Package org.epics.archiverappliance.retrieval.postprocessors

Class OptimizedWithLastSample

java.lang.Object

org.epics.archiverappliance.retrieval.postprocessors.OptimizedWithLastSample

All Implemented Interfaces:

FillNoFillSupport, PostProcessor, PostProcessorWithConsolidatedEventStream

public class OptimizedWithLastSample
extends Object
implements PostProcessor, PostProcessorWithConsolidatedEventStream, FillNoFillSupport

OptimizedWithLastSample expects one parameter at initialization, which is the number of requested points. If there are less samples in the time interval than requested (with a certain deadband), all samples will be returned. If there are more samples than requested, the samples will be collected into bins. Mean, std, min, max and count of each bin is calculated and returned as a single sample. This differs from the Optimized post processor in that if a bin is empty, instead of repeating the last bin with samples, the bin uses the last value of the last recorded sample instead (as mean, min and max; stddev is zero and number of samples is also zero).

Field Summary

Fields

Modifier and Type	Field	Description
protected LinkedHashMap<Long,org.epics.archiverappliance.retrieval.postprocessors.SummaryStatsPostProcessor.SummaryValue>	consolidatedData

Constructor Summary

Constructors

Constructor	Description
OptimizedWithLastSample()

Method Summary

Modifier and Type	Method	Description
Callable<EventStream>	customStatsWrap(Callable<EventStream> callable)
void	doNotInheritValuesFromPrevioisBins()
long	estimateMemoryConsumption(String pvName, PVTypeInfo typeInfo, Timestamp start, Timestamp end, javax.servlet.http.HttpServletRequest req)	Estimate the amount of memory required for the data generated by the post processors.
LinkedList<TimeSpan>	getBinTimestamps()
EventStream	getConsolidatedEventStream()
long	getEndBinEpochSeconds()
String	getExtension()	This is the full form of the identity for the post processor and includes any parameters for the post processor.
String	getIdentity()	The string used by clients to identify this post processor when making retrieval requests.
long	getStartBinEpochSeconds()
void	initialize(String userarg, String pvName)	Initialize this post processor for the given PV and request parameters.
Callable<EventStream>	wrap(Callable<EventStream> callable)	Primary data generation method.
boolean	zeroOutEmptyBins()	For some post processors, we do fill empty bins but with zeroes instead.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

consolidatedData

protected LinkedHashMap<Long,org.epics.archiverappliance.retrieval.postprocessors.SummaryStatsPostProcessor.SummaryValue> consolidatedData

Constructor Detail

OptimizedWithLastSample

public OptimizedWithLastSample()

Method Detail

initialize

public void initialize(String userarg,
                       String pvName)
                throws IOException

Description copied from interface: PostProcessor

Initialize this post processor for the given PV and request parameters.

Specified by:

initialize in interface PostProcessor

Parameters:

userarg - This is the full form (extension) of the identity for the post processor.

pvName - The name of PV

Throws:

IOException -  

estimateMemoryConsumption

public long estimateMemoryConsumption(String pvName,
                                      PVTypeInfo typeInfo,
                                      Timestamp start,
                                      Timestamp end,
                                      javax.servlet.http.HttpServletRequest req)

Description copied from interface: PostProcessor

Estimate the amount of memory required for the data generated by the post processors.

Specified by:

estimateMemoryConsumption in interface PostProcessor

Parameters:

pvName - The name of PV

typeInfo - PVTypeInfo

start - Timestamp

end - Timestamp

req - HttpServletRequest

Returns:

Estimated Memory comsumption

doNotInheritValuesFromPrevioisBins

public void doNotInheritValuesFromPrevioisBins()

Specified by:

doNotInheritValuesFromPrevioisBins in interface FillNoFillSupport

zeroOutEmptyBins

public boolean zeroOutEmptyBins()

Description copied from interface: FillNoFillSupport

For some post processors, we do fill empty bins but with zeroes instead.

Specified by:

zeroOutEmptyBins in interface FillNoFillSupport

Returns:

boolean True or False

getConsolidatedEventStream

public EventStream getConsolidatedEventStream()

Specified by:

getConsolidatedEventStream in interface PostProcessorWithConsolidatedEventStream

getStartBinEpochSeconds

public long getStartBinEpochSeconds()

Specified by:

getStartBinEpochSeconds in interface PostProcessorWithConsolidatedEventStream

getEndBinEpochSeconds

public long getEndBinEpochSeconds()

Specified by:

getEndBinEpochSeconds in interface PostProcessorWithConsolidatedEventStream

getBinTimestamps

public LinkedList<TimeSpan> getBinTimestamps()

Specified by:

getBinTimestamps in interface PostProcessorWithConsolidatedEventStream

getIdentity

public String getIdentity()

Description copied from interface: PostProcessor

The string used by clients to identify this post processor when making retrieval requests. For example to identify the FirstSamplePP postprocessor, users would add a pp=firstSample to the request for data. The situation is a little more complex; if a post processor has parameters then it needs to combine these into a string and offer that as an extension. The identity is just the starting part of this.

1. For example, pp=firstSample_600 asks the server to sparsify with an interval of 600 seconds.
2. identity is firstSample.
3. extension is firstSample_600.
4. User specifies firstSample_600.
5. ETL caches the data as firstSample_600 if asked to.

Specified by:

getIdentity in interface PostProcessor

Returns:

identify  

getExtension

public String getExtension()

Description copied from interface: PostProcessor

This is the full form of the identity for the post processor and includes any parameters for the post processor. The exact format and interpretation of this is left to the post processor; however the convention is to use underscores to separate the params and have them in a specific order.

Specified by:

getExtension in interface PostProcessor

Returns:

extension  

See Also:

PostProcessor.getIdentity()

wrap

public Callable<EventStream> wrap(Callable<EventStream> callable)

Description copied from interface: PostProcessor

Primary data generation method. Using the event stream provided, do your magic and generate processed data

Specified by:

wrap in interface PostProcessor

Parameters:

callable -  

Returns:

EventStream  

customStatsWrap

public Callable<EventStream> customStatsWrap(Callable<EventStream> callable)