Logtools

Added the 'logfilter' utility with support for regexp/aho-corasick based matching. See README/documentation for examples
Added the 'percentiles' utility for quickly computing percentiles out of input data
Added unit-tests for the logfilter functionality
Added dependency on the acora string matching library (Implements Aho-Corasick in CPython)
Minor fixes to qps utility (formatting and parameter naming in README
example)
Updated README to be markdown based, cleaned up and edited

Added missing -d and -f parameters to logsample for enabling weighted reservoir sampling
Bugfix to geoip --filter default value triggering a filtering by default (regression introduced in in 0.7.1)

Small fix with geoip --filter default value

Added -q flag to urlparse (See README/documentation for examples)
Added -f flag to geoip, allowing one to filter inplace for a specific country (e.g -f 'United States')
Added --decode options to urlparse that allows one to url-decode (unquote) input stream
Added sumstat utility/API for generating some common summary statistics for a given histogram-like logfile. See README for examples
Added 'colsum' script utility that does an arithmetic summation over a column of data
Added stripping of leading whitespace from output of 'aggregate' script
Updated documentation, Added more examples for new utilities in README
Updated unit-tests to reflect additions/changes to tools
Unicode encoding in logplot line-plot using matplotlib backend (would crash with encoding problems in some cases)
Unicode encoding for the logparse, fixes some cases of errors, e.g when using JSONParser for a JSON that has non-string types (e.g integers)
Fixed bug in use of --ignore for logparse
Default value for --delimiter is now ' ' across commands (logjoin etc.)

Added --header flag to logparse, will prepend output with field names. useful e.g for dumping to CSV file
Added --parser and --ip-ua-fields options to filterbots, allows routing input through a parser for filtering bots
Updated logmerge to support indexed/named field-based merge-sorting when using --parser. See README for examples
Added Histogram plot type support to logplot (--type hist)
Added urlparse utility/API. See README for details
Added logjoin utility/API. See README for details
Added logtools.join_backends module with SQLAlchemy backend implementation (added sqlalchemy to install_requires)

Added qps utility - Parse timestamps from log stream and calculate qps over sliding time window
Added Prefix/Suffix matching to filterbots - Specify p'<prefix>' or s'<suffix>' in useragents file to use this feature
Added comment support for filterbots useragents/ips files - use '' in start of line
Added JSONParser to parsing module, can now use JSON-formatted logs e.g for logmerge, logparse etc.
Added --ignore flag to logparse, surpresses missing field errors and skips those lines
Added the 'timeseries' plot type for plotting timeseries data
Added the matplotlib plotting backend supporting pie/line/timeseries charting
Added data/examples subdirectory, included example bots_useragents.txt files
Can now specify fields by field name in -f parameter for the logparse method/console script
Can now specify multiple comma-separated field names/indices in logparse method
Fixed memory bug resulting from un-needed use of readlines() on file descriptors
Changed programmatic interface for most methods to use explicit parameter naming instead of generic (opts, args)
Updated unit-tests, examples and README to reflect most of latest functionality