fudong1127 / pylimitbook

Python implementation of fast limit-order book.

PyLimitBook

PyLimitBook is an implementation of a fast limit-order book for level-2 US equities data writen in Python. It includes some tools to output sampled data as well as a curses-based application to view the book and move forward or backward in time.

Input files must be for a single day, symbol, and exchange.

This code is aimed at other developers looking for a limit-book implementation to include in their own trading projects.

Requirements

• Python libraries bintrees (v0.4 used) and pandas.
• curses and cPickle libraries are also used, but are often installed by default.
• This code has thus far only been tested on Ubuntu Linux with Python 2.7.2.

Input Data Format

PyLimitBook applications expect the following input format (in a CSV file):

• bids:

  B,<symbol>,<exchange>,<id>,<quantity>,<price>,<timestamp>

• asks:

  A,<symbol>,<exchange>,<id>,<quantity>,<price>,<timestamp>

• trades: (optional)

  T,<symbol>,<exchange>,<quantity>,<price>,<timestamp>

Symbol and exchange values are currently unimportant and can be dummy values. Lines should be in timestamp-order from earliest to latest (the order it is sent by the exchange).

Application Descriptions

• parse.py - Simple illustration of using the limit book
• tseries.py - Parse an input file and export 1-second snapshots of the book (Uses pandas library to fill in gaps, etc)
• convert.py - Convert from an input format that uses ', ' as a CSV separator and different column order to the correct input format.
• bookViewer.py - Graphically view the book state at any point in time

Known Issues

• There are some resize issues when using bookViewer.py in a (uselessly) small terminal.