Orange3 - 开源机器学习和数据可视化工具

7,885 阅读1分钟
原文链接: github.com

build: passing codecov

Orange is a component-based data mining software. It includes a range of data visualization, exploration, preprocessing and modeling techniques. It can be used through a nice and intuitive user interface or, for more advanced users, as a module for the Python programming language.

This is a development version of Orange 3. The stable version 2.7 is still available (binaries and sources).

Installing

This version of Orange requires Python 3.4 or newer. To build it and install it in a development environment, run:

# Install some build requirements via your system's package manager
sudo apt-get install virtualenv git python3-dev g++ gfortran \
                     libblas-dev liblapack-dev libatlas-base-dev

# Also install Qt dependencies for the GUI
sudo apt-get install python3-pyqt4

# Create a separate Python environment for Orange and its dependencies,
# and make it the active one
virtualenv --python=python3 --system-site-packages orange3venv
source orange3venv/bin/activate

# Clone the repository and move into it
git clone https://github.com/biolab/orange3.git
cd orange3

# Install the minimum required dependencies first
pip install numpy
pip install scipy
pip install -r requirements-core.txt  # For Orange Python library
pip install -r requirements-gui.txt   # For Orange GUI

pip install -r requirements-sql.txt   # To use SQL support
pip install -r requirements-opt.txt   # Optional dependencies, may fail

# Finally install Orange in editable/development mode.
pip install -e .

Installation of SciPy and qt-graph-helpers is sometimes challenging because of their non-python dependencies that have to be installed manually. More detailed, if mostly obsolete, guides for some platforms can be found in the wiki.

Anaconda Installation

First, install Anaconda for your OS (Python version 3.5+). Create virtual environment for Orange:

conda create python=3 --name orange3 

In your Anaconda Prompt add conda-forge to your channels:

conda config --add channels conda-forge

This will enable access to the latest Orange release. Then install Orange3:

conda install orange3

Starting Orange GUI

Orange GUI requires PyQt, which is not pip-installable in Python 3. You have to download and install it system-wide. Make sure that the virtual environment for orange is created with --system-site-packages, so it will have access to the installed PyQt4.

To start Orange GUI from the command line, assuming it was successfully installed, run:

orange-canvas
# or
python3 -m Orange.canvas

Append --help for a list of program options.

Windows dev setup

Windows + GCC:

python setup.py build_ext --inplace --compile=mingw32