Shark Logging Facilities

The Shark machine learning library offers a framework for logging purposes as a part of its default installation (see Logger.h). Its purpose is twofold: On the one hand, debug, status and error information can be shown and stored persistently in an abstract manner. On the other hand, external components like GUI framework can rely on the framework to redirect Shark logging messages to custom handlers, e.g., GUI log views.

The design of the logging framework is simple and straightforward. An instance of class Logger is required to report messages to the user. Every logger instance relies on a set of handlers (see Logger::AbstractHandler) for formatting and reporting purposes. Loggers are announced within the whole library and queryable by their name by means of the singleton LoggerPool.