Installation and Getting Started

scikit-chem is easy to install and configure. Detailed instructions are listed below. The quickest way to get everything installed is by using conda.

Installation

scikit-chem is tested on Python 2.7 and 3.5. It depends on rdkit, most of the core Scientific Python Stack, as well as several smaller pure Python libraries.

Dependencies

The full list of dependencies is:

The package and these dependencies are available through two different Python package managers, conda and pip. It is recommended to use conda.

Using conda

conda is a cross-platform, Python-agnostic package and environment manager developed by Continuum Analytics. It provides packages as prebuilt binary files, allowing for straightforward installation of Python packages, even those with complex C/C++ extensions. It is installed as part of the Anaconda Scientific Python distribution, or as the lightweight miniconda.

The package and all dependencies for scikit-chem are available through the defaults or richlewis conda channel. To install:

conda install -c richlewis scikit-chem

This will install scikit-chem with all its dependencies from the author’s anaconda repository as conda packages.

Attention

For Windows, you will need to install a dependency, fuel, separately. This will be made available via conda in the future.

Using pip

pip is the standard Python package manager. The package is available via PyPI, although the dependencies may require compilation or at worst may not work at all.

pip install scikit-chem

This will install scikit-chem with all available dependencies as regular pip controlled packages.

Attention

A key dependency, rdkit, is not installable using pip, and will need to be installed by other means, such as conda, apt-get on Linux or Homebrew on Mac, or compiled from source (not recommended!!).

Configuration

scikit-chem

Currently, scikit-chem cannot be configured in a config file. This feature is planned to be added in future releases. To request this feature as a priority, please mention it in the appropriate github issue

Fuel

To use the data functionality, you will need to set up fuel. This involves configuring the .fuelrc. An example .fuelrc might be as follows:

data_path: ~/datasets
extra_downloaders:
- skchem.data.downloaders
extra_converters:
- skchem.data.converters

This adds the location for fuel datasets, and adds the scikit-chem data downloaders and converters to the fuel command line tools.