PRX Podcast User-Agent Parser

prx.github.io/prx-podagent

Description

Basic User-Agent string parser that includes some basic podcasting apps. This project is intended to help parse/group requests for analytics purposes, not for browser feature detection.

The included agents.lock.yml also includes name/type/os IDs, in case you want to normalize the strings in your database.

Install

Just npm install --save prx-podagent. Or to use outside of node, just grab a database file out of the db/ directory.

Usage

const podagent = require('prx-podagent');
const agent = podagent.parse('some-string');
if (agent) {
  console.log('Match:', agent.name, agent.type, agent.os);
} else {
  console.log('Did not match any known agents');
}

Or to DIY use the json lock file:

  $.getJSON('agents.json', function(db) {
    db.agents.forEach(function(a) {
      a.regex = new RegExp(a.regex, a.ignorecase ? 'i' : undefined);
    });

    var str = decodeURIComponent('some-agent-string');
    var agent = db.agents.find(function(a) { return a.regex.test(str); });
    if (agent) {
      console.log('matched!', agent);
    } else {
      console.log('no match for:', str);
    }
  });

Or in Ruby:

require 'yaml'

DB = YAML.load_file('db/agents.lock.yml')
def match_agent(str)
  str = URI.unescape(str)
  DB['agents'].find { |a| Regexp.new(a['regex'], a['ignorecase'] ? 'i' : nil).match(str) }&.tap do |match|
    %w(name type os).each { |k| match[k] = DB['tags'][match[k]] }
  end
end

puts match_agent('Pandora/1812.2 Android/5.1.1 ford (ExoPlayerLib2.8.2)').inspect
# {"regex"=>"/^HardCast.+CFNetwork/", "name"=>"HardCast", "type"=>"Mobile App", "os"=>"iOS"}
puts match_agent('blah blah blah').inspect
# nil

Or in PHP:

URI Encodings

Note: it's fairly common to see URI encodings in user agent strings. And often depends on your particular server setup. The regexps in this library are intended to be used on the fully-uri-decoded User-Agent string. So you should always decodeURIComponent() the value before attempting to match.

I've also seen a mix of encoded/decoded spaces within a single user agent string. Where the first space has been encoded to %20, but subsequent ones are not. So URI decoding is probably a good idea anyways.

Development

Basic tests are located in the /test directory, and can be run with npm test.

Additionally, there is a test/support/testagents.csv file containing some actual production logs. The "coverage" and "omissions" tests use this file to check that the database file accounts for all the major known user agents.

To add a new user agent:

1. Edit the db/agents.yml file to include your new regular expression, plus an example user-agent string or two.
2. Run npm test (or just mocha test/examples-test.js) to test that your example strings match the regexp.
3. Run npm run lock to regenerate the db/agents.lock.yml. This file normalizes the text tags/labels shared between the various matchers. Check that your change didn't add any unexpected new tags (if you accidentally changed the case of a label or something).
4. Create a pull request to this repo.

To release a new version:

1. Get or set a GITHUB_TOKEN ENV (needed for release-it)
2. Run npm run release
3. Select whether this is a major/minor patch, according to semantic versioning
4. Select "yes" on publishing to NPM, Pushing, and creating a Github Release

License

MIT License

Contributing

1. Fork it
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create new Pull Request