Skip to main content

Stochastic Neighbor and Crowd Kernel (SNaCK) embeddings: Quick and dirty visualization of large-scale datasets via concept embeddings

Project description

Stochastic Neighbor and Crowd Kernel (SNaCK) embedding
======================================================
Quick and dirty visualization of large-scale datasets via concept embeddings

Installation
============
The following platforms are supported:
- Python 2.7 on Linux
- Binary packages available on Conda
- Source packages available from Pip
- Python 2.7 on OSX
- Binary packages for Yosemite available on Conda
- Source packages available from Pip (Homebrew-GCC required)

Linux: Install from Conda
-------------------------
Just run:
$ conda install -c https://conda.anaconda.org/gcr snack

Mac OS X: Install from Conda
----------------------------
TODO.

Linux: Install from Pip
-----------------------
Just run:
$ pip install snack

You need to install Python 2.7, Numpy, and Cython. You also need a
working compiler, CBLAS, and the Python development headers, which are
installable from your distribution's package manager.

To install SNaCK on a clean Ubuntu Trusty x64 system, run:

# sudo aptitude install \
build-essential \
python-dev \
libblas3 \
libblas-dev \
python-virtualenv
$ virtualenv venv; source venv/bin/activate
$ pip install numpy
$ pip install cython
$ pip install snack

OS X: Install from Pip and Homebrew
-----------------------------------
If you are on Mac OS X, you must install the real "not-clang" version
of gcc because it has OpenMP support. At the time of writing, clang
does not support OpenMP, and Apple has unhelpfully symlinked clang to
`/usr/bin/gcc`. This is not sufficient.

Using Apple-provided GCC is NOT supported. If `gcc-5 --version`
contains the string `clang` anywhere in its output, you do not have
the correct version of gcc.

Using Apple-provided Python is NOT supported.

The recommended installation method on OS X is with Homebrew:

$ brew install gcc
$ brew install python
$ virtualenv venv; source venv/bin/activate
$ pip install numpy
$ pip install cython
$ pip install snack

You may need to edit `setup.py` and change `GCC_VERSION` to point to
the correct version, if you are not using `/usr/local/bin/gcc-5`.


Just build without installing
-----------------------------

To simply build Snack without installing it, run:

$ python setup.py build_ext --inplace

This builds `snack/_snack.so`. You can move the `snack` folder to your
project's directory and then `import snack`. This should work as long
as the `snack` folder is inside your current directory.

How to use
----------

Examples
--------

See also
--------

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

snack-0.0.3.tar.gz (155.7 kB view details)

Uploaded Source

File details

Details for the file snack-0.0.3.tar.gz.

File metadata

  • Download URL: snack-0.0.3.tar.gz
  • Upload date:
  • Size: 155.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No

File hashes

Hashes for snack-0.0.3.tar.gz
Algorithm Hash digest
SHA256 98996b14811a9ce5efad938c32823bff998de104901982d98f214c73f9407096
MD5 566ee458296b7c1650a8d848b5f5ac86
BLAKE2b-256 15527889541bad6bd053c6554e3158c91cf0d4092b251b34de5fa35e20040968

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page