Low Latency incremental Technical Analysis
Project description
RTTA
Purpose
The purpose of this package is to implement a very low latency incremental technical analysis toolkit. Most technical analysis tool-kits work in a "batch mode" where you hand them a blob of data and in a pandas series and they return a series with the computed data. Incremental updates for these require O(n) work. There is one tool, talipp that is designed to support incremental updates, but it is implemented in pure python and is a little more than an order of magnitude slower than rtta. On a 5995WX talipp's exponential moving average requires 465ns; rtta's requires 36ns. A bare python function call requires 35ns, so we're about as fast as fast can be.
Installation
From a repository checkout
Installation is still a little bit rough. We use cython and compile against numpy so you have to have both installed just to perform a pip install from source. In the coming days I'll upload binary packages for everything, but for now here's how you build the package.
pip install -U numpy cython
git clone git@github.com:adamdeprince/rtta.git
cd rtta
pip install -U .
We'll figure out how to make this pypi install-able in a few days.
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
Built Distribution
Hashes for rtta-0.0.1-cp311-cp311-macosx_12_0_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | ffcb71c7c32913ff0add61eb0d132fd30e3a211ec157728747b055b66eca6c0c |
|
MD5 | a2fe262364bb8c896fef6668deb7ea64 |
|
BLAKE2b-256 | b43a31bbafd78f5a50414c447e6cd58c903ca783f8098d920458d892cfcc9d14 |