github

RHACK is Ruby Http ACcess Kit: curl-based web-client framework created for developing web-scrapers/bots.

• Asynchronous, still EventMachine independent. Synchronization can be turned on with fine of 1sec per [bulk] request
• Fast as on simple queries as on high load. Can process up to thousands (limited by a net interface, of course) different parallel requests of any HTTP method with no penalty
• Flexibly configurable on 3 levels:
  • Curl::Easy (simplest request configuration, inherited from curb gem)
  • ::Scout (Curl::Easy wrapper with transparent cookies and extendable anonimization processing, detailed request/response info, callbacks and retry configuration)
  • ::Frame (Scout array wrapper with smart request interpretor, load balancing and extendable response processor)
• Support of javascript processing on loaded html pages is included (johnson gem)
• Web-service-client abstraction implementing some examples of how to use this library

It's still randomly documented since it's just my working tool.

• Documented examples with Scout, Frame and Client
• Tests for Scout and Frame to interpret requests
• Tests for Scout, Page and Page subclasses to process fictive results
• Add some transparent control on user-agents: desktop, mobile, randomly predefined...

• ::Scout

  • Added #retry! method and fixed raise/retry workflow

• ::Page

  • Added #retry? method to organize retry behaviour in subclasses
  • RHACK#ReloadablePage is deprecated now

• Config

  • rhack.yml generator: rake rhack:config
  • Commented out "db" part and suppress require "rhack/storage" without Redis being loaded

• Curl

  • 5xx HTTP response will call on_server_error callback instead of on_failure
  • Fixed Segmentation Fault on `void curl_multi_free()'

• ::Page

  • Fixed #expand_link for partial links

• Make Curl.status catch any Exception subclass

• ::Frame

  • Moved Curl.execute from initialize to on after request added
  • #initialize option :scouts aliased as :threads

• ::ScoutSquad

  • Finally stabilized #next and #rand time management for parallel recursive execution

• ::Frame

  • Added #anchor

• ::Scout

  • Fixed #update
  • Catch weird Curl::Err::CurlOK being thrown on some pages

• Fixed some exceptions messages

• ::OAuthClient < ::Client

  • A full set of abstract OAuth2 authorizaztion and API methods
  • Per-user key-value oauth_token storage
  • Handling of tokens expiration
  • Fits for, at least, facebook.com, linkedin.com and vk.com

• ::Storage

  • Wrapper of Redis-based storage to handily store/cache scrapers data

• ::Frame

  • #initialize: ::ScoutSquad size can be specified by :scouts option (default still is 10)
  • @static with Hash value is now essentially a default route. :protocol and :host values are used for request where only "path" url is given
  • Fixed weird #run_callbacks! and :raw behaviour. @res of a Page now always get result of a block passed as :proc_result or block passed to #exec itself

• ::Scout

  • Added explicit cacert loading, cacert.pem by curl.haxx.se lies in /config
  • Provided support of curl HTTP DELETE and PUT verbs: #loadDelete and #loadPut. From a frame, add :verb => (:delete|:put) as an option to #run.

• ::ScoutSquad

  • Automatically Curl.execute on #next and #rand if Carier Thread is exited without an exception

• ::Service

  • Is renamed to Client what is more sensible. RHACK::Service is still usable as alias
  • require 'rhack/clients' <-> require 'rhack/services'

• Structural changes

  • Updated and documented rhack.yml.template that now lies in /config
  • All initialization moved to /lib/rhack.rb, rhack_in.rb stays there for compatibility
  • The gem is now being produced in the bundle style: added Gemfile, .gemspec, etc
  • ::Frame#get_cached and all its methods related to #dl (downloading) moved to rhack/dl
  • Global variables is replaced by module attributes of respective names

• Persistence

  • Removed every reference to SQL: /cache.rb, /words.rb (removed) and /extensions/declarative.rb (moved to rmtools project as optional part)
  • Made a foundation for Redis support, redis-based storage itself is coming the very next version

• Added rake redis:config: generate rhack.yml -> redis.conf

• Сhanged ::Frame @static behaviour, :static option now accept hash with :procotol key (see ::Frame#validate comment)
• Changed log level in curl-global.rb
• Described the library and marked down this readme

• Fixed bugs

  • idle execution in Rails application thread
  • Curl::Easy default callback
  • some misspelling-bugs

• ::ScoutSquad

  • Minified #next waiting time

• ::Service

  • added meta-methods #login (sync only) and #scrape!(<::Page>)

• ::Frame

  • made new cache prototype. Call #use_cache!(false?) for (in)activate and #drop_cache! for clearance
  • added :xhr exec option

• ::Page

  • #title returns full title by default
  • #html is auto-encoded to UTF-8 during #process

• Adjusted cookie processor in accordance with web-servers and entrust redirection process to ::Scout
• Added some shortcuts to Frame and Curl modules
• Сonfig defaults are now taken from rails
• Removed crappy database usage from lib/words.rb
• curb_multi.c: Moved callbacks out of rb_rescue so that I could know wtf was happen there

• Nastily pulled down curb-0.8.1 extension sources and harshly patched by changes made long before, so that the core will be as modern as possible and with necessary features
• Fixed syntax for Ruby 1.9

• A long time ago in a galaxy far, far away...
• A library had been created based on Net::HTTP
• In a few months its base had been changed by curb-0.4.4 because of poorness and incovinience of Net::HTTP
• Had been made background-mode for Curl::Multi and multipart body setting for Curl::Easy so that Curl could be both sync and async
• Had been added a couple of wrappers for Curl::Easy and its results, proxy lists processor, scrapers for a few web-services, and plugin for libxml-ruby that lives at rmtools gem now

RHACK is copyright (c) 2010-2013 Sergey Baev tinbka@gmail.com, and released under the terms of the MIT license. See the LICENSE and CURB-LICENSE files for the details. Rhack includes slightly modified Curb gem extension source code. For original Curb gem code you may want to check ext/curb-original directory or visit http://github.com/taf2/curb/tree/master.