Skip to main content

Parse binary files by describing their structure.

Project description

Parse Binary File

Install with python -m pip install parse_binary_file

Parse binary files by describing their structure.

Description file

The parse_binary_file package is based on the idea that binary file formats should be parsed from a description of their format.

If the description file is a list, pbf assumes that the elemnts are fields. One can also pass a dictionary, though, to describe more details about the file format.

There are three top-level items that can be specified:

  • info: Describes information about the file as a whole (e.g. endianess).
  • default_options: Describes defaults options for each field type.
  • fields: Fields of the file format.


Describes information about the file.


  • byte_order: Byte order of the file. Valid values: ['little', 'big', 'native'] native byte ordering uses the byte ordering of the current machine. [Default: 'native']

  • encoding: File encoding. Used to encode the description file's values to bytes when matching against the file being parsed. See Python's standard encodings for valid values. [Default: 'utf-8']

  • word_size: Bytes per word of the file. [Default: 4 bytes] [Inactive]

Default Options

Describes parsing information for each type of field. All types can specify default values for a particular field property (discussed below). To specify a default field property first specify the type of the field, then the property.

e.g. To specify strings are null terminated (\x00) by default

        terminator: "\x00"


Describes the fields of the file format. Each field must one and only one termination condition. Termination conditions are determined by the size, value, and terminator fields. for some types these can be interpreted. For instance, an int is assumed to have a size of 4, so a termination condition does not need to be explicitly stated.

All fields can have the properties:

  • type: Data type of the field. If a type is enclosed in brackets ([]) it marks the field as an array-like field. Array like fields are assumed to contain multiple instances of the specified type. For more information see the Array Fields section.
  • size: Sets the size of the field in bytes. This may be inferred depending on the field's type. A size of -1 indicated the field extends until the end of the file. Only the final field may have a length of -1.
  • value: Sets the size of the fiel based on the expected value.
  • format: How to format the field. This is type dependent. For str fields, this represents the encoding of the string and defaults to the file's encoding provided in the info section. For other types this represents how Python's struct.unpack should interpret the field.
  • terminator: Termination string.
  • is_null: Used for bytes data type to indicate all bytes should be null. Should be used in conjunction with the size property to indicate the termination condition.
  • name: Name of the field. Used for accessing the data by name.
  • description: Description of the field.
  • fields: Describes the subfields of the field. If a field is made up of subfields, its type must be bytes or [bytes].
  • exec: Execution hooks for logical processing. [Inactive]


If a field is not provided a type it defaults to bytes.

Valid types are:

  • bytes: Interpret the field as raw bytes. [Default]
  • char: Single character. (1 byte)
  • str: String data.
  • bool: Boolean. (1 byte)
  • short: Short integer. (2 bytes)
  • u_short: Unsigned short integer. (2 bytes)
  • int: Integer.(4 bytes)
  • u_int: Unisigned integer. (4 bytes)
  • long: Long integer. (1 word)
  • u_long: Unsigned long integer. (1 word)
  • long_long: Double long integer. (2 words)
  • u_long_long: Unsigned double long integer. (2 words)
  • float: Floating point number. (4 bytes)
  • double: Double precision floating point number. (2 words)
  • Any type may be wrapped in brackets [] to indicate the field is an array of that value type. Details of the elements are set using the elements property.

For more information read about Python struct's formatting strings.

Termination condition

The termination condition of a field tells the parser how far to read. There are three explicit and one implit termination condition.

  • size: Sets the size of the field in bytes. This may be inferred depending on the field's type.
  • terminator: Termination string as bytes.
  • value: Sets the size of the field based on the expected value.
  • subfields: [Inactive] The parent field can be terminated when the last child field is terminated.


Each field can be made up of subfields. Each subfield follows the same pattern as top-level fields. This is a recursive concept, so fields can be nested as deeply as needed.

Fields that are made up of subfields must have their type set to bytes or [bytes]. If the type is [bytes] it is assumed that the subfields repeat in a array-like manner with all sufields being repeated an equal number of times. i.e. If the data for the parent field ends before all subfields have been read an exception is raised.

Array Fields

A field may consist of multiple elements. This is where array fields come into play. For instance, imagine a field that consists of a list of floats. To indicate a field is array-like, enclose its type in square brackets ('[]') (e.g. [bytes], [float]).

Execution Hooks

:warning: These fields allow arbitrary Python code to be executed.

