Showing posts with label Python-for-science. Show all posts
Showing posts with label Python-for-science. Show all posts

Monday, December 9, 2013

Domino, a Python PaaS for data science

By Vasudev Ram


Domino (dominoup.com) is a new Platform as a Service (PaaS) for data science. It supports Python, the R language and Matlab.

I saw it via a retweet by Cameron Laird (@Phaseit).

I signed up and tried Domino, using their sample Python project. It prints "hello", and plots a simple graph using matplotlib.

To do this, I had to download their command-line tool called domino; this is a one-time task. Then I used it to download the project to my PC.

Then I ran it on Domino, with "domino run main.py". This gives you a message that the run has started on their server, and prints a link where you can check the status of the run. The run command will also upload the project files if they are newer on your machine than on Domino. Running "domino" gives the help for the command, and they also have a help section on their web site with a lot of info.

I did encounter a couple of issues, one is fixed, the other is minor and will soon be fixed.

Here is a screenshot of the results:


They have an interactive web wizard that guides you through the steps of creating and running your first project, even though the steps are quite simple. And that is one thing I like about Domino. They seem to be trying to take the pain (and routine work) out of deployment and running of apps in the cloud. In fact, that's what they say on their home page, or words to that effect. Also, I've exchanged several emails with the founders, regarding various questions I had about Domino and its capabilities, and their responses have been prompt and helpful.

Domino also emails you a notification when the run ends, with a link you can click to see the results page, including the output, which you can download.

I was interested to see that Domino uses Docker, which has been in the news a lot lately.

Docker on Wikipedia.

An article on Linux.com about Docker: Docker: A 'Shipping Container' for Linux Code

Domino uses Docker - a post by Domino on the Docker blog.

[ Update: I checked a bit more and saw that Docker uses LXC (LinuX Containers), which in turn use cgroups, a Linux kernel feature originally written by two people at Google. according to Wikipedia. ]

Domino Data Lab on Twitter.

One of their tweets (in response to a query) says, "Perl coming soon. C on the list".

- Vasudev Ram - Dancing Bison Enterprises

Contact Page




Thursday, December 20, 2012

Continuum gets $3 mil from DARPA for Python science and big data tools

Technical Discovery: Passing the torch of NumPy and moving on to Blaze

Seen via Python Weekly newsletter.

I had blogged somewhat recently about Continuum Analytics and Wakari:

http://jugad2.blogspot.no/2012/11/wakari-scientific-python-in-cloud.html

- Vasudev Ram

Friday, December 7, 2012

PyRSS2Gen, to create RSS feeds Pythonically

PyRSS2Gen

It is by Andrew Dalke of Dalke Scientific, a company that develops Python tools for science:

http://www.dalkescientific.com/company.html

I saw PyRSS2Gen used on this blog:

http://blaag.haard.se/

- Vasudev Ram
www.dancingbison.com

Thursday, November 29, 2012

Spyder, a Python development environment for science

spyderlib - Spyder is the Scientific PYthon Development EnviRonment - Google Project Hosting

http://code.google.com/p/spyderlib/wiki/Features

http://packages.python.org/spyder/editor.html

Spyder uses pylint, pyflakes and rope, all of which are Python development utilities.

I used lint (which is for C) a lot earlier, and it is quite useful for improving your code.

Got to check out pylint and pyflakes.

Vasudev Ram
www.dancingbison.com