Commits

Show all
Author Commit Message Labels Comments Date
Mark Roddy
Updated to reflect new constructor signature of the MarkovChain class.
Branches
logparse-refactor
Mark Roddy
Updated MarkovChain class to use new pidginlog class.
Branches
logparse-refactor
Mark Roddy
Added tests for the PidginLogParser class.
Branches
logparse-refactor
Mark Roddy
Added logs property and the ability to iterate over all log records to the PidginLogParser class.
Branches
logparse-refactor
Mark Roddy
Reorder the appearance of the PidginLogRecord and PidginLog classes so that classes appear in the order in which they consume eachother. Did the same with the test cases.
Branches
logparse-refactor
Mark Roddy
Merging trunk changes 87-93 into the logparse-refactor branch
Branches
logparse-refactor
Mark Roddy
Updated test target to reflect rename of test modules.
Branches
logparse-refactor
Mark Roddy
Implemented class PidginLog and added new class PidginLogRecord which represents a single entry in a Pidgin log file.
Branches
logparse-refactor
Mark Roddy
Added two new classes PidginLog and PidginLogParser. PigdinLog represents a single pidgin log container (i.e. - a log file), and the PigdinLogParser takes PidginLog objects and extracts converstations from them. These classes are currently only partly implemented.
Branches
logparse-refactor
Mark Roddy
Creating branch to work on refactorying how log files are parsed to make this more flexible, eaiser to report issues processing log files, and eaiser to fix parsing bugs as they are found. Working on this in a branch as existing functionality will likely break in the course of this work."
Branches
logparse-refactor
Mark Roddy
Added 'test_' prefix to each module containing unit tests.
Mark Roddy
Removed control characters added by a windows based text editor.
Mark Roddy
Refactored code to prepend a list of None values to the word list so into a single line of code which makes the intention clearer. Also refactored the while loop into a for loop.
Mark Roddy
Reinstanting function removed for v0.1.0 release.
Mark Roddy
Ran spell check on readme file. I don't know how to spell.
Tags
v0.1.0
Mark Roddy
Removed the _loadmap() funciton as it is not part of v0.1. It will be reinstated for version v0.2 for which work will begin in a few hours possibly.
Mark Roddy
Changed the project name to be all lower case.
Mark Roddy
Changed the default source distribution type to be built by the 'installer' target to zip instead of gztar. This will make it easier on Windows.
Mark Roddy
Added a short section of implementation notes to the readme file.
Mark Roddy
Added description of the required command line argument 'screenname'
Mark Roddy
Pulling out explination of Markov Chain from the command line usage section. This is better suited in its own section.
djtumolo
Mark Roddy
Filled in basic info in README file with program usage and installation instructions.
djtumolo
added dummy function for offline processing of logs
Mark Roddy
Added tokenization of the beginning of sentences. This is done via a list of tokens (None value) equal to precision. The first word of each sentence is then mapped to this list. This also simplifies an amount of the code as the special case of sentences shorter then precision is no longer an issue.
Mark Roddy
Added message to test so that the actual type of the piece of data is shown
Mark Roddy
Added explicit type for the chain length command line argument. Without this, a string type was being passed into the precision value for the markovchain class resulting in incorrect behavior.
Mark Roddy
Added inspection of the entire wordmap to check that keys/value pairs are of the correct type.
Mark Roddy
Added unit test to check the response.MarkovChain._addSentenceToMap() function.
Mark Roddy
Made all words in the sentence lower case before passing it to be mapped. This ensures that a word is matched regardless of its source case.
  1. Prev
  2. Next