binaryfile

A library for defining the structure of a binary file and then reading or writing it.

import binaryfile

def png(b):
	b.byteorder = 'big'
	b.skip(16)
	b.uint('width', 4)
	b.uint('height', 4)
	b.uint('depth', 1)

with open('image.png', 'rb') as fh:
	data = binaryfile.read(fh, png)
print(f"Image is {data.width} pixels wide, {data.height} pixels tall, and {data.depth} bits deep.")

Getting Started

Requirements

You will need Python 3.6 or later.

Installing

Windows with Python launcher:

py -3 -m pip install binaryfile

Linux with python3-pip:

pip3 install binaryfile

How to use

If you want to read or write to a binary file, first you will need to define the file structure. You do this by writing a function that takes a single argument, which is a subclass of binaryfile.fileformat.BinarySectionBase. The file structure is then defined by calling methods on said argument:

import binaryfile
import io

# Define the file structure
def file_spec(f):
	size = f.count('size', 'text', 2)  # A two-byte unsigned integer
	f.bytes('text', size)  # A variable number of bytes

if __name__ == '__main__':
	# Read the file and print the text field
	with open('myfile.dat', 'rb') as file_handle:
		data = binaryfile.read(file_handle, file_spec)
	print(data.text.decode('utf-8'))

	# Modify the text field
	data.text += ' More Text!'.encode('utf-8')

	# Then write back to file
	with open('myfile.dat', 'wb') as file_handle:
		binaryfile.write(file_handle, data, file_spec)

You can break the definition into reusable sections:

def subsection_spec(f):
	f.struct('position', 'fff')  # Three floats, using a format string from Python's built-in struct module

def section_spec(f):
	f.int('type', 1)  # A one-byte signed integer
	f.section('subsection1', subsection_spec)  # Three floats, as specified in subsection_spec
	f.section('subsection2', subsection_spec)

def file_spec(f):
	f.section(f'section1', section_spec)
	f.section(f'section2', section_spec)
	f.section(f'section3', section_spec)

if __name__ == '__main__':
	with open('myfile2.dat', 'rb') as file_handle:
		data = binaryfile.read(file_handle, file_spec)
	print(data.section2.subsection1.position)

And you can declare fields to be arrays and use loops:

def file_spec(f):
	f.array('positions')  # Declare "positions" to be an array
	count = f.count('count', 'positions', 4)
	for i in range(count):
		f.struct('positions', 'fff')  # Each time "positions" is used, it's the next element of the array

Reference

The reference documentation for this module is in the source at binaryfile/fileformat.py.

• Look at the BinarySectionBase class for all the methods available when writing a specification.
• Then check out the read and write functions for how to use the specification to read and write file-like objects. The read and write functions are also available from the binaryfile namespace.

Configuration

Result type

By default, a file is read into a binaryfile.utils.SimpleDict, which allows you to access the fields by dot notation (e.g. foo.bar.baz). This means you cannot use names that are invalid field names in Python.

To override the result type, pass the desired type to result_type in the read call, e.g.:

binaryfile.read(fh, spec, result_type=dict)

The desired type must be a dict-like type that implements __getitem__, __setitem__ and __contains__.

Byte order

The default byte order is big-endian. You can change the endianness either by setting byteorder on the BinarySectionBase object, or in individual methods that support it. Valid byteorders are 'big' and 'little', which is also the possible values returned by sys.byteorder.

def spec(b):
	b.byteorder = 'little'
	b.int('a', 4)  # Little-endian
	b.int('b', 4, byteorder='big')  # Big-endian
	b.int('c', 4)  # Little-endian again

Automated tests

Setting up the environment

1. Create and activate a Python virtual environment.
2. From the project root, run ./setup.py develop to install a binaryfile package linked to the project source into the venv.

Running the tests

Make sure that the venv is active, then run the Python files in the tests folder.

License

This project is licensed under MIT License, see LICENSE for details.