ricroberts

Deploying a sub-folder of a git repository with Capistrano

Sat, 05 Jun 2010 12:53:07 +0100

Sometimes it makes sense to have several Rails applications (or other ruby projects) in a single git repository. For example you might want to fork, tag or track changes on all the projects together (or you might just be too stingy to pay for a bigger Github plan).

The problem is that if you want to deploy just one of the applications to your server, unfortunately Capistrano doesn’t support deploying just sub-folders of git repositories. You can get round this by adapting the deploy_via :remote_cache option.

In the vendor/plugins folder of your project create the following folder structure:


/vendor
|- /plugins
   |- /remote_cache_with_project_root
      |- /recipes
      |- /lib
      |  |- /capistrano
      |     |- /recipes
      |        |- /deploy
      |           |- /strategy
      |              |- remote_cache_with_project_root.rb
      |- recipe.rb

Here are the contents of the files:

recipe.rb:

remote_cache_with_project_root.rb:

In your project’s deploy.rb file, set remote_cache_with_project_root as the deployment method, and set the name of the subfolder as the project_root option.

Video: TechCrunch/Techfluff.tv interview at EIE10

Sun, 23 May 2010 14:03:36 +0100

At the Engage Invest Exploit 2010 event in Edinburgh a couple of weeks ago, I was interviewed by Hermione Way about Swirrl’s new service, Publish My Data.

My interview is at around 4h 06m into the video. (Note: Safari seems to have to buffer the whole video before letting you skip, but Firefox lets you jump straight to any location.) I was completely unprepared, so please forgive my waffling!

PublishMyData: A New Platform for Publishing Linked Data

Wed, 31 Mar 2010 23:10:58 +0100

As Bill mentioned on his blog today, at Swirrl we’re working on a new platform for publishing Linked Data, called PublishMyData.

Please check out Bill’s post and let us know your thoughts on our initial ideas.

Problems with this blog's feed

Mon, 15 Feb 2010 10:08:48 +0000

You may have noticed a bunch of articles about the semantic web appear in this feed over the last couple of days. Apologies for any confusion.

It happened because my Swirrl co-founder, Bill, moved his blog over to this blog-engine at the weekend, and we had some caching/redirect issues on our server, so my blog was serving up Bill’s feed for a while!

I think that the feed should be sorted out now, but I’ll keep an eye on it for a bit to make sure. :)

Create Your Own Semantic Web-enabled Blog

Sun, 17 Jan 2010 18:20:49 +0000

A few weeks ago I moved this blog over to Swirrl’s own open-source Rails+CouchDB blog engine, SemanticJournal (At Swirrl, we like to call it ‘Semjo’ for short).

As well as the standard blogging features, Semjo includes some Semantic Web features, in the form of helpers to aid with marking up elements on public-facing pages with RDFa (such as the date, author, title of articles etc).

Earlier this week, I wrote a bit of documentation and made the Github repository public. So now, you can grab the code at github.com/Swirrl/SemanticJournal, and run your own Semjo blog. The readme file in the repo explains how to get up and running, including designing your own theme.

Current Features:

CouchDB Storage. Each blog has its own CouchDB database.
Helpers for marking up elements with RDFa
Host multiple blogs using the same rails app
Write articles with Textile markup
Design the theme used for each blog, using HTML and ERB
Caching, using Rack/Cache
Just provides the blog engine only but…
Easy to extend with 3rd party plugins (e.g. Disqus for comments, ShareThis for social bookmarking, Google for search etc., Gists for Code/Syntax highlighting, S3 for storing other assets used on your blog).

Semjo is still very much work in progress. Please let me know if you encounter any bugs, or if there are any features you would like to see (the README contains my current list of todos).

SemanticJournal blog engine

Wed, 30 Dec 2009 17:07:07 +0000

Sorry if I’ve introduced lots of unread articles into your feed reader… I moved this blog over to a new blog engine yesterday.

Previously, it was running on Typo, an open source Rails-based blogging app, which has served me well for a few years, but at times I found it a little frustrating and inflexible to work with. It’s now running on my own blogging platform codenamed SemanticJournal – still based on Rails, but using CouchDB for storage rather than MySql.

SemanticJournal was written partly as an exercise in using CouchDB, and partly to serve as a testing ground and vehicle for some new technology that we’re working on at Swirrl. As the name suggests, SemanticJournal combines some Semantic Web features with a blogging platform.

If you view the source for this page (or use an RDFa-inspector like this bookmarklet), then you’ll notice that certain elements are marked-up with RDFa.

RDFa is just a way of enhancing web pages with additional metadata by combining html with RDF-triples. (i.e. machine-readable statements about resources consisting of subject, predicate and object). For example, one such statement embedded in this page describes when the article was published.

What’s the point of adding RDFa to a web page? Using RDFa helps other computers understand its content more easily, as it gives the marked-up text a context. Terms mentioned in the page can be made less ambiguous (e.g. If I mention “Paris Hilton”, am I talking about the celebrity or the hotel?). By providing some of the content as RDF it becomes more easily accessible by other services on the web, making it more “mashup-able” and providing potential SEO benefits.

At the moment, SemanticJournal only adds RDFa to the basic blog article metadata, such as the title, author and date, but at Swirrl, we’re working on providing ways to make it easier to publish machine readable content on the Web.

Once I’ve had a chance to tidy up the code a bit (and make it a bit more generically useful) I plan on making the code for SemanticJournal openly available. Watch this space – I’ll announce it on this blog when we do.

DailyJS: a new Javascript blog

Thu, 05 Nov 2009 00:00:00 +0000

Alex Young and I (with a bit of help from Justin (Gus) Knowlden – of Riot fame), have started up a new Javascript blog called DailyJS.

Although still chiefly a Ruby developer, I’ve had much more to do with Javascript recently, but I’ve had trouble finding a technically-focused blog (i.e. along the lines of RubyInside, but for Javascript). Blogs like Ajaxian are cool, but they’re quite general, and focus on high-level topics.

So, DailyJS.com was born. Check it out for regular Javascript-related code, tips and news. It might not turn out to actually be updated daily, but we’ll try to keep posts coming fairly frequently …it will certainly be available daily, anyway! :)

If you want to get involved with contributing to, or sponsoring, DailyJS please get in touch.

Using a different Couch DB for each web-request with CouchRest

Sun, 01 Nov 2009 00:00:00 +0000

If data for each different account in your web application is stored in a separate CouchDB database, you need a way to tell your web app which database to use for that request.

Here’s solution using CouchRest and Ruby, based on a discussion from the CouchRest Google Group.

class RequestSpecificDocument < CouchRest::ExtendedDocument
  def self.database
    Thread.current[:request_db]
  end
end

All models which need to use the request-specific database should derive from this:

class Article < RequestSpecificDocument
  # normal stuff for your CouchRest document here.
end

So in your application controller (if you’re using Rails), you might add a before_filter which does something like this:

def set_database
  # determine the database_name from the host, url or whatever. 
  svr = CouchRest::Server.new("http://127.0.0.1:5984")
  couch_db = CouchRest::Database.new(svr, database_name)        
  Thread.current[:request_db] = couch_db
end

Stardotstar blog article: Relax, Stop worrying about the structure of your data

Sun, 25 Oct 2009 00:00:00 +0100

I’ve been working as a Ruby developer with Manchester digital agency, Stardotstar, for a few months now. Last week, I wrote my first post on the new Stardotstar blog:

When thinking about web application design, it doesn’t take long before the realisation hits you: to achieve any kind of efficiency you need to ignore or abuse key principles of relational database design, employing strategies such as sharding your information across multiple databases, abandoning the use of transactions, and the wholesale denormalisation of data…

Read the full article: Relax: Stop worrying about the structure of your data

Blink #3 - 24th October 2009

Sat, 24 Oct 2009 00:00:00 +0100

It’s been a while since my last link-blog entry, so this will be quite a big list!

Ruby

Mustache – for logic-free views in your Ruby web apps
Watchr – a replacement for autotest
New Thoughtbot-designed Gemcutter site launched
Getting ImageMagick and Passenger to play nicely together on OS X
MacRuby approaching a Milestone – Blog post by Matt Aimonetti.
Loofah – HTML sanitizer based on Nokogiri.
Old, but still neat and useful: Subdomain routing in rails
Tilt – a generic interface to various Ruby templating engines.
Riot – A fast unit test framework for Ruby
Ruby code-beautifier TextMate Bundle

General Web/Software

Google ChromeFrame for sabotaging IE (this feels like old news now!)
We are colorblind. About 8% of the male population has some sort of color blindness (like me). The patterns and examples on this site help you create websites that we color blind can use without problems
Pimpin’ (Software Projects) Ain’t Easy – Great marketing-software-for-non-marketers presentation by Amy Hoy
There’s only Metadata and URLS – Interesting blog post by Tom Scott.

Git

Github moves to rackspace and ditches gem building
Tips from @Cauis on .gitignore

CouchDB

Why CouchDB is the filesystem for the Web
CouchDB goes beta!
Mozilla Raindrop – a messaging system built on CouchDB
Some CouchDB performance tips from Till Klampaeckel