Reports

A set of interfaces for reporting perceiver outcomes during processing. More...

Collaboration diagram for Reports:

Classes
class	Always
	Class that always triggers a report. More...
class	Announcement
	Base/abstract announcer class. More...
class	Assignment
	Assignment class for communication between BeatReporter and Editor. More...
class	BeatReporter
	Specialized reporter class that works with an editor. More...
class	BuildCfgReporter
	Configuration instance for a Reporter. More...
class	BuildCfgTrigger
	Configuration instance for building a Trigger. More...
class	CfgAnnouncement
	Configuration instance for an Announcement. More...
class	CfgBeatReporter
	Configuration instance for a Reporter. More...
class	CfgChannel
	Configuration instance for a Channel. More...
class	CfgCommentary
	Configuration instance for Commentary. More...
class	CfgDistTrigger
	Configuration instance for a Trigger. More...
class	CfgReporter
	Configuration instance for a Reporter. More...
class	CfgRunningCommentary
	Configuration instance for RunningCommentary. More...
class	CfgToFile
	Configuration instance for a Channel. More...
class	CfgTrigger
	Configuration instance for a Trigger. More...
class	Channel
	Base/abstract channel class. More...
class	Commentary
	Commentary class, which does not transcribe the signal but passes it along in some non-text format. More...
class	Editor
	Editor class manages multiple reporters and curates information going to the output channel based on reporter commentary. More...
class	Falling
	Class that triggers a report when binary state changes, on falling edge. More...
class	onChange
	Class that triggers a report when the state changes. More...
class	onMatch
	Class that triggers a report when the state matches a target state. More...
class	Reporter
	Base/abstract reporter class. More...
class	Rising
	Class that triggers a report when binary state changes, on rising edge. More...
class	RunningCommentary
	Commentary class that also keeps track of past information and reports it all out when requested. At that point, it should wipe the past information and start collecting anew. More...
class	toCSV
	Save to CSV formatted file channel. More...
class	toFile
	Save to file channel. More...
class	Trigger
	Base/abstract trigger class. More...
class	whenClose
	Class that triggers a report when current state is close to target state. More...
class	whenDiffers
	Class that triggers a report when current state differs from previous state. More...
class	whenFar
	Class that triggers a report when current state is far fromo target state. More...
class	whenSimilar
	Class that triggers a report when current state is similar to previous state. More...

Detailed Description

A set of interfaces for reporting perceiver outcomes during processing.

Synopsis : The reporting package implements a news or event reporting system whose output is similar to logging but customized for downstream data analysis. In principle, it extends or supports the work of a Monitor by providing interpretable output of Monitor status. The language surrounding this implementation comes from that of a news publication and its organizational approach to investigating, reporting, and ultimately publishing news/information.

Details : A collection of interfaces that establish flexible reporting schemes modeled after the python logging API. Unlike logging, reporting is considered part of the standard processing routine when a reporter is instantiated. Reports provide state or signal dependent outputs that support analysis, action responses, etc. In contrast, logging is generally meant to output information needed to understand the state of code execution during debugging analysis or for post-execution study. There is some overlap in the post-execution part, however when logs are studied after the fact, the stream needs to be decomposed to be interpreted. In contrast, a report already outputs the data needed and can immediately be studied. It is even possible to respond in real-time as the reporting may be through a channel that can be captured in real-time. An example of this is if the output is to a ROS message.

Similar to the logging API, constructing a report consists of building out and linking several components. The components are a channel, a trigger, and an announcer. A channel implements a reporting scheme. In the python logging API, a channel is similar to the combination of a handler and a formatter. When a report is triggered, then the channel needs to take care of outputting the proper message to the reporting output "stream."

A trigger determines when a particular state or signal should be reported. In the python logging API, a trigger is similar to a filter and is tested before constructing the message or moving forward on state/signal reporting. Usually in logging, some external process requests the log to happen, and the filter applies pre-established logic to determine if the logging should be permitted.
Here, there is a continuously submitted state/signal and the choice must be made as to whether the state should be reported or not.

An announcer generates the output that will be sent through the reporting channel. The most basic version should be plain text, but we can envision other versions like JSON or csv compliant outputs. Any data storage scheme is valid as long as it permits some form of sequential outputting and does not require backtracking.

While the code is attached to the perceiver, it is general enough that other classes or packages may implement a reporting scheme. These classes/packages should provide their own derived implementations.