\x00This is a test'))
>>> encoded
b'\x0cHello world\x0fThis is a test'
>>> cobs.decode(encoded)
b'Hello world\x00This is a test'

Note that the encode function does not add any framing 0x00 byte at the end (or beginning) of the encoded data. Similarly, the decode function does not strip or split on any framing 0x00 bytes, but treats any 0x00 bytes as a data input error. The details of data framing is application-specific, so it is the user’s application’s responsibility to implement the framing and deframing that is suitable for the needs of the application.

Supported Python Versions

Python >= 3.6 are supported, and have both a C extension and a pure Python implementation.

Python versions < 3.6 might work, but have not been tested.

Installation

The cobs package is installed using distutils. If you have the tools installed to build a Python extension module, run the following command:

python setup.py install

If you cannot build the C extension, you may install just the pure Python implementation, using the following command:

python setup.py build_py install --skip-build

Unit Testing

Basic unit testing is in the test sub-module, e.g. cobs.cobs.test. To run it:

python -m cobs.cobs.test
python -m cobs.cobsr.test

Documentation

Documentation is written with Sphinx. Source files are provided in the doc directory. It can be built using Sphinx 0.6.5. It uses the pngmath Sphinx extension, which requires Latex and dvipng to be installed.

The documentation is available online at: http://packages.python.org/cobs/

License

The code is released under the MIT license. See LICENSE.txt for details.

Consistent Overhead Byte Stuffing–Reduced (COBS/R)

A modification of COBS, which I’m calling “Consistent Overhead Byte Stuffing–Reduced” (COBS/R), is provided in the cobs.cobsr module. Its purpose is to save one byte from the encoded form in some cases. Plain COBS encoding always has a +1 byte encoding overhead. See the references for details [1]. COBS/R can often avoid the +1 byte, which can be a useful savings if it is mostly small messages that are being encoded.

In plain COBS, the last length code byte in the message has some inherent redundancy: if it is greater than the number of remaining bytes, this is detected as an error.

In COBS/R, instead we opportunistically replace the final length code byte with the final data byte, whenever the value of the final data byte is greater than or equal to what the final length value would normally be. This variation can be unambiguously decoded: the decoder notices that the length code is greater than the number of remaining bytes.

Examples

The byte values in the examples are in hex.

First example:

Input:

2F

A2

00

92

73

02

This example is encoded the same in COBS and COBS/R. Encoded (length code bytes are bold):

03

2F

A2

04

92

73

02

Second example:

The second example is almost the same, except the final data byte value is greater than what the length byte would be.

Input:

2F

A2

00

92

73

26

Encoded in plain COBS (length code bytes are bold):

03

2F

A2

04

92

73

26

Encoded in COBS/R:

03

2F

A2

26

92

73

Because the last data byte (26) is greater than the usual length code (04), the last data byte can be inserted in place of the length code, and removed from the end of the sequence. This avoids the usual +1 byte overhead of the COBS encoding.

The decoder detects this variation on the encoding simply by detecting that the length code is greater than the number of remaining bytes. That situation would be a decoding error in regular COBS, but in COBS/R it is used to save one byte in the encoded message.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

cobs-1.2.1.tar.gz (14.6 kB view hashes)

Uploaded Source

Built Distributions

cobs-1.2.1-cp312-cp312-win_amd64.whl (27.3 kB view hashes)

Uploaded CPython 3.12 Windows x86-64

cobs-1.2.1-cp312-cp312-win32.whl (25.9 kB view hashes)

Uploaded CPython 3.12 Windows x86

cobs-1.2.1-cp311-cp311-win_amd64.whl (27.3 kB view hashes)

Uploaded CPython 3.11 Windows x86-64

cobs-1.2.1-cp311-cp311-win32.whl (25.8 kB view hashes)

Uploaded CPython 3.11 Windows x86

cobs-1.2.1-cp310-cp310-win_amd64.whl (27.3 kB view hashes)

Uploaded CPython 3.10 Windows x86-64

cobs-1.2.1-cp310-cp310-win32.whl (25.8 kB view hashes)

Uploaded CPython 3.10 Windows x86

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page