This is a set of experiments to see ways in which Python and Javascript can be combined.

At the moment, it implements a simple function to perform a 2D affine transformation on 128,000 points in a number of different ways in order to compare the performance and benefits of different approaches.

In the future, this test example should be made more complex and involve classes and complex interactions between them.

Below is a list of the different benchmarks available, and how to build and run them.

Implemented in pure Python.

Running::

./test_pure_python.py

Implemented using vector operations on Numpy arrays.

Requirements:

• Numpy <http://www.numpy.org>_

Running::

./test_numpy.py

Implemented using a C extension. The assumption here is that the data is in native C format (by, for example, using Numpy arrays. Actually placing values in Python objects inside of Python lists would be much slower, but this test provides a "best case" scenario.

Requirements:

• Python development headers (apt-get install python-dev or yum install python-devel). The location of the headers is hard-coded in the build file.

• The gcc compiler

Building::

./make_c

Running::

./test_c.py

This tests a hand-written JavaScript implementation.

Running:

Open test_raw_js.html in your browser. Press the "Run" button and the benchmark results will be logged to the Javascript console.

Skulpt <http://www.skulpt.org>_ is a Python interpreter written in Javascript that runs in the web browser.

Requirements:

• A copy of skulpt is included in the py-js-blending-experiments source tree.

Building::

./make_skulpt

Running:

Open test_skulpt.html in your browser. Press the "Run" button and the benchmark timing will be output to the JavaScript console.

PyJS <http://www.pyjs.org>_ (formerly Pyjamas) is a Python-to-Javascript converter, as well as a framework for developing client-side web applications in Python.

Requirements:

• Download PyJS and run the bootstrap.py script.

Building::

export PYJS_HOME=/path/to/pyjs_checkout ./make_pyjs

Running:

• Start a simple webserver using the serve.py script.

• Open http://127.0.0.1:8890/output/test_pyjs.html in your browser.

PyV8 <http://code.google.com/p/pyv8/>_ embeds the open source V8 Javascript engine (from Google Chrome) in the CPython interpreter. It allows for running Javascript code from Python, calling into Python code from Javascript and sharing objects somewhat transparently between the two sides.

Requirements:

• Download and install PyV8

Running::

./test_v8.py