* Retrieve members by processing MemberEvents to counter API change

* Removed the request caching layer. Requests are cached in the persister
* Change default DB isolation to REPEATABLE READ for stronger isolation
* Finer-grained (commit level) transactions when processing forks
* More accurate and uniform logging
* Tool to push logs to InfluxDB for monitoring
* Drop ext_ref_id from all tables
* More efficient retrieval of events, 100 in one go
* Tool to retrieve all user details and support for marking users deleted
* Support for retrieving repo events when using ght-retrieve-repo
* Non-recursive retrieval of pull requests leads to 1/3 API calls
* Custom rate limits for Github API tokens
* Tooling for MySQL dumps in CSV files
* General bug fixes and cleanups

v0.10

  * Base class for multiprocess queue clients
  * Make retrieval of pull request commits faster
  * Drop table forks
  * Remove merged and user_id fields from pull_request
  * Ruby 2.0.0 support
  * Fast(er) path to project forking. Copy commits from base repo
  * instead of getting them from       Github.
  * Github API tokens for authentication
  * Support for disabling the cache
  * A new real-time log analyzer
  * Mark fake users as such
  * Organization members are now retrieved
  * Authenticated MongoDB connections
  * General bug fixes and cleanups

v0.9

  * Remove dependency to EventMachine-based AMQP client. We now use bunny.

v0.8.1

* New tool to retrieve specific entities and their dependencies
* New tool to retrieve repositories en masse
* Support for resuming when exception occurs while processing items in
* loops
* Support for finer grained transactions when processing large entities
* Commit comments are now indexed per owner/repo (was just by comment id)
* Remove the unused daemon mode
* Various exception fixes and more detailed logging

v0.8

* Retrieve and process issue labels
* Retrive and process actors for pull request events
* Retrieve pullreq comments for projects with no issuetracker
* Better heuristics for avoiding dulpicate entries in pull request
* histories
* The event loader now loads events by IDs to reduce pressure to the
* queue
* Compound indexes in MongoDB by default
* X-RateLimit-Reset header support
* Remove lots of dead code, general cleanups

Version 0.7.2

Version 0.7.1

Version 0.6

Changes

* Generic methods for retrieving items that are bound to repositories
* Processing of pull requests with commits, comments and history
* Processing of project forks
* New tool (ght-load) to filter and load events to the queue
* New tool (ght-rm-dupl) to delete duplicate entries from
* collections (events & commits supported)
* Project wide requesting result caching for multi-page requests
* Better logging in various places
* Better defaults for MySQL (UTF8 + InnoDB tables)
* Commits are now seperated from projects. Project forks can
* share commits.
* Support for setting the IP address to use for retrieval
* on multi-homed
* Compatibility with Ruby 1.9 (now default) and JRuby
* Proper modularization, following the cake design
* pattern
* Never retrieve arrays of results from MongoDB