I write scientific software in Numpy/Scipy/Matplotlib. Getting developed programs on my small desktop computer, I'm now thinking about writing simple web programs. Example: user uploads image or audio file, my program processes it using Numpy/Scipy, and output is shown on the browser using Matplotlib, or possibly the consumer can download a processed file.

I already purchase hosting that comes with Python 2.4.3 installed, but no Numpy/Scipy. I do not have spend access via command line, either. Just drag-and-drop FTP. Pretty limited, however i could possibly get simple Python/CGI scripts working.

Remarkably, an internet search revealed couple of appropriate choices for website hosting with one of these abilities already built-in. (Please guide me should i be wrong.) I'm researching the Google Application Engine, however i still do not have a complete understanding about its tools and restrictions. Exactly what the web did let me know is the fact that others have similar concerns.

Wishing for solutions, I figured I'd request these a quick question towards the awesome SO community:

  1. It is possible to simple method of setting up numpy (or any third-party package/library) onto my already located space? I understand the Python path on my small located space, and that i be aware of relevant Python/Numpy sites on my small desktop computer. Can One simply copy files over and also have it work? Both local and remote systems run Ubuntu.

  2. What hosting sites exist (either free or compensated) that have Numpy/Matplotlib installed or, otherwise installed, the potential of setting up it? What are the recorded sites that you could reference with working programs, regardless of how simple?

  3. Can Google Application Engine assist me to by any means? Or perhaps is it totally for another thing? Perhaps you have varieties tried on the extender to create scientific programs in Python/Numpy? If that's the case, would you reference them?

Appreciate your help.

EDIT: Following the helpful solutions below, I purchased the $20 plan at Slicehost, and that i like it to date! (When i first attempted Amazon . com EC2. I have to be stupid, however i just could not get results.) Establishing the Ubuntu server with Apache required mere hrs (and I am an Apache novice). It enables me to complete precisely what I needed with Python and even more. I are in possession of my very own remote repository for version control, too. Many thanks!

Application Engine doesn't support some of numpy, scipy, or matplotlib, alas.

Knowing precisely what OS and CPU your host is applying, you can make the same installation on your own, download and install exactly the same version of Python they are using, download the causes of packages you need and make them into .so (or .pyd, with respect to the platform) files, and upload individuals -- seems like a genuine tour p pressure.

The many, many sites that provide normal virtual hosting (an online machine, typically Linux, with modest HW assets, but root rights for you personally, ssh spend access, along with a gcc you should use particularly) is going to be much simpler to utilize -- basically, you'll download and install the program you'll need nearly exactly the same way you'd do by yourself Linux workstation!