cs.binary 20181231

Facilities associated with binary data parsing and transcription.

Facilities associated with binary data parsing and transcription.

Note: this module requires Python 3 and recommends Python 3.6+ because it uses abc.ABC, because a Python 2 bytes object is too weak (just a str) as also is my cs.py3.bytes hack class and because the keyword based Packet initiialisation benefits from keyword argument ordering.

In the description below I use the word "chunk" to mean a piece of binary data obeying the buffer protocol, almost always a bytes instance or a memoryview, but in principle also things like bytearray.

The classes in this module support easy parsing of binary data structures.

The functions and classes in this module include:

• PacketField: an abstract class for a binary field, with a factory method to parse it, a transcription method to transcribe it back out in binary form and usually a .value attribute holding the parsed value.
• several presupplied subclasses for common basic types such as UInt32BE (an unsigned 32 bit big endian integer).
• struct_field: a factory for making PacketField classes for struct formats with a single value field.
• multi_struct_field and structtuple: factories for making PacketFields from struct formats with multiple value fields; structtuple makes PacketFields which are also namedtuples, supporting trivial access to the parsed values.
• Packet: a PacketField subclass for parsing multiple PacketFields into a larger structure with ordered named fields.

You don't need to make fields only from binary data; because PacketField.__init__ takes a post parse value, you can also construct PacketFields from scratch with their values and transcribe the resulting binary form.

Each PacketField subclass has the following methods:

• transcribe: easily return the binary transcription of this field, either directly as a chunk (or for convenience, also None or an ASCII str) or by yielding successive binary data.
• from_buffer: a factory to parse this field from a cs.buffer.CornuCopyBuffer.
• from_bytes: a factory to parse this field from a chunk with an optional starting offset; this is a convenience wrapper for from_buffer.

That may sound a little arcane, but we also supply:

• flatten: a recursive function to take the return from any transcribe method and yield chunks, so copying a packet to a file or elsewhere can always be done by iterating over flatten(field.transcribe()) or via the convenience field.transcribe_flat() method which calls flatten itself.
• a CornuCopyBuffer is an easy to use wrapper for parsing any iterable of chunks, which may come from almost any source. It has a bunch of convenient factories including: from_bytes, make a buffer from a chunk; from_fd, make a buffer from a file descriptor; from_file, make a buffer from a file-like object; from_mmap, make a buffer from a file descriptor using a memory map (the mmap module) of the file, so that chunks can use the file itself as backing store instead of allocating and copying memory. See the cs.buffer module for further detail.

Cameron Simpson [e-mail] 22jul2018

Class BSData

MRO: PacketField, abc.ABC
A run length encoded data chunk, with the length encoded as a BSUInt.

Class BSString

MRO: PacketField, abc.ABC
A run length encoded string, with the length encoded as a BSUInt.

Class BSUInt

MRO: PacketField, abc.ABC
A binary serialsed unsigned int.

This uses a big endian byte encoding where continuation octets have their high bit set. The bits contributing to the value are in the low order 7 bits.

Class BytesesField

MRO: PacketField, abc.ABC
A field containing a list of bytes chunks.

The following attributes are defined:

• value: the gathered data as a list of bytes instances, or None if the field was gathered with discard_data true.
• offset: the starting offset of the data.
• end_offset: the ending offset of the data.

The offset and end_offset values are recorded during the parse, and may become irrelevant if the field's contents are changed.

Class BytesField

MRO: PacketField, abc.ABC
A field of bytes.

Class BytesRunField

MRO: PacketField, abc.ABC
A field containing a continuous run of a single bytes value.

The following attributes are defined:

• length: the length of the run
• bytes_value: the repeated bytes value

The property value is computed on the fly on every reference and returns a value obeying the buffer protocol: a bytes or memoryview object.

Class EmptyPacketField

MRO: PacketField, abc.ABC
An empty data field, used as a placeholder for optional fields when they are not present.

The singleton EmptyField is a predefined instance.

Function fixed_bytes_field(length, class_name=None)

Factory for BytesField subclasses built from fixed length byte strings.

Function flatten(chunks)

Flatten chunks into an iterable of bytes instances.

This exists to allow subclass methods to easily return ASCII strings or bytes or iterables or even None, in turn allowing them to simply return their superclass' chunks iterators directly instead of having to unpack them.

Class Int16BE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '>h'.

Class Int16LE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '<h'.

Class Int32BE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '>l'.

Class Int32LE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '<l'.

Class ListField

MRO: PacketField, abc.ABC
A field which is a list of other fields.

Function multi_struct_field(struct_format, subvalue_names=None, class_name=None)

Factory for PacketField subclasses build around complex struct formats.

Parameters:

• struct_format: the struct format string
• subvalue_names: an optional namedtuple field name list; if supplied then the field value will be a namedtuple with these names
• class_name: option name for the generated class

Class Packet

MRO: PacketField, abc.ABC
Base class for compound objects derived from binary data.

Class PacketField

MRO: abc.ABC
A record for an individual packet field.

Function struct_field(struct_format, class_name)

Factory for PacketField subclasses built around a single struct format.

Parameters:

• struct_format: the struct format string, specifying a single struct field
• class_name: the class name for the generated class

Example:

>>> UInt16BE = struct_field('>H', class_name='UInt16BE')
>>> UInt16BE.__name__
'UInt16BE'
>>> UInt16BE.format
'>H'
>>> UInt16BE.struct   #doctest: +ELLIPSIS
<Struct object at ...>
>>> field, offset = UInt16BE.from_bytes(bytes((2,3,4)))
>>> field
UInt16BE(515)
>>> offset
2
>>> field.value
515

Function structtuple(class_name, struct_format, subvalue_names)

Convenience wrapper for multi_struct_field.

Class UInt16BE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '>H'.

Class UInt16LE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '<H'.

Class UInt32BE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '>L'.

Class UInt32LE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '<L'.

Class UInt64BE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '>Q'.

Class UInt64LE

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format '<Q'.

Class UInt8

MRO: PacketField, abc.ABC
A PacketField which parses and transcribes the struct format 'B'.

Class UTF8NULField

MRO: PacketField, abc.ABC
A NUL terminated UTF-8 string.