browser.rb

Read the bookmarks and history of the web browsers on a machine — Brave, Chrome, Chromium, Firefox, Safari and Tor Browser — and render them to CSV, JSON or a property list. Bookmarks come through as one common representation whatever the browser; history as each browser stores it.

Add this line to your application's Gemfile:

gem 'browser.rb'

And then execute:

$ bundle

Or install directly:

$ gem install browser.rb

Note the .rb. The plain browser gem is someone else's — a user-agent parser — so this one is browser.rb. The require is the plain name, as it is for http.rb:

require 'browser'

The gem 'browser.rb' line is what makes that require load this gem rather than the other; the two ship the same file name.

require 'browser'

chrome = Browser.new(:chrome)
chrome.bookmarks.to_objects   # => [{title:, url:, folder:}, ...]
chrome.bookmarks.tree         # => Browser::Bookmark, the folders and all
chrome.bookmarks.to_csv       # title,url,folder rows
chrome.bookmarks.to_json
chrome.history.to_objects
chrome.profile_path           # => the directory the profile sits in

Name a browser by string or symbol, in any case:

Browser.new(:brave)
Browser.new('TorBrowser')
Browser.new('tor_browser')

Firefox and Tor Browser open the profile their profiles.ini names as the default, so none need be named. To reach another:

Browser.new(:firefox, profile_name: 'abcd1234.default-release').bookmarks.to_objects

Or opt into top-level names:

require 'browser-classes'
Chrome.new.bookmarks

The renderers are to_objects (an array of {title:, url:, folder:}), to_csv, to_json and to_plist. dump(format:, path:, filename:) writes one to a file and returns the path. readable? says whether the data can be read right now.

Underneath the renderers, bookmarks.tree is the bookmarks as Browser::Bookmark: a titleless root whose children are the browser's own roots, each folder holding its children. The renderers flatten that to leaves and drop the folders, so the tree is where a folder, or Bookmark#name, is to be had.

profile_path is the directory the profile sits in, which a Selenium front end wants in place of a throwaway one. Safari, having no profiles, answers nil.

• macOS only, for now. Every location is a macOS path; Linux and Windows are on the Todo.
• A running browser locks its database. History, and Firefox's bookmarks, are read from SQLite, which the browser locks while it is open — so reading a browser you are using raises Browser::Unreadable. Ask readable? first, or quit it.
• Which Firefox profile. Firefox and Tor Browser keep profiles in hashed directories, and the default is the one profiles.ini names, which is the one Firefox itself would open. A directory named plainly default is usually a leftover, the live one being <hash>.default-release, and the install entry in profiles.ini knows which. Two installations, say release and Nightly, each name a default, and then profile_name: is required.
• Tor Browser is looked for in ~/Library/Application Support/TorBrowser-Data. That is where an install in /Applications keeps its data. One installed elsewhere keeps it beside the application, and wants bookmarks_location: and history_location: passed to the constructor.
• Safari's Bookmarks.plist is a mirror. With iCloud bookmark sync on, a sync agent maintains it and writes it on change, so it can be absent while Safari is working perfectly. It is a binary property list, read here via CFPropertyList.

I realised when wanting to dump all bookmarks from any browsers on one machine for import to another that I'd already written something of the sort for history called dump_browser_history_to_csv and that it would probably be a good idea to combine those efforts into a single browser library to handle both bookmarks and history, as well as any other similar browser data extraction effort.

1. Fork it (https://github.com/thoran/browser/fork)
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 a new pull request

MIT