It may be necessary to add some basic logic to your description. For instance, if a field named data may have a variable length offset buffer specified by a field named data_offset. The exec field allows you to specify code before (pre) or after (post) data for the field has been loaded and has access to all previous fields. Each of these fields should contain a lambda function with signature (data, fields) where data is byte string of the current field and fields is a parse_binary_file.Data object representing the file with all previous fields loaded.



Describes a field.


  • type: The type of the field being described. When loading data into the field it will be parsed based on the type.

    • [Default: 'bytes']
  • name: Name of the field. This is used to label the field.

  • value: The expected value of the field. This can be usedto indicate a specific value is expected for the field.

  • size: Number of bytes in the field. The last field can have a size of -1 which indicates the field continues to the end of the file.

  • terminator: Sequence of values indicating the end of the field.

  • is_null: Indicates the field only contains null bytes. Settings to true is equivalent to setting value equal to all null bytes. Must set size if True.

  • description: Description of the field.

  • fields: Used to specifiy the structure of the elements of an array.


  • FieldDescription(**properties): Initializes a new FieldDescription with the provided properties.


Container for FieldDescriptions to describe the structure of a file.


  • fields: List of FieldDescriptions.

  • info: Dictionary of information on the file.

  • Fields can be accessed by name or index using brackets ([])


  • from_dicts(desc, info, defaults): @staticmethod Converts a list of dictionaries into a FileFormat.

  • keys(): Returns a list of the keys of named fields.


Used for parsing files in a given format.


  • format: A FileFormat used to parse files.


  • parse(stream): Returns a Data object representing the data from stream.


Contains information about a field, including its description and loaded value.


  • desc: FieldDescription that was used to create the Field.

  • value: Parsed value.

  • expected_value: self.desc.value. Provided for convenience.

  • fields: Tuple of children Fields.

  • size: If the Field's value is set, returns its size. If the Field's value has not yet been set return the Field's expected size if available.

  • data: Original data as bytes.

  • value_is_valid: Whether the parsed value matches the expected value. If a specific value was not specified by the FieldDescription this will always return True.

  • Values of the Field's FieldDescription are accessible as properties, as well.


  • Field(desc: FieldDescription, fields: tuple[Field]): Initializes a new Field. desc should describe the field, and fields are the children Fields.

  • from_data(data: bytes, desc: FieldDescription): Creates a new Field using desc as its description to parse data.

  • parse_data(val: bytes, **options): Sets the value of the Field based on its description and the provided value.


Represents data from a file.


  • fields: Fields containing the parsed data.

  • is_loaded: If data has been loaded yet.

  • value: Values of the parsed data.

  • named_field_values: Returns a dictionary of name-value pairs for named Fields. If multiple fields have the same name a tuple is returned with the values in order.

  • Fields are accessible by name and index using brackets ([]). If multiple Fields have the same name, they are returned as a tuple in order.


This library is intended to be used by describing the struture of a binary file format in a configuration file. That file is then loaded and used to create a Data instance. The Data instance can then load() data from a file and assort it into the correct fields.


We would like to describe a new binary file format of .msg.

Offset Size Name Description
0x00 0x02 head 0xff 0x00 Indicates the file is a .msg file.
0x02 0x04 size Total size of the file.
0x06 0x04 -- Null bytes to separate header from body.
0x0a -- greeting Message greeting, null terminated.
-- -- message Message content.

We can describe this structure in a YAML file. msg.yaml

    byte_order: 'little'

    - name: 'head'
      value: "\xff\x00"

    - type: 'int'
      name: 'file_size'
      description: 'Size of file in bytes'

    - size: 4
      is_null: true

    - type: 'str'
      terminator: "\x00"
      name: 'greeting'

    - type: 'str'
      size: -1
      name: 'message'

To read a .msg file we can use the description of it.

import os
import yaml  # use pyyaml to load `msg.yaml`
import parse_binary_file as pbf

# load description
with open('msg.yaml') as f:
    msg_desc = yaml.safe_load(f)
# create file format
msg_format = pbf.FieldFormat.from_dicts(msg_desc['fields'], info = msg_desc['info'])

# create parser
parser = pbf.Parser(msg_format)

# create data
with open('my_message.msg', 'rb') as f:
    data = parser.parse(f)

is_corrupt = (data['file_size'] != os.path.getsize('my_message.msg'))
greeting = data['greeting'].value
message = data[-1].value

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

parse_binary_file-0.0.2.tar.gz (18.8 kB view hashes)

Uploaded Source

Built Distribution

parse_binary_file-0.0.2-py3-none-any.whl (19.1 kB view hashes)

Uploaded Python 3

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