An RKHS based module for numerics, statistic and machine learning
Project description
Overview
Codpy is a kernel based, open source software library for high performance numerical computations relying on the RKHS theory.
It contains a set of core tools that we use for machine Learning, statistics and numerical simulations, see our introduction to codpy for a review of the method, as well as several examples running this library.
Please refer to ReadTheDoc for the technical documentation.
Warning: codpy versions 0.1.XX are alpha versions in early development stage and will be subject to rapid changes without down compatibilities.
Technical requirement
This version of the library is multi-core CPU architectures, and is tested on
- windows / amd64 platforms
Installation
Note: this installation process has been tested on
- windows / amd64 platform
prerequisite
Minimum installation
- python3.9.XX: a valid python python 3.9 installation.
NOTE : Python installations differ from one machine to another. The python root folder is denoted "<path/to/python39>" in the rest of this document. The software Everything, or other finding files tools can be useful locating the file python.exe on windows machine...
Dev installations
For information, we list the softwares that we are using for our dev configuration :
- GitHub Desktop
- R: use any CRAN mirror for download
- RStudio: see the download link, then choose the free version
- MiKTEX: see the download tab
- Everything
- Visual Studio Code
Those installations should be fine using the latest (64 bits) version and the default settings for each software .
Note Once R and RStudio are installed, open the latter. In the console, enter "install.packages("rmarkdown")" to install RMarkdown.
Cloning repo
Download the codpy repo at codpy alpha to your location <path/to/codpyrepo>
Installation
prerequisite
We suppose that there is a valid python installation on the host machine. The reader can
- either use its main python environment
<path/to/python39>
- or create a virtual python environment
<path/to/venv>
, generally an advisable practice.
First open a command shell cmd
, create a virtual environment and activate it using the commands
python -m venv .\venv
.\venv\Scripts\activate
NOTE : In the rest of the installation procedure, we consider a virtual environment <path/to/venv>. One can replace with <path/to/python39> if a main environment installation is desired, for dev purposes for instance.
pip install codpy
Open a command shell cmd
, and pip install codpy
pip install codpy==0.XX.XX
or from the local repository
pip install <path/to/codpyrepo>/dist/codpy-XXXX.whl
The installation procedure might take some minutes depending on your internet connection.
Test codpy
open a python shell and import codpy
python
import codpy
Testing with Visual Studio Code
You can your visual studio installation.
-
With Visual Studio Code, open the
<path/to/codpyrepo>
folder and select for instance the file<path/to/codpyrepo>/test/1NN_estimation_rate.py
-
If required, select your python interpreter to the virtual environment one (Shift+P)
-
Hit F5. If everything works, you should have some figures after one or two minutes.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
Hashes for codpy-0.1.13-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | da5da3e9dea1631a7590462d4f175801e8e51300a854d063963a4871dc2898fb |
|
MD5 | 63e4ec8721886472478c415de32fb185 |
|
BLAKE2b-256 | 3f3d0275deef06a695cfc64485dbb659d1d2ef40eafd65fa99859e0d4152bdf9 |