sbe-python

Easy to use, fast, pure Python FIX SBE encoder and decoder.

Install

pip install sbe

Usage

Simple Decoding

Decode SBE to Python dictionaries in one line. Good for exploratory data analysis.

import sbe

with open('your-schema.xml', 'r') as f:
  schema = sbe.Schema.parse(f)

wtih open('your-data.sbe', 'rb') as f:
  buf = f.read()

# Get a Python dict in one-line
x = schema.decode(buf)

x.name  # The template message name

x.value
# {'userId': 11,
# 'timestamp': 1598784004840,
# 'orderSize': 0,
# 'price': 5678.0,
# ...

# If you need an initial offset, apply it Pythonically
schema.decode(buf[19:])

# decode_header to filter out messages based on header to avoid decoding
# message bodies that are not needed
schema.decode_header(buf)['templateId']

High Performance Decoding (Wrapping)

This gives you decent performance while still retaining high code readability.

import sbe

with open('your-schema.xml', 'r') as f:
  schema = sbe.Schema.parse(f)

wtih open('your-data.sbe', 'rb') as f:
  buf = f.read()

# Wrap the buffer without decoding it, fields are converted to Python variables
# on demand
x = schema.wrap(buf)

x.header['templateId']
x.body['price']
x.body['someGroup'][2]['price']

Direct Access with Pointers

get_raw_pointer gives you the required information to unpack a variable from memoryview / bytes. This gets you very close to the fastest achievable performance in Python:

import sbe

with open('your-schema.xml', 'r') as f:
  schema = sbe.Schema.parse(f)

header_pointer = schema.header_wrapper.get_raw_pointer('templateId')

# Let's say we are only interested in messages of templateId == 3
price_pointer = schema.message_wrappers[3].get_raw_pointer('price')

wtih open('your-data.sbe', 'rb') as f:
  buf = f.read()

# pass `memoryview` to `unpack` to avoid copying
buf = memoryview(buf)[initial_offset:]
template_id = header_pointer.unpack(buf)  # calls buf[offset:offset+size].cast("I")[0] directly

if template_id == 3:
  print(price_pointer.unpack(buf))

Encoding

import sbe

with open('./your-schema.xml', 'r') as f:
  schema = sbe.Schema.parse(f)

# message_id from the schema you want to encode
message_id = 3

obj = {
  'userId': 11,
  'price': 5678.0,
  # ...
}

# Encode from Python dict in one-line
schema.encode(schema.messages[3], obj)

# You can supply your header values as a dict
schema.encode(schema.messages[3], obj, headers)

Metadata

Release files for sbe 0.4.3

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for sbe 0.4.3
File	Size	Uploaded
sbe-0.4.3.tar.gz	12.3 kB	Dec 20, 2024	Details

Release files / sbe-0.4.3.tar.gz

Download URL	sbe-0.4.3.tar.gz
Size	12.3 kB
Tags	Source
SHA-256 checksum How to use checksums	`8913b543302e89353f1fb90556606883e795b5c6367b2c27fb039450c0118f7a`
BLAKE2b-256 checksum How to use checksums	`37f48fb98b65eb24d4a02b21fc06895bf6586b042dcc590e2b556efdd6e062db`
Upload date	Dec 20, 2024
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/5.1.1 CPython/3.11.5`