perspective-python 0.4.2

Python bindings and JupyterLab integration for Perspective

perspective-python

Python APIs for perspective front end

Install

Dependencies

PyArrow==0.15.1 and Numpy are required build-time dependencies, and can both be installed via pip.

pip install pyarrow==0.15.1

It is recommended to have https://github.com/intel/tbb installed as a system dependency:

On MacOs:

brew install tbb

On CentOS/RHEL:

yum install tbb-devel

On Ubuntu:

apt-get install libtbb-dev

On Windows:

choco install tbb

Or install from source.

All other dependencies will be vendored if missing as part of Perspective's CMake build, but you can install them yourself and CMake will detect them via the usual mechanisms.

Installation

To install the base package from pip:

pip install perspective-python

To install the JupyterLab extension:

jupyter labextension install @finos/perspective-jupyterlab

Getting Started

Python User Guide Python API Examples

Developing

To build perspective-python from source, you'll need the following C++ dependencies:

• Python 3.7
• numpy
• CMake
• PyBind11
• tbb

On MacOS, you should be able to install Boost, PyBind11, and tbb from brew:

brew install pybind11 tbb

And then install Python dependencies using pip:

python3 -m pip install -r requirements-dev.txt

Then run yarn build, and if a .perspectiverc config file has not been created yet, you'll be taken through the yarn setup script. On the Focus section, type p for python, and continue through the script. Run yarn build again, and the python build should begin.

If you already have a .perspectiverc and want to reset your build configuration, simply run yarn setup.

Once the build successfully completes, run yarn test_python from the project root in order to verify operation.

Windows

By default, perspective attempts to build utilizing Microsoft Visual Studio 2017 (MSVC 14.1). You may change this to older versions by editing setup.py.