There are plenty of libraries that can parse and detect user agents in a generic way.
PodIdent
takes a different approach here: it puts a lot of effort into getting the most used Podcast clients right, leaving the job of detecting browsers and more obscure
devices to generic user agent detecting libraries.
Version numbers might be interesting for some use cases, but are not relevant at all to podcasters - podcasters want to know where their listeners are, and what platforms / clients work the best.
There is a great deal of confusion about what is a device, an operating system or a platform in general. To keep things simple, PodIdent
uses platform
as a generic term to describe the place where software runs. It can be an iPhone (device), iOS (operating system) or even just an "Apple device" (category).
PodIdent
has no production dependencies. Keep it simple, and all that.
User agents are known to be confusing and difficult to understand. Most vendors don't document their user agents explicitly, and most of the detection rules are created by trying and collecting information from other sources.
This is why PodIdent
tries to be always as precise as possible (if it knows that a client runs on an iPad, it will return platform: 'iPad'
, instead of a generic iOS
.
The detection rules always try to match substrings first, before using full-fledged regular expressions. This should make detection a lot faster. Also, since there are not even 100 podcast clients out there, there is no need to support a billion different browsers and devices.
Add this line to your application's Gemfile:
gem 'pod_ident'
And then execute:
$ bundle
Or install it yourself as:
$ gem install pod_ident
require 'pod_ident'
user_agent = 'AppleCoreMedia/1.0.0.10B500 (iPod; U; CPU OS 6_1_6 like Mac OS X; en_gb)'
result = PodIdent::Detector.detect(user_agent)
puts result.app
=> 'Apple Podcasts'
puts result.platform
=> 'iPod'
require 'pod_ident'
user_agent = 'Mozilla/5.0 (compatible; archive.org_bot http://archive.org/details/archive.org_bot)'
result = PodIdent::Detector.detect(user_agent)
puts result.app
=> 'Archive.org'
puts result.bot?
=> true
After checking out the repo, run bin/setup
to install dependencies. Then, run rake spec
to run the tests. You can also run bin/console
for an interactive prompt that will allow you to experiment.
To install this gem onto your local machine, run bundle exec rake install
. To release a new version, update the version number in version.rb
, and then run bundle exec rake release
, which will create a git tag for the version, push git commits and tags, and push the .gem
file to rubygems.org.
The detection rules are found as a yaml file under lib/detection_rules.yml
.
Please, only edit this file, and afterwards run bin/parse-rules
in order to generate both the detection rules as a rb
file for production usage, and the complete rules with their corresponding test cases for testing.
For the bot detection rules we are using user-agents-v2, to import the json of this project and transform it into a yaml file readable by pod-ident you can run the following task: rake update_bot_rules
.
Once the yml file created run bin/parse-rules
to generate all rb detection rules files for bots.
Bug reports and pull requests are welcome on GitHub at https://github.com/podigee/pod_ident.rb.
It would be nice to have the following features in future versions:
- bot & crawler detection (instead of having to use a generic user agent parser)
- adding custom detection rules (this can be achieved today by pushing rules into the rules array, but there should be a better way)