BinField 0.7.0

binfield

Python binfield implementation for binary data manipulation.

Why? Python supports binary data manipulation via binary operations out of the box and it’s fast, but it’s hard to read and painful during prototyping, especially for complex (nested) structures.

This library is desined to fix this issue: it allows to operate with binary data like dict with constant indexes: you just need to define structure class and create an instance with start data. Now you can use indexes for reading and writing data

Pros:

• Free software: Apache license

• Open Source: https://github.com/penguinolog/binfield

• Self-documented code: docstrings with types in comments

• Tested: see bages on top

• Support multiple Python versions:

Python 2.7
Python 3.4
Python 3.5
Python 3.6
PyPy
PyPy3
Jyton 2.7

Usage

Not mapped objects could be created simply from BinField class:

bf = BinField(42)

Data with fixed size should be created as new class (type): Example on real data (TCP header, constant part):

# Describe
class TCPHeader(binfield.BinField):
    _size_ = 160
    _mask_ = 0xFFFFFFFFFFFFFF8FFFFFFFFFFFFFFFFFFFFFFFFF
    source_port = [0, 16]
    destination_port = [16, 32]
    sequence_number = [32, 64]
    ack_number = [64, 96]
    data_offset = [96, 100]
    flags = {
        '_index_': [103, 112],
        'NS': 0,
        'CWR': 1,
        'ECE': 2,
        'URG': 3,
        'ACK': 4,
        'PSH': 5,
        'RST': 6,
        'SYN': 7,
        'FIN': 8
    }
    window_size = [112, 128]
    checksum = [128, 144]
    urgent_pointer = [144, 160]

# Construct from frame
# (limitation: endianless convertation is not supported, make it by another tools)
header = TCPHeader(0x0000BD1A043708050000078B000007F601BBAF0A)

# Do not print header due to huge length (will be printed all bits)
>>> repr(header)
'TCPHeader(x=0x0000BD1A043708050000078B000007F601BBAF0A, base=16)'

>>> repr(header.source_port)
'<source_port(x=0xAF0A, base=16) at 0x7F890C8B9348>'

>>> print(header.source_port)
<44810 == 0xAF0A == (0b1010111100001010 & 0b1111111111111111)>

>>> header.source_port == 44810  # Transparent comparsion with integers
True

>>> int(header.source_port)  # Painless conversion to int
44810

>>> print(header.destination_port)
<443 == 0x01BB == (0b0000000110111011 & 0b1111111111111111)> # Request multiple bytes

>>> print(header.data_offset)  # Request multiple bits
<5 == 0x05 == (0b0101 & 0b1111)>

>>> print(header.destination_port[1: 3])  # Request several bits from nested block too
<1 == 0x01 == (0b01 & 0b11)>

>>> print(header.flags)  # Request nested mapping block
<16 == 0x0010 == (0b000010000 & 0b111111111)
  NS  = <0 == 0x00 == (0b0 & 0b1)>
  CWR = <0 == 0x00 == (0b0 & 0b1)>
  ECE = <0 == 0x00 == (0b0 & 0b1)>
  URG = <0 == 0x00 == (0b0 & 0b1)>
  ACK = <1 == 0x01 == (0b1 & 0b1)>
  PSH = <0 == 0x00 == (0b0 & 0b1)>
  RST = <0 == 0x00 == (0b0 & 0b1)>
  SYN = <0 == 0x00 == (0b0 & 0b1)>
  FIN = <0 == 0x00 == (0b0 & 0b1)>
>

>>> print(header.flags.ACK == 0x01)  # Request single bit from nested mapping
True

>>> print(header[: 4])  # Ignore indexes and just get few bits using slice
<10 == 0x0A == (0b1010 & 0b1111)>

# Modification of nested data (if no type conversion was used) changes original object:
header.flags.FIN = 1
>>> print(header.flags)
<272 == 0x0110 == (0b100010000 & 0b111111111)
  NS  = <0 == 0x00 == (0b0 & 0b1)>
  CWR = <0 == 0x00 == (0b0 & 0b1)>
  ECE = <0 == 0x00 == (0b0 & 0b1)>
  URG = <0 == 0x00 == (0b0 & 0b1)>
  ACK = <1 == 0x01 == (0b1 & 0b1)>
  PSH = <0 == 0x00 == (0b0 & 0b1)>
  RST = <0 == 0x00 == (0b0 & 0b1)>
  SYN = <0 == 0x00 == (0b0 & 0b1)>
  FIN = <1 == 0x01 == (0b1 & 0b1)>
>

# Indexes is accessible from base class
>>> print(TCPHeader.source_port)
slice(0, 16, None)

# But remember, that nested blocks has it's own classes
>>> header.flags.__class__
<class 'binfield.binfield.flags'>

# Fields could be set only from integers
>>> header2 = TCPHeader()
>>> header2.flags = header.flags
Traceback (most recent call last):
...
TypeError: BinField value could be set only as int

>>> header2.flags = int(header.flags)
>>> header2.flags
<flags(x=0x0110, base=16) at 0x7FBC9A21FFC8>

Note: negative indexes is not supported by design!

Testing

The main test mechanism for the package binfield is using tox. Test environments available:

pep8
py27
py34
py35
py36
pypy
pypy3
jyton
pylint

CI systems

For code checking several CI systems is used in parallel:

1. Travis CI: is used for checking: PEP8, pylint, bandit, installation possibility and unit tests. Also it’s publishes coverage on coveralls.

2. coveralls: is used for coverage display.

CD system

Travis CI: is used for package delivery on PyPI